Anyone interested in Shakespeare must care about collaboration. Modern attribution scholarship agrees that Shakespeare’s writing can be found in at least forty plays: the thirty-six in the First Folio, plus Pericles, The Two Noble Kinsmen, Edward III and Sir Thomas More. Of those forty plays, the four not included in the Folio are undeniably collaborative.¹ Within the Folio itself, another four – Timon of Athens, Henry VIII/All Is True, Titus Andronicus and 1 Henry VI – are now accepted as collaborative by all the leading attribution specialists.² Eight plays out of forty: that’s twenty per cent of the canon. Hugh Craig and John Burrows have produced compelling new statistical evidence that the other two parts of Henry VI are also collaborative, as most attribution scholars in the eighteenth, nineteenth and early twentieth century contended.³ Ten out of forty: that’s twenty-five per cent of the canon. If we accept the growing consensus that Shakespeare wrote the additions to the Spanish Tragedy published in 1602, and parts of Arden of Faversham and The History of Cardenio, then he collaborated in 13 out of the extant 43 plays he worked on: that’s 30 per cent.⁴ Those who accept the claims of the 2013 RSC edition of Collaborative Plays would add five more.⁵ Modern scholarship gives us a larger Shakespeare canon, but also a larger proportion of collaborative work. Moreover, two plays originally written by Shakespeare alone – Macbeth and Measure for Measure – were apparently adapted after Shakespeare’s death.⁶ That leaves just twenty-eight plays that survive in texts written entirely by Shakespeare.⁷ Shakespeare’s is the only hand in less than two-thirds of the plays that Shakespeare had a hand in.

We can be interested in collaboration without mastering Principle Component Analysis, Fisher’s Exact Test, chi-square, degrees of freedom, historical sociolinguistics, plagiarism software, palaeo-graphy, chainlines or watermarks. Rowe did not believe that all of Pericles was written by Shakespeare, but ‘some part of it certainly was, particularly the last Act’;⁸ Coleridge denied that Shakespeare could have written the first speech of 1 Henry VI but conjectured that he did write the additions to The Spanish Tragedy; Tennyson identified John Fletcher as the author of some scenes of Henry VIII; Swinburne insisted on Shakespeare’s presence in Arden of Faversham. On the basis of their own sensitivity to verse style, each poet floated an intuitive hypothesis, which has subsequently been tested and confirmed repeatedly, by a variety of independent empirical experiments, conducted by people who are not poets. Attribution scholarship is a determinedly dull technical discipline, like physical archaeology or the chemical analysis of pigments. Caravaggio signed only one of his paintings, but you can admire, teach and write about Caravaggio’s art without becoming an expert in the scientific techniques that established his canon. Likewise, I will here take for granted the consensus of the leading living experts about what Shakespeare wrote, and begin with a different, critical question: why did he collaborate?

Both prevailing answers to that question are economic. The postmodernist answer celebrates collaboration because, it claims, proprietary individual authorship was a capitalist ideology not written into law until the Enlightenment, and therefore irrelevant to the cooperative mentalité of early modern playwrights. Shakespeare collaborated because he didn't know any better.⁹ He belonged to the innocent race before the bourgeois flood. Undeniably, the evolution of copyright and the economics of the book trade affected the subsequent history of Shakespeare’s reputation. But although theatre since the time of the Athenians has required actors, musicians, dancers, choreographers, painters, carpenters, costume-makers, financiers and all manner of back-stage crew, before Shakespeare’s lifetime it had never routinely required more than one playwright per play. Even then, as Jeffrey Knapp has demonstrated, ‘Collective play-writing was never the norm for Renaissance drama, practically or conceptually.’¹⁰

The collapse of the historically indefensible postmodernist hypothesis has led to a resurgence of the only available alternative: the neoclassical, formalist explanation, which denigrates collaboration as itself a capitalist intrusion upon the natural and desirable state of individual artistic autonomy. According to this theory, the economic dominance of actors and proprietors forced playwrights to collaborate. Shakespeare’s plays include material by other writers because the men who paid the piper fiddled with the tunes. ‘If we give into this opinion’, Pope declared in 1725, ‘how many low and vicious parts and passages might no longer reflect upon this great Genius, but appear unworthily charged upon him?’¹¹ Likewise, Bart van Es, in 2013, explains six of Shakespeare’s collaborations in terms of ‘Shakespeare’s working conditions in the early 1590s’, dominated by the ‘financial pressure’ and ‘constant haste’ of a system where writers ‘were the employees of the acting companies’.¹²

But the neoclassical economic claim is as suspect as the postmodernist one. Since most commercial plays of the period were apparently written by a single author, the theatres were unsuccessful in imposing their alleged collaborative agenda. Anyway, why would theatres want multiple authors? When Knapp claims that ‘collective playwriting helped speed up the process of satisfying’ the commercial theatre’s ‘demand’ for new material, he is simply echoing the assertion by Brian Vickers that ‘the need to keep the theatrical companies supplied with material must have been one reason for co-authorship’.¹³ Vickers himself supports this thesis by citing, more than once, a 1927 article by Charles Sisson.¹⁴ Sisson discovered legal documents about a lost play called ‘The late Murder in White Chappell, or Keepe the Widow Waking’, which ‘was contrived and written by Wm Rowley, Jon ffoord, John Webster, and Tho: Decker’. This indisputably collaborative play was based upon two recent (‘late’) and local (‘White Chappell’) events, which took place between April and August 1624; the play was licensed in September, and both Sisson and Vickers agree that it must therefore have been written ‘at great speed’ by all four authors. Vickers then notes, on the basis of Henslowe’s account books, that the six weeks allowed for Keep the Widow Waking was ‘a not-unusual period of time’ for writing a play. Combining the evidence of Henslowe and Sisson, Vickers generalizes that collaborative dramatists ‘must have worked in permanent haste’.

