Gary Taylor and John Lavagnino (eds), Thomas Middleton, Vol. 2: Thomas Middleton and Early Modern Textual Culture: A Companion to the Collected Works

Contents
Find Location in text

Main Text

pg 80EARLY MODERN AUTHORSHIP: CANONS AND CHRONOLOGIESMacDonald P. Jackson

I

'Those who study plays want to know who wrote them.' The statement with which in 1966 S. Schoenbaum began his survey of investigations of the authorship of Elizabethan, Jacobean, and Caroline drama (xv) remains true long after literary theorists first announced the 'death of the author'. Indeed, reports of this demise, as of Mark Twain's, were 'an exaggeration'.1 There is a sense, obviously, in which a play is the expression of a Zeitgeist, a recipient of the current ideologies, a fabrication of the available language, a product of intertextual transactions. Postmodernist theory has increased our awareness of those elements that, entering a text beyond the author's conscious control, become symptomatic of its time and place. But as well as a 'tradition' that encompasses everything from theatre practice to the prevailing cosmology there is the 'individual talent'. Middleton's plays result from the workings of his own creative imagination within a social, literary, and philosophical milieu. They bear the stamp of his distinctive genius.

Writing about The Revenger's Tragedy in the Cambridge Companion to English Renaissance Drama, Robert N. Watson claims that both traditional literary historians and 'their theoretical antagonists' may legitimately agree that the play 'cannot be safely attributed to any single author'. Traditionalists, he argues, find insufficient evidence to ascribe it either to Tourneur or to Middleton (or to Marston). 'Theoreticians increasingly advocate a new model in which the attribution of any play to any person is misleading, either because plays were not proprietary to their authors in the Renaissance, or because the entire concept of authorship is a sentimental modern fiction disguising the fact that authors (like other individuals) are not really autonomous agents, but products of their cultures' (329–30). The Revenger's Tragedy, he points out, is indebted to the literary and dramatic traditions, and expresses the obsessions, of its age.

Watson's remarks themselves betray their specific cultural origins. They could not have been written in the palmy days of attribution studies, when scholars such as H. Dugdale Sykes and E. H. C. Oliphant were busily finding parental homes for the waifs and strays of Elizabethan and Jacobean drama. But Watson's lively chapter, though of its time, is also his. He deserves the credit for it. It belongs to his CV. Qualms about his not being an 'autonomous agent' would be misplaced. Had it been written by somebody else, it would be different, in style and substance. Renaissance writers were real people too. Middleton earned his living by his writing. It is true that when he handed over his script of The Chester Tragedy, let us say, to theatre entrepreneur Philip Henslowe, he retained no authorial copyright, but on 21 October 1602 he received four pounds in exchange and on 9 November a further two pounds (Foakes and Rickert, 205–6). Middleton, whatever he owed to his culture, was the person paid. Henslowe, confronted with the insistent demands of his ever-needy team of playwrights, may often have wished that 'the concept of authorship' could be dismissed as 'a sentimental fiction', but the writers had mouths to feed and came to him with saleable property. And however much The Revenger's Tragedy is a product of its period, it either was or was not conceived in Middleton's brain, first penned by him, and sold by him to some theatrical company. Somebody mixed the culturally derived ingredients into the play's new and individual blend.

But does it matter who was, in this sense, its author? I think that it does—that, while the facts of literary history are not always fully ascertainable, we have an obligation to get them as right as we can; that the reasons for assigning The Revenger's Tragedy to Middleton are overwhelmingly strong; that the traditional ascription to Tourneur has led critics to misrepresent the play's nature; that it gains in interest from being placed within the Middleton canon; and that Middleton's own insouciance about his reputation with posterity ought not to prevent us, as historians of drama and literature, from examining the evolution of his art, evaluating his achievement, and recognizing his greatness. Commenting on Shakespeare, T. S. Eliot held 'that the full meaning of any one of his plays is not in itself alone, but in that play in the order in which it was written, in its relation to all of Shakespeare's other plays, earlier and later' (136). We may not want to go so far as that—with regard either to Middleton or to Shakespeare—but the sense we receive of 'one significant, consistent, and developing personality' (150) behind the diverse works associated with a major, or even minor, author's name remains an important part of our experience of literature and drama. Which is why there is still a demand for writers' 'collected works'. And, as Schoenbaum went on to remark apropos of wanting pg 81to know who wrote things, 'The editor needs to decide which works to include in his edition' (xv).