Can these grand claims be supported by a few self-serving depositions in a lawsuit about a lost late-Jacobean play? The timetable of composition is less certain than Sisson and Vickers assert.¹⁵ But even if we accept their conjectures, none of the very specific circumstances that encouraged rapid production of Keep the Widow Waking is relevant to any Shakespeare play that attribution specialists have identified as collaborative. Vickers is correct when he claims that, in Henslowe’s accounts, ‘plays were normally finished in four to six weeks’ (Co-author, 43). But that sentence, tellingly, is about all plays, not just collaborative ones. Vickers cites Neil Carson for this statistic, but he omits Carson’s preceding and following sentences: ‘However organized, the playwrights worked with considerable speed. Henslowe’s accounts indicate that plays were normally finished in four to six weeks. Drayton promised to complete a book in a “fortnyght”.’¹⁶ Likewise, Vickers ignores the fact that Sisson’s book also contains a chapter about another lost play produced in haste to exploit a topical scandal: The Old Joiner of Aldgate, written by George Chapman, alone. Nor were Drayton and Chapman the only playwrights capable of writing quickly. Ben Jonson ‘fully penned’ the very long text of Volpone in five weeks (85 lines per day), and the biggest hit of the entire period, Middleton’s A Game at Chess, must have been written in five weeks or less (375 lines per week).¹⁷ Shakespeare allegedly wrote The Merry Wives of Windsor in two weeks.¹⁸ Noël Coward wrote his most admired play, Private Lives, in four days.¹⁹ Alan Ayckbourn began writing his enormously popular Bedroom Farce on a Wednesday, completed it that Friday, typed it all up on Saturday, and began rehearsals on Monday.²⁰ Compositional velocity is not a function of the number of playwrights involved. Neither is commercial success.

Why should a theatrical demand for new material create a demand for collaboration? Carson pointed out that seven playwrights did all the writing for Henslowe’s crowded fall and winter season of 1599–1600; an eighth actor-playwright (the older Robert Wilson) joined them to collaborate on one play. Seven playwrights working alone on separate plays should, theoretically, be able to produce as many plays as seven playwrights working together on collaborative plays. The only obvious gain in productivity here, created by collaboration, is the single collaborative contribution of Robert Wilson. But Wilson worked on fifteen other Henslowe plays from spring 1598 to summer 1600, so he clearly belongs to the same ensemble of writers. From the perspective of theatre management, why not have eight playwrights writing separately, instead of eight playwrights writing collaboratively with each other?

With professional playwrights writing for commercial theatres, collaboration cannot be explained by simple economies of time or personnel. The motive cannot be quantitative. It’s not about the numbers. It must be qualitative, and therefore phenomenological. Collaboration in some way improved the quality of the human experience. Analysing Henslowe’s records, Carson could detect only one statistical difference between single-author plays and plural-author plays: the collaborative ones were more likely to get finished (57–8). If theatres had an economic motive for encouraging collaboration, it was not because co-authors worked faster, but because they were more often able to achieve closure. Why? Carson does not venture an answer. Finishing a play for which you were the sole author (and therefore the sole payee) would have provided a greater financial incentive to finish. Therefore, the economic motive cannot have been paramount: playwrights apparently had greater or more effective incentives to finish plays for which they received smaller, divided payments. Why? Historically, some playwrights, some of the time, have found it stimulating, socially and imaginatively, to work with what Nashe called a ‘fellow writer’.²¹ Apparently, at least some of the time, the social relationship of one Elizabethan playwright to his fellows mattered more to him than his economic relationship to Henslowe. Some of the time, collaboration created a different, more satisfying working experience for the playwrights themselves.²²

Collaboration might also, theoretically, have increased the quality of the product. Acting companies could have believed that collaboration produced scripts that improved the experience of acting in them, and therefore improved the experience of audiences watching and hearing them. Economic pressure might thus, theoretically, have created an incentive to produce a better product. This possibility may seem counter-intuitive, and many critics reject it out of hand. Sisson had nothing but contempt for The Late Murder in Whitechapel, or Keep the Widow Waking: ‘Incongruous as was the linking together of these two stories into one play, in which no possible dramatic connection could give them any artistic unity, it was evidently sufficient for the dramatist exploiting topical interest that the two wretched criminals involved lay in the same gaol together and were led forth on the same day to stand at the bar of judgment’ (82). Vickers, likewise, asserts that ‘the speed with which the play was staged meant that the four dramatists had little time for consultation’ with each other (315) – thus explaining what he sees as a lamentable lack of artistic unity in all collaborative plays.

It should be obvious that we can say nothing intelligent about the artistic unity of a lost play. Nor can we say anything useful about the lost conversations of one playwright with another. How does Vickers know that four professional playwrights, all living within the much smaller space of early modern London, in easy walking distance from each other, had ‘little time’ for interaction? In six weeks none of them could find any time to talk to each other? Are we to imagine them, walled up in separate rooms for a month and a half, never venturing out to share a meal, an ale or a chat? Should we assume that playwrights, people who make a living writing dialogue, are by nature anti-social? Isn't it likely that some playwrights, then as now, were capable not only of ‘empathic listening’ but also of mutually productive and interactive ‘dialogic listening’?²³ The fact that such conversations were not recorded does not mean that they never took place.

Vickers jumps immediately from Keep the Widow Waking to Sir Thomas More (Co-author, 34–43). Both plays provide documentary evidence of commercial theatre practice. Like the lost 1624 play, the manuscript adaptation of More contains the work of four playwrights, one of whom is Thomas Dekker; the others are Chettle, Heywood and Shakespeare.²⁴ However, the adaptation of More has not been linked to any topical scandal that needed to be exploited quickly. Even if speed had been necessary, adapting the play required much less work than writing a new one from scratch, and should have taken much less time to write. Moreover, for most of the period from spring 1603 to the end of 1604 (when Jowett and other recent scholars date the adaptation), plague closed the London theatres. No public performances means no urgent demand for new material. Time pressure, that catch-all economic explanation for collective writing, cannot explain the manuscript of Sir Thomas More. So, why four playwrights, instead of one? Why collaboration at all?

We could ask that question, and distinguish those four hands in the manuscript, even if we could not connect those hands to particular playwrights working simultaneously in the commercial theatres of early modern London. In the manuscript, authorship is not a theory about cultural authority. The manuscript does not contain Shakespeare’s name, and the British Library originally acquired it without knowing that Shakespeare had anything to do with it. Of course, the value of the manuscript rocketed once scholars began to identify Hand D as Shakespeare, just as sales of a crime novel called The Cuckoo’s Calling rocketed when its author, ‘Robert Galbraith’, was outed as a pseudonym for J. K. Rowling. Like Rowling, Shakespeare is a lucrative brand name. In the twenty-first century, many more people will buy an edition of Sir Thomas More, or buy tickets to a performance of Sir Thomas More, because the trademark ‘Shakespeare’ is attached to it. But that is a fact about the subsequent history of the text. It tells us nothing about Shakespeare, or collaboration, in the early modern London theatre.

Let’s begin, therefore, with ‘Hand C’, which remains anonymous, but does appear in other extant playhouse documents. Hand C might be a theatrical scribe, or (less likely) an unidentified playwright, or some combination of the two. There needs no ghost, come from the grave, to tell us that theatre is a collaborative art-form, but Hand C usefully incarnates the commercial and intrinsically social institution of a joint-stock theatre company.²⁵ His handwriting illustrates one particular kind of collaborative interaction.

Shakespeare’s three pages of the manuscript – the smallest of his known contributions to a collaborative play – contain 1266 words in his own handwriting, including stage directions and speech prefixes, but excluding the eighteen words that he himself deleted in the course of his writing. Hand C subsequently altered Shakespeare’s 1266 words thirteen times. In nine places he changed Shakespeare’s speech prefixes to bring them into line with the rest of the play, replacing Shakespeare’s anonymous crowd with the specific individuals established by the other playwrights. Once, Hand C added the word ‘Enter’ before a speech prefix, to clarify the stage logistics. These ten changes clearly belong to the necessary business of performing a play: telling actors when to enter, identifying which lines are spoken by which actors. Another change corrects Shakespeare’s tautology ‘letts us’ to ‘letts’; this necessary correction might have been made by any scribe or editor. Together, these eleven interventions alter the text no more than Shakespeare himself did, deleting words when he changed his own mind. More significant, from an editorial or dramaturgical point of view, is Hand C’s deletion of 26 consecutive words:

In context, in the manuscript, with interlineations and deletions and an unpunctuated relationship to what goes before and after, this is a confusing and superfluous passage. Hand C replaced it with four unexciting but clear transition words of his own: ‘tell me but this’. This is the kind of intrusion that, we can imagine, would make Ben Jonson furious. Nevertheless, these twenty-six deleted words constitute only 2 per cent of Shakespeare’s original handwritten text, and Hand C’s four added words are less than one third of one per cent. Hand C tinkers with what Shakespeare wrote – and with what the other playwrights wrote. Whether scribe or playwright, his function was to coordinate the work of all the other hands in the manuscript.

Notoriously, Shakespeare’s own contribution to More is not well connected to the work of the three other adapters; therefore, at the time he wrote his three surviving pages, he was not intensely interacting with Chettle, Dekker or Heywood, and his primary motive for writing those pages does not seem to have been social. Either he wanted to write that particular episode, or someone else thought that the play would be improved if he wrote that particular episode. We cannot know whether the initiative came from Shakespeare or someone else, but it hardly matters, because either way the motive was aesthetic, and either way Shakespeare was willing. He was not forced. Although for twenty years Shakespeare was what Gerald Bentley called the company’s ‘attached dramatist’, he did not write the company’s additions to their expropriation of Marston’s Malcontent, which were provided instead by John Webster, a younger playwright with, at the time, no known previous connection to the company. If Shakespeare in 1603–4 chose not to write additions to The Malcontent, Shakespeare in 1603–4 could also have chosen not to write additions to Sir Thomas More. In fact, by 1603–4 Shakespeare had more economic and artistic freedom than any other professional playwright in London. Consequently, the best explanation for Shakespeare writing those three pages is that something about one episode in Sir Thomas More was particularly appealing or appropriate for him to write – by contrast with the material added to Sir Thomas More by Chettle, Dekker and Heywood, which seemed appealing or appropriate for each of them, but not for him.

So, what is it about that scene that seemed to him, or someone else, particularly Shakespearian? To begin with, it is not the beginning of the play. Shakespeare’s three pages contribute to the sixth scene of the play; editors with a fetish for act divisions have placed it somewhere in the middle of Act 2. The play’s original first scene of urban unrest had been so thoroughly eviscerated by the censor that it had to be replaced, or abandoned. Shakespeare did not write a replacement. Instead, Heywood added new material in scene 4, and a new scene 5 was supplied by Hand C (perhaps copying and modifying something written by Chettle). Only then does Shakespeare appear.

Shakespeare’s contribution to More fits a pattern found in all his known or suspected collaborations from the beginning of his career until the early Jacobean period.²⁶ According to the most recent attribution scholarship, Shakespeare did not begin Arden of Faversham, Edward III, Titus Andronicus, any of the three Henry VI plays or Pericles. For most of his career, Shakespeare was less interested, or less accomplished, in setting up a situation than in developing one. Playwright David Edgar, without any knowledge of this pattern, contends that ‘Shakespeare wasn't skilled at exposition.’²⁷ No modern Shakespeare scholar would dare say so, but the Royal Shakespeare Company apparently agrees: their 2013 productions of All’s Well That Ends Well, Richard II and Othello (the only ones I saw) all interpolated new material to jump-start the play. So did the 2013 Chicago Shakespeare Theatre’s Henry VIII and the Goodman Theatre’s 2013 Measure for Measure. All productions at Shakespeare’s Globe now begin with an interpolated, energetic musical performance; in their outstanding 2013 Midsummer Night’s Dream director Dominic Dromgoole also added an introductory dumbshow of the war between Theseus and the Amazons. We might perhaps agree that ‘Shakespeare wasn't as skilled at exposition’ (much virtue in as). Certainly, he owes more of his global reputation to an extraordinary gift, demonstrated in More, for writing scenes of intense conflict.

In Shakespeare’s three pages, More single-handedly quells a rioting mob, one which shouts down a sergeant, a mayor and two earls. Scholars have compared this episode to other mob scenes in Shakespeare’s works, and in certain respects it does resemble two scenes in Julius Caesar and, especially, the opening scene of Coriolanus. Those parallels help establish Shakespeare’s authorship of the episode, but they have also been used to explain Shakespeare’s participation in the project: he had ‘a specific expertise in staging popular uprisings’ (Jowett, 379), in a way that excited spectators but did not disturb censors.