II

During the last few decades the body of literature in English that is deemed worthy of academic study has undergone considerable modification. The security of Shakespeare's place within this changing canon is in part due to there being such a clearly defined 'Shakespeare canon'—a substantial and coherent body of work that we know to be his. Had not his fellow actor-shareholders in the King's Company, John Heminges and Henry Condell, gathered his plays for publication in the First Folio of 1623, seven years after his death, the situation would have been very different. Eighteen of his plays— about half the total—would be unknown to us, including such masterpieces as Julius Caesar, As You Like It, Twelfth Night, Measure for Measure, Macbeth, Anthony and Cleopatra, Coriolanus, The Winter's Tale, and The Tempest. Henry V and The Merry Wives of Windsor would exist only in corrupt, abbreviated texts, barely recognizable as Shakespearean. Moreover, since the 1600 'bad quarto' of Henry V fails to attribute it to Shakespeare, the only item of 'external evidence' for his authorship would be the Epilogue's promise of a sequel to 2 Henry IV. The manner in which the anonymous The First Part of the Contention (1594) and Richard Duke of York (1595)— now commonly held to be memorially reconstructed texts of the Folio's 2 and 3 Henry VI—are connected with Shakespeare would be altogether mysterious. The Folio is crucial to our sense of Shakespeare's achievement. Modern editors accept its definition of his dramatic canon with only minor additions (Taylor 1987). Also, it is widely acknowledged that at least one or two of the plays in the Folio (and in subsequent editions of The Complete Works) are collaborative, containing material by dramatists other than Shakespeare. But these adjustments to the Folio's canon have little effect on our sense of 'Shakespeare'.

Heminges and Condell collected Shakespeare's plays in order 'to keep the memory of so worthy a friend and fellow alive', and succeeded beyond all expectation. They may never have undertaken the venture had it not been for the precedent set by Ben Jonson, who fashioned his own memorial with his famous Works of 1616. The Jonson Folio was an altogether more self-consciously ambitious volume than the Shakespeare, and there were jibes at the author's pretensions. With meticulous attention to the details of typography and layout and to the selection and ordering of texts, Jonson presented his plays, poems, and masques, to a reading rather than theatre-going public, as the authorized opera of an instant classic. In his treatment of speech prefixes and stage directions he imitated classical Latin literature, associating his texts with venerable conventions while perpetuating them through the technology of print. Omitting all his apprentice work and collaborations (including several for Henslowe in the period 1597–1602), suppressing from Sejanus all trace of his original co-author, and thoroughly revising such plays as Every Man in His Humour, he projected the image of himself as writer by which he wished to be remembered—independent and in control. He supervised the volume's progress through the printing-house. Jonson constructed a cultural icon. His Works bore witness to his individual genius. The volume was designed as the physical monument of a solid authorial persona (Brady and Herendeen; Butler; Dutton 61–9).

The third great Folio collection of Renaissance drama provides a striking contrast. This, published in 1647 by Humphrey Moseley, purported to contain the comedies and tragedies of Francis Beaumont and John Fletcher. There were thirty-five dramatic pieces, including one masque. A Second Folio of 1679 added another eighteen plays. Moseley's Folio, unlike the Shakespeare and Jonson folios, was a purely commercial venture; moreover, the plays were mustered under the aegis not of a single author but of 'Beaumont and Fletcher'. In fact—if for the sake of argument we accept the findings of the latest and most careful investigator—Fletcher wrote on his own fifteen of the Second Folio's fifty-two plays and he had a share in all but three of the remainder, but only nine result from his straightforward dual partnership with Beaumont, who wrote or participated in no more than five others (Hoy). The masque was also Beaumont's. This Folio is essentially a witness to the copious industry of Fletcher, alone and as co-author, not just with Beaumont but, slightly more often, with Massinger, and also (sometimes unwittingly, as provider of a text later revised) with Field, Shirley, Rowley, Webster, Shakespeare, and Ford. Several plays are by three or four hands.

In the second half of the seventeenth century, after the Restoration, Shakespeare, Jonson, and Fletcher (to whom Beaumont's name was sometimes conjoined) were the English Renaissance dramatists to be admired and discussed. For Dryden, in his 'Of Dramatic Poesy: An Essay' (1668) and elsewhere, these are 'the big three' (or 'four') among his predecessors, the most worthy of his critical attention. Dryden weighs the comparative strengths and weaknesses of Shakespeare and Fletcher or of Shakespeare and Jonson. About Middleton he has nothing to say. Middleton's corpus of plays, poems, entertainments, and pamphlets was dispersed over sundry quartos, octavos, and manuscripts of widely varying dates. Uncollected, Middleton disappeared from view, 'untalked of and unseen'.