But the episode in More also differs from the mob scenes in Caesar and Coriolanus in one crucial respect, which connects it to Shakespeare’s aesthetic much more broadly. Shakespeare’s three pages are entirely dominated by the play’s charismatic male protagonist, an ‘abnormally interesting’ person.²⁸ Thomas More speaks 788 of the 1213 words that Shakespeare wrote for actors to speak: 65 per cent of the dialogue, including one speech 44 lines long.²⁹ The consensus of attribution scholarship is that Shakespeare also wrote More’s first meditative soliloquy, transcribed by Hand C in scene 8, a turning point for the play and for More’s career. More recently, Jowett has argued that Shakespeare wrote parts of a second soliloquy for More, in scene 9, also transcribed by Hand C.³⁰ But even if we disregard those two soliloquies, the pattern is clear. What interested Shakespeare was Thomas More. From Richard III to The Tempest, Shakespeare wrote a succession of exceptionally long, exceptionally dominant roles for male protagonists. As Scott McMillin first pointed out, the part of More is comparably long, and must have been written for one of the few Elizabethan actors capable of memorizing and mastering so many lines. Edward Alleyn and Richard Burbage are the most plausible candidates.³¹ But for Othello (1603–4) and Volpone (1605–6) the King’s Men required two such actors, to play the paired protagonists Othello–Iago and Volpone–Mosca. This change in company practice might well be connected to the arrival of John Lowin, who at some point in the second half of 1603 left Worcester’s Men (working for Henslowe at the Rose) to join the King’s Men.³² Perhaps he brought the manuscript of More with him.³³

More’s dominance is nowhere more evident than in the three pages Shakespeare wrote. Like many of Shakespeare’s most famous roles, More in this scene enacts, embodies, the political, imaginative and charismatic power of male eloquence. By contrast with Shakespeare’s history plays, much of the rest of Sir Thomas More presents, as Jowett says, ‘a strong sense of a London locality’, of London as ‘a city of the people’, and of More himself as a Londoner among Londoners. Vittorio Gabrieli and Giorgio Melchiori go so far as to claim that London is the ‘protagonist’ of Act 2.³⁴ Twenty specific London localities are mentioned by name. But not in Shakespeare’s three pages, which do not even contain the word ‘London’. The original play, and the other additions, can be clearly linked to emergent genres of city comedy and of history plays with a strong local London interest, like Heywood’s Edward IV and Dekker’s Shoemaker’s Holiday. The presence of Dekker and Heywood among the adapters makes perfect sense. They specialized in citizen pride and civic humanism. Shakespeare did not. His three pages echo, instead, with the names ‘Surrey’ and ‘Shrewsbury’, and evocations of ‘the majesty of England’. Although editors describe the mob as citizens, in Shakespeare’s pages they are addressed, instead, as ‘countrymen’, they first refer to their home not as London but as ‘our country’, and the imagery is not urban either, but imported instead from the natural world: herring, butter, beef, roots, parsnips, dung, pumpkins, a river’s ‘bank’, shark, ravenous fishes, a hound, dogs, mountainish. Thomas More may be a Londoner speaking to Londoners, but Shakespeare still warbles his native woodnotes wild.

You can see the same pattern in Timon of Athens, another early Jacobean collaboration, written not long after the additions to Sir Thomas More. Shakespeare creates almost the entire long part of the eloquent tragic male protagonist Lord Timon, and Shakespeare completely dominates the play once Timon leaves the city for the countryside; Shakespeare’s ‘poesy is as a gum, which oozes / From whence ’tis nourished’. By contrast, his younger collaborator, the life-long Londoner Thomas Middleton, dramatizes the satiric, comic, urban ensemble world of servants, creditors and so-called ‘senators’ who are indistinguishable from the oligarchic commercial aldermen who ruled London. Shakespeare wrote most of the play, but MacDonald P. Jackson observes that Middleton ‘created the scenes on which the plot pivots’, and that ‘Middleton’s satirical cameos in Act 3…mingling verse and prose, are the only scenes by a collaborator that Shakespeare could not have written better himself’.³⁵ Theatrically, Middleton’s fast, energetic, urban scenes have always worked better than the magnificently metaphysical poetry of the long, slow, self-indulgent, emotionally static monologues of Timon in the woods. The National Theatre’s award-winning 2012 production of Timon, directed by Nicholas Hytner, demonstrated how powerful and pertinent the play can be in performance. And if, for some critics, the play is a failure, that failure has to be attributed to the dominant playwright: Shakespeare’s excessive focus on the protagonist and Shakespeare’s lack of interest in the rest of the plot, including its conclusion.

The collaborative adaptation of More and the collaborative creation of Timon both recognize two things: first, that London audiences had a growing appetite for the city comedies and city histories being written by Shakespeare’s younger contemporaries, and secondly, that Shakespeare himself was not the man to satisfy that appetite and needed a collaborator to do so. Did Shakespeare personally recognize his limitations? Or did the recognition come from Richard Burbage and the rest of the King’s Men? Who knows? What matters is that we can see, here, the artistic logic of collaboration. An actor is cast in one role, and not another, because every actor does certain things especially well, and other things not quite so well; ideally, the producer or director or actor-manager or someone in the company aligns the skills of a particular actor with the requirements of a particular role. Casting is, according to a widespread theatrical and cinematic axiom, ninety per cent of directing. Casting is also, I would propose, ninety per cent of collaboration. In a collaborative work, each contributor is cast in a particular role; ideally, each is cast in a role that suits his or her particular talents. The achievement of West Side Story depends, for instance, on the music of Leonard Bernstein, the lyrics of Stephen Sondheim, the choreography of Jerome Robbins, the script by Arthur Laurents – and even, a little, on the precursory author, William Shakespeare.

Shakespeare was an actor, but he certainly knew – indeed, everyone knew – that Burbage was a better actor. Burbage was also a painter, which Shakespeare was not. Robert Johnson and Thomas Morley were accomplished composers, which Shakespeare was not. We are willing to admit that Shakespeare collaborated with other people, like Burbage and Johnson, because they were better at something than he was. Why then are we so resistant to accepting that another writer might have been better at some aspect of writing than Shakespeare was? We accept that Shakespeare incorporated traditional song lyrics, written by other people, into his plays. Why then has it taken Shakespeare’s editors three centuries to recognize or accept the evidence of his collaborations?

Part of the explanation must be that human beings are hard-wired to seek the simplest possible cause of an effect, and therefore we typically focus on a single agent, even when we know there is more than one. Everybody talks about Verdi’s operas, or Sondheim’s musicals, even though neither Verdi nor Sondheim ever worked alone, or wrote what theatre credits call ‘the book’ of a musical play. Thus, modern productions and editions advertise ‘William Shakespeare’s’ Timon of Athens, even when the inside of the book, or the programme, acknowledges that Middleton wrote parts of the play. Likewise, as David Nicol has pointed out, critics routinely praise the collaborative plays of Middleton and Rowley as though they had been written entirely by Middleton.³⁶ Jeffrey Knapp recognizes that Pericles and The Two Noble Kinsmen are collaborative plays, and indeed he interprets both as metatheatrical meditations on collaboration – but only in terms of Shakespeare’s thoughts about collaborating.³⁷

Another part of the explanation must be another, related illusion: what Thomas Carlyle called ‘hero-worship’, what Daniel Kahneman and other cognitive psychologists call ‘the halo effect’.³⁸ This can be seen clearly in the first edition of Shakespeare’s works to pay serious attention to the problems of attribution and collaboration. In 1725 Alexander Pope correctly denied that Shakespeare had written ‘Locrine, Sir John Oldcastle, Yorkshire Tragedy, Lord Cromwell, The Puritan’ or The London Prodigal. But in the very next sentence he conjectured that in some other plays ‘(particularly Love’s Labour Lost, The Winter’s Tale and Titus Andronicus)’, Shakespeare wrote ‘only some characters, single scenes, or perhaps a few particular passages’ (xx). In the edition itself Pope marks as un-Shakespearian particular scenes in other comedies (Two Gentlemen of Verona, The Comedy of Errors, Much Ado about Nothing, The Taming of the Shrew) and the comic Porter’s scene in Macbeth. Pope, a great satiric poet with a brutally sharp sense of humour, did not think this comic material was funny, and accordingly could not believe that Shakespeare wrote it.

As it happens, modern scholarship has found plentiful evidence of collaboration in Shakespeare’s histories and tragedies, but no collaborator has been identified in any of his comedies – and although Middleton adapted Macbeth, Shakespeare created the Porter. In this respect, Samuel Johnson was a more accurate judge of Shakespeare’s achievement than Pope. Shakespeare’s ‘natural disposition’, Johnson famously intoned, ‘led him to comedy. In tragedy he often writes with great appearance of toil and study, what is written at last with little felicity; but in his comick scenes, he seems to produce without labour, what no labour can improve. In tragedy he is always struggling after some occasion to be comick, but in comedy he seems to repose, or to luxuriate, in a mode of thinking congenial to his nature.’³⁹ On the basis of what we now know about Shakespeare’s collaborations, we can modify Johnson’s summary slightly by concluding that (like John Lyly) Shakespeare created his own mode of Elizabethan romantic comedy, which was particularly and happily self-sufficient. His histories, tragedies and tragicomedies, by contrast, contain more writing by other people, presumably because Shakespeare or someone else felt that he would profit from the partnership.

The halo-effect led Pope to deny Shakespeare’s authorship of passages that Pope himself found aesthetically unsatisfying; in Pope’s case, hero-worship preceded and directed attribution. That still happens; it’s found, for instance, in the Folger Shakespeare’s systematic denial of collaboration or adaptation; it’s evident in the particular refusal to acknowledge a second author in the most recent Arden edition of Titus Andronicus, or the most recent Cambridge editions of Pericles, Timon of Athens and Measure for Measure; it’s visible when Lukas Erne attributes to Shakespeare all of the ‘first tetralogy’ and calls it ‘the most ambitious project the professional stage had yet seen’.⁴⁰ But hero-worship can also follow, and respond to, attribution. The growing empiricism of attribution scholarship as a technical discipline, beginning in the nineteenth century, inevitably impacted the preferences of Shakespeare fandom. Shakespearians now routinely deny the aesthetic achievement of passages in Shakespeare’s works that Shakespeare turns out not to have written. Why would Shakespeare choose to collaborate with writers who were as incompetent as many of Shakespeare’s editors and critics presume them to be?

Perhaps because Shakespeare did not consider his collaborators incompetent. At the beginning of his career, Shakespeare was learning from more experienced craftsmen, in the kind of apprenticeship relationship normal throughout medieval and early modern Europe. Coleridge recognized that the first lines of 1 Henry VI do not sound like Shakespeare’s verse – but Coleridge did not dismiss them as rhythmically or rhetorically monotonous, as Vickers does. Nashe’s sequential short sentences and end-stopped lines in that scene can also be found elsewhere in his work, in passages that have been widely anthologized and admired:

The choric repetition of ‘Lord, have mercy on us’ in this famous lyric is echoed in 1 Henry VI, when the dying Salisbury’s ‘O Lord, have mercy on us, wretched sinners!’ is immediately followed by the dying Gargrave’s ‘O Lord, have mercy on me, woeful man!’ (1.4.70–1). G. R. Hibbard, describing the blank verse in Summer’s Last Will and Testament, the source of this lyric, acknowledges that ‘The lines are largely end-stopped’, but he also recognizes that much of it ‘surprises by its lyrical grace and easy flow’. He praises the ‘the argumentativeness’ and ‘sheer virtuosity’ of Nashe’s dialogue, ‘the way in which the blank verse is handled to express and contain a veritable torrent of abuse and misrepresentation’, ‘the insistent use of hammering alliteration to convey scorn and contempt’ and ‘the way in which all the detail is integrated into a long verse paragraph, building up to the climax’. All these observations are just as relevant to Nashe’s verse in Act 1 of 1 Henry VI, as is Hibbard’s recognition of ‘two conflicting impulses’ in Nashe’s work, ‘an affection for the past and an impulse to laugh at it. It is both naive and sophisticated.’⁴¹

In the 2013 production at Shakespeare’s Globe, the first scene, by Nashe, was the most powerful in the play, much more memorable than the Temple Garden scene (2.4), or the scenes leading up to Talbot’s death (4.2–4.5), which are all attributed to Shakespeare.⁴² We may blame this disparity on the weakness of a particular production; earlier stagings, directed by Terry Hands (1977) and Michael Boyd (2008), better demonstrated the theatrical potential of those later Shakespeare scenes. But in those productions, too, the first scene was more brilliantly dramatic. And Pope, who preferred poetry to theatre, and who was not biased by our knowledge of who wrote what, degraded to the bottom of the page passages in both 2.4 and 4.2.