It was not till 1840 that Alexander Dyce conferred some kind of corporate identity on Middleton's writings with the publication of The Works. A. H. Bullen's reprint of 1885–6 made Dyce's canon more readily available, but added almost nothing in the way of new scholarship; its most significant alteration was to omit both parts of The Honest Whore, whereas Dyce had included both. Bullen omitted them, simply on the grounds that both had recently been included in his edition of Dekker. Bullen's edition—the form in which most readers encountered Middleton's œuvre for more than a century—was a defective pg 82reprint of Dyce. But even the canon defined by Dyce was seriously misleading, excluding some of Middleton's finest plays, besides other writings of considerable interest, and including items for which Middleton was not responsible. Dyce omitted some works—The Two Gates of Salvation, for instance, and Honourable Entertainments—simply because he did not know of their existence. Otherwise, he basically collected anything that had ever been attributed to Middleton in the early modern period. He paid no attention to internal evidence. He did not recognize, in practice, the prevalence in Middleton's time of anonymous publication. The principles adopted by Dyce, and followed by Bullen, would, if applied to Shakespeare, admit to the canon the various apocrypha printed in the Third Folio of 1664, while excluding 'Hand D's' contribution to Sir Thomas More. The inadequacies of Dyce's intellectual procedures, and of Bullen's careless reprint, distorted the Middleton canon. Readers, historians, and theatre practitioners interested in Middleton need a better definition of his corpus.

III

How, then, are decisions about what rightly belongs in a Collected Works of Thomas Middleton to be made? The evidence for particular inclusions and exclusions in the Oxford edition is described in detail, elsewhere in this volume, in 'Works Included in this Edition: Canon and Chronology' (p. 335). My aim here is not to repeat that information, but to provide an overview of the issues raised, and of the different categories of evidence deployed, in any attempt to establish the canon of an early modern author.

Renaissance autograph manuscripts are rare; printed texts were often published anonymously; when title-pages do make ascriptions they cannot always be trusted; other sources of information—such as the Stationers' Register, Henslowe's Diary, official documents, booksellers' catalogues, and the recorded statements of a playwright's contemporaries—may have nothing pertinent to offer on a given case and are of variable worth; the collaborative preparation of new playscripts and the refurbishing of old ones were features of the Jacobean entertainment industry that complicate any attempt to identify authorial provenance.

Works that have been associated with Middleton differ enormously in the quality of the external evidence for regarding them as his. At one extreme, the authorship of A Game at Chess is better attested than that of any other English play of the period. It is preserved in a manuscript wholly in Middleton's own handwriting (including his 'by T.M' on the title-page), in two manuscripts with autograph title-pages displaying the formula 'by Tho: Middleton', and in a manuscript containing dedicatory verses written out and signed 'T.M.' by Middleton himself, as well as in other early manuscripts and quartos. A Game at Chess was licensed by Sir Henry Herbert on 12 June 1624 as 'written by Middleton'. Moreover, Middleton's responsibility for this bold anti-Catholic and anti-Spanish allegory apparently earned him a term in jail: on 18 August 1624 a warrant directed him to appear before the Privy Council, and Middleton, having failed to do so, is mentioned as the 'poet' of the offending theatrical piece in several further State Papers (Howard-Hill, 1–33, 192–213). At the other extreme, for some of the works admitted to the new Oxford Collected Works there is no external evidence whatsoever that links them to Middleton. When such evidence does exist, it can seldom be taken on trust, but requires careful assessment.

In such circumstances, resort to the 'internal evidence' of stylistic and sub-stylistic features of the texts themselves is absolutely necessary. But of course this kind of investigation would be impossible did not external evidence establish a core of undoubtedly, or almost undoubtedly, Middletonian writings to serve as basis for the detection of the writer's distinguishing peculiarities. Fortunately, there has been general agreement that at least a dozen plays are of Middleton's unaided authorship, and these span almost the full chronological range of his playwriting career. Admittedly, some of the attributions, on printed title-pages or in catalogues, are quite late, made after Middleton's death in 1627, but scholars have accepted their accuracy because of the clear internal connections among all twelve plays.

I have already mentioned A Game at Chess. The first quarto of Your Five Gallants (undated, but perhaps 1608) and