What Nashe brought to the history play, beyond a different music and a mind better equipped to start texts than to close them, was the vigorous xenophobia that we can also see in The Unfortunate Traveller.⁴³ We may not praise this, but English audiences still respond to the Francophobia (often with laughter), and it was an essential ingredient to the growth of history as an Elizabethan dramatic genre. We can easily distinguish Nashe’s French-baiting from the humanist defence of (Catholic) ‘strangers’ in the speeches Shakespeare wrote for More. Shakespeare’s scenes in 1 Henry VI concentrate instead on flowers and hunting, on destructive factional divisions among the English themselves, on the tragedy of the charismatic male protagonist Talbot, on the relationship between fathers and sons, and on men generally. Indeed, no women appear in Shakespeare’s scenes of 1 Henry VI. This, too, is typical. Women play a much larger and more important role in the three (collaborative) Henry VI plays and the (collaborative) All Is True than in Shakespeare’s single-authored ‘second tetralogy’.⁴⁴ Nashe, the author of ‘Choice of Valentines’, introduces the play’s first sexualized and ambiguous woman, Joan. Nashe is also more interested than Shakespeare in modern mechanized warfare: 1.4 is the only scene in Shakespeare where characters are killed, onstage, by artillery fire, which produces the grotesque mangling of bodies (‘One of thy eyes and thy cheeks’ side struck off!’) that we can also find in Nashe’s prose, fascinated as it is with the dissected human body. Nashe provides, too, the populist anti-Catholicism of the play’s third scene, explicitly set in a London of rioting apprentices, which twice mocks Winchester’s ‘cardinal’s hat’ (1.3.36, 49), and rebels from ‘Pope and dignities of Church’ (50). Winchester, historically, should have been a bishop in that scene, which notoriously contradicts 5.1.28–33, where Winchester’s elevation to cardinal seems to have just occurred. But Nashe, who had written anti-Marprelate pamphlets in defence of English episcopacy, would have known that attacking Winchester as a ‘bishop’ would have come dangerously close to Marprelate’s puritan position; attacking a cardinal, by contrast, was perfectly safe, indeed an endorsed public pleasure. Anti-Catholicism and Xenophobia go hand-in-hand and, together with their good friend Misogyny, they made 1 Henry VI a huge popular success. The play is, if anything, too unified by these interrelated otherings; its failure, from our perspective, is primarily political, not aesthetic.

‘It takes all sorts of playwrights’, Harold Pinter acknowledged, ‘to make a world.’⁴⁵ The Elizabethan history play also created a world, and often did so by combining different authorial voices in what Nina Levine calls ‘a community constituted by difference’ and by collaborative ‘reciprocities of plurality’.⁴⁶ Michael Morpurgo, describing the National Theatre’s acclaimed and beloved adaptation of his novel War Horse, was particularly struck by the company’s ability to ‘yoke people together with a common purpose’, and thereby ‘create a sense of community’. War Horse is, of course, an English history play, and its success depends less on puppets than on what Morpurgo called ‘a togetherness about the whole thing’.⁴⁷ A history play answers the question ‘Who were we?’ and its causal corollary ‘Why are we?’, defining ‘we’ not as ‘human beings in general’, but as ‘a particular community to which the spectators belong’.⁴⁸ The genre explores collective identity, not individuality. It should not surprise us that, of the twelve history plays in current definitions of the Shakespeare canon, half are collaborative – a larger proportion than for any other genre.

Shakespeare’s last history, All is True (a.k.a. Henry VIII), decisively identified as a collaboration more than a century and a half ago, has suffered from the negative side of the halo effect longer than any other play. Pope never doubted Shakespeare’s responsibility for the play, and he did not degrade a single passage to the sewer at the bottom of the page that he reserved for interpolations. Indeed, he singled out, typographically, four passages in the play as ‘beauties’ deserving of particular commendation and attention: three in speeches by Wolsey after his fall (3.2), and some lines in Queen Katherine’s final scene (4.2). All four passages, it turns out, were written by Fletcher. The notoriously anti-theatrical Pope thought that the best poetry in the play was written by Fletcher. By contrast, modern critics consider Fletcher a poor poet, and compare these passages unfavourably with the complex versification, syntax and imagery of Shakespeare’s scenes in the play.⁴⁹ But theatrical performances of the play, like the stunning 2013 production at the Chicago Shakespeare Theatre, directed by Barbara Gaines, always keep these speeches by Fletcher, and often heavily cut Shakespeare’s incomprehensible poetry.⁵⁰ In All is True, as in The Two Noble Kinsmen, ‘Shakespeare’s poetry…invests the story with a sense of real gravitas’, particularly in the portrayal of royal families and gods. But in performance Fletcher’s scenes in both plays are more emotionally and theatrically powerful, and – as in Timon of Athens – Shakespeare’s collaborator provides a stronger narrative drive.⁵¹ In each case, the collaborator provides something valuable, and valuably different from Shakespeare.⁵²

Fletcher is also accused of ‘inconsistencies in characterization’, and critics are particularly contemptuous of his ‘trademark sudden reversals’ when a character abruptly switches ‘from one position, expressed in extravagant terms, to its opposite’.⁵³ But the sudden reversals of Fletcher’s characters are compellingly true to my own experience of other people. Perhaps I’m particularly dense, but haven't you ever had the experience of being completely stunned when someone you think you know does something or says something that contradicts, radically, who you think they are? Eventually I may, or may not, re-interpret their personality in a way that reconciles position A with position B, and therefore ‘unifies’ their character; but that retrospective intellectual hypothesis does not erase, or replace, the vivid, disturbing experience of inconsistency, which is also found often in the tragedies of Euripides.

The modern objection to Fletcher’s characters belongs to a larger critique of his work which originates with Coleridge: ‘the plays of Beaumont and Fletcher’, declared the Romantic sage, ‘are mere aggregations without unity’.⁵⁴ This Romantic celebration of the ‘organic’ unity of Shakespeare’s plays, opposed to the merely ‘mechanic’ unity of Beaumont and Fletcher, has become a more general critique of all collaborative plays, including Shakespeare’s. The pioneering attribution scholar Cyrus Hoy declared: ‘The crucial issue for the aesthetic appraisal of [a collaborative play] is how satisfactorily the multiple dramatic visions have fused into a single coherent one.’ The postmodernist critic Jeffrey Masten, who disdains Hoy, insists that collaborators were dedicated to ‘erasing the perception of difference’ between them. The attribution scholar Brian Vickers, who disdains Masten, systematically catalogues and laments the inconsistencies of plotting and characterization in each of Shakespeare’s collaborative plays.⁵⁵ Despite their other differences, these three influential experts on collaborative drama all agree that plays should be unified, all agree that collaboration is an apparent obstacle to unity, and all agree that collaborative plays can and should be judged, aesthetically, by the standard of unification.

In Shakespeare, Co-Author Vickers provided a valuable, monumental, polemical synthesis of the collaborative achievement of hundreds of nineteenth and twentieth century scholars, who established the presence of a collaborator in five of Shakespeare’s plays. Unfortunately, that synthesis culminates in a chapter that denigrates the aesthetic achievement of those five plays, and indeed of all collaborative plays, because all such plays fail to deliver the ‘unity’ demanded by Horace, and by all the Elizabethan grammar schoolmasters who shoved Horace down their students’ throats. Why would Shakespeare collaborate, when collaboration inevitably damaged the aesthetic unity of a play?

Perhaps because Shakespeare did not value certain kinds of unity. Horace was not a playwright. Ben Jonson translated Horace’s Ars Poetica, compared himself to Horace, and was compared by his contemporaries to Horace, sincerely or satirically. But Shakespeare was compared to Ovid, not Horace. Shakespeare’s favourite poet was Ovid, and his favourite work was Ovid’s Metamorphoses.⁵⁶ No one ever read the Metamorphoses for unity. Ovid was not only far more popular than Horace in Western Europe throughout the middle ages and the Renaissance; he was also more characteristic of classical culture generally, as anyone familiar with The Odyssey, Pindar or Herodotus should recognize. The classical scholar Malcolm Heath observes that ‘poikilia’, diversity, is invoked more frequently in ancient Greek criticism than unity, ‘and is an important principle of artistic construction in Greek tragedy’.⁵⁷ The Renaissance rediscovery of Aristotle’s Poetics quickly led Italian critics to defend Ariosto’s Orlando Furioso by challenging the Aristotelian insistence on unity. The foundational defence of Ariosto was written by a writer that Shakespeare certainly read, Giovambattista Giraldi Cinzio (who wrote the source of Othello). Whereas Homer in the Iliad and Vergil in the Aeneid had set out to describe ‘a single action of one knight’ (‘una sola attione de un cavaliero’), Ariosto – like the authors of French, Provencal and Spanish romances – wished to treat ‘many [actions] of many [men]’ (‘molte de molti’).⁵⁸ Neither Aristotle nor Horace is of any use in understanding modern vernacular poets, who with their multiple plots ‘relieve the satiety caused by always reading one same thing’ (‘levare la satietà al lettore di sempre leggere una medesima cosa’).⁵⁹ Variety, not unity, was the fundamental criterion.⁶⁰

Whether or not Shakespeare read Orlando Furioso and its Italian fans, Sidney and Spenser certainly did. Whether or not Shakespeare had ever seen a commedia dell'arte performance, he began his career in the century when licensed companies of professional players, first in Italy and then in England, subordinated the plots of humanist writers to the unpredictable onstage interactions of an ensemble of character-actors and clowns. After the work of Madeleine Doran on ‘multiplicity and sequential action’, after David Bevington’s work on the sixteenth-century evolution of dramatic form to achieve ‘panoramic inclusiveness’ by means of doubling, after Richard Levin’s work on the evolution of multiple plots from Terence to Middleton, after Scott McMillin and Sally-Beth MacLean on the Queen’s Men, the dominant acting company of the 1580s, with a repertory defined not by recognizable classical genres but by what the actor-playwright Robert Wilson called a ‘medley’ – after six decades of collaborative scholarship, it should no longer be necessary to insist that the aesthetic of Shakespeare and his collaborators valued Ovidian variety more than Horatian unity.⁶¹

None of these critics was writing specifically about collaborative authorship. But the cultural logic that values romance narratives, stylistic variation, mixed genres, doubling actors and double plots also applies to the variety created by mixed authorial voices. The first Elizabethan collaborative plays were not written by hurried hacks, whipped into submission by the vile capitalist imperatives of Philip Henslowe. From 1562 to 1588, Gorboduc, Jocasta, Gismond of Salerne, Estrild and The Misfortunes of Arthur were all collaboratively written by humanist gentlemen, apparently for no financial reward.⁶² Before he began collaborating in the London theatres, Thomas Nashe wrote one part of a student ‘shew’ at the University of Cambridge. Dr Johnson speaks of ‘confederate authors’ and, although that idiom disappeared after the American Civil War gave ‘confederate’ particularly negative connotations, Johnson’s adjective does capture the sense that a collaborative work of art does not seek ‘unity’, but instead presupposes a ‘federal’ structure that allows for individual difference.⁶³ Early modern playwrights made temporary and shifting alliances, based on the needs of a particular narrative. The division of plays into acts, scenes, or sections of scenes, written by separate authors, gives each author imaginative autonomy within an agreed framework.

But how do we distinguish between a desirable multiplicity and an undesirable inconsistency? In highlighting the inconsistencies in Shakespeare’s collaborative plays, Vickers violates one of the cardinal rules of attribution scholarship. Before we can attribute responsibility for a disputed work to a particular author, we must first examine the uncontested canon of candidate authors, in order to establish the constants that characterize all their work. In this case, we are trying to identify which characteristics, if any, distinguish collaborative plays from single-author plays. Vickers says, rightly, ‘Scholars have long noticed many inconsistencies in Shakespeare’s co-authored plays.’⁶⁴ But in fact scholars have noticed many inconsistencies in all Shakespeare’s plays. Between 1982 and 1993 Kristian Smidt wrote four books on ‘unconformities’ in Shakespeare’s histories, tragedies and comedies, collecting examples noticed by earlier critics.⁶⁵ In Henry V Pistol begins married to Nell, but ends married to Doll; Exeter is specifically ordered by the King to ‘remain’ behind in Harfleur (3.3.135), but then reappears on the battlefield at Agincourt (4.3, 4.6, 4.7, 4.8), without explanation. In Two Gentlemen Sylvia’s father is both a Duke and an Emperor, living simultaneously in both Milan and Verona; Othello is only the most famous example of the incompatible time schemes found throughout Shakespeare’s plays. A.C. Bradley took an interest in just this sort of contradiction in the notorious notes to Shakespearean Tragedy, and Tolstoy’s condemnation of Lear shows what happens when a realist critic encounters Shakespeare’s basic violations of narrative logic. Wherever we look in English Renaissance drama, we will find rampant inconsistency.

Attribution scholars are particularly likely to notice, and to be disturbed by, the kinds of inconsistency industriously collected and rhetorically displayed by Vickers. After all, the technical discipline of attribution scholarship consists of the systematic reading of recurring material signs; computers are often preferred because a computer reads material signs more consistently and systematically than humans do. The attribution scholar or the scholar’s computer is particularly good at noticing small details that have hitherto been overlooked; in fact, details of which the author himself might have been unconscious provide particularly valuable evidence. Moreover, attribution scholarship depends on pattern recognition, the ability to connect details in different parts of a text. All these admirable skills and procedures enable attribution scholars to identify, empirically and objectively, who wrote a play or a significant part of one. But these same skills and procedures also make attribution scholars hypersensitive – indeed, obsessive compulsive – about inconsistencies of plot and characterization. Those inconsistencies are not likely to be noticed in the theatre; nor are they likely to be noticed by what we might call the common reader, that endangered species of creature who reads Shakespeare for the plot or the poetry, outside the professionalized ‘domain dependence’ of specialist attribution scholarship.

Spectators don't notice the inconsistencies that bother scholars. In the theatre, or in other unscripted environments, we ‘only perceive a contradiction if we juxtapose a present moment with an incompatible antecedent moment; how many spectators perceive a contradiction therefore depends on the probability of this present moment being juxtaposed with that one particular incompatible antecedent moment, and no other’.⁶⁶ Our ability to make such juxtapositions, in real time, is a function of immediate memory span. ‘A great deal of experience does not survive the instant of its passing and is irretrievably forgotten the moment it is over’, according to cognitive psychologists; ‘in the next instant, the circumstances of which we need to take account are different.’⁶⁷ A book, by contrast, is an artificial memory device, which enables us to juxtapose any moment of a play with any other moment, and to notice the kinds of self-contradictions that have fuelled deconstructionist philosophy and criticism.

Long before laboratory experiments in the psychology of perception demonstrated our hard-wired liability to overlook such contradictions, professional theatre-makers knew how easy it was to trick an audience. A professional magician knows that we won't see the sleight of hand by which he exchanges one card for another. The protagonist of Titus Andronicus cuts off his own hand (in a scene written by Shakespeare, not Peele), an action that must be convincingly horrific, and that contributes to the pathos of the character’s ‘lamentable action of one arm’ for the rest of the play.⁶⁸ But the actor playing Titus does not really cut off his hand at every performance, any more than the actor playing Lavinia has her hands and tongue amputated. The 2013 Royal Shakespeare Company production of Titus included, among its Creative Team, a professional ‘illusionist’, Richard Pinner, who was ‘Stage and Close Up Magician of the Year’ in 2003 and has worked in more than thirty countries. The success of magic tricks depends not only on the practised skills of the magician but on the human susceptibility to particular kinds of cognitive illusion.⁶⁹ We routinely tell our students and ourselves that a play is not ‘real’, that it is ‘artificial’, and we congratulate ourselves on our metatheatrical sophistication. We celebrate Shakespeare’s ‘magic of bounty’ (Timon, 1.1.6) and recognize the playwright’s self-description when Rosalind describes ‘a magician, most perfect in his art and yet not damnable’ (As You Like It, 5.2.58–9). But we are less happy about acknowledging that as a matter of routine actors and playwrights successfully con us, trick us, pick the pockets of our minds, dazzle and deceive us with what Middleton, in a speech he added to Macbeth, calls ‘magic sleights’ – tricks that ‘raise such artificial sprites’ as the supernatural figures in Shakespeare’s plays, which wow us with ‘the strength of their illusions’ (3.5.26–8). Unfortunately, the magic tricks that create the illusion of narrative logic, psychological depth and aesthetic unity no longer work if we record them with a fixed camera, and then play them back, over and over again, at very slow speed – which is what happens when scholars read, over and over again, the printed text of a play.

The inconsistencies created by collaboration do not matter in the theatre; the variety created by the particular styles and talents of different collaborators do matter. That’s perhaps why collaboration is not common in novels, but common enough in both theatre and film. The screenplay for Birth of a Nation, the first full-length motion picture, was co-written by D. W. Griffith and Frank E. Woods. Casablanca, consistently ranked among the top films of all time, had three credited screenwriters, and a fourth uncredited, and was based on a ‘plot’ co-written by two playwrights. Some Like It Hot, widely regarded as one of the greatest film comedies in English, was co-written by Billy Wilder and I. A. L. Diamond; Wilder, in fact, ‘quite clearly preferred to collaborate’, and all his great films were co-written.⁷⁰ The film of David Mamet’s Glengarry Glen Ross, a ‘constructive collaboration’ based on a ‘productive intersection’ of talents, was a critical and commercial success; by contrast, the ‘one-sided’ film of Oleanna, completely dominated by the singular authorial intelligence of Mamet, was a critical and popular failure.⁷¹ Screenplays that are collectively written, like screenplays with a single author, can ‘produce the most engaging and the most turgid films’.⁷² The same is true of early modern plays. Sometimes it works, sometimes it doesn't. But the fundamental motive is always the same: the recognition that, sometimes, ‘collaboration has an explosive upside, what is mathematically called a superadditive function, i.e. one plus one equals more than two, and one plus one plus one equals much, much more than three’.⁷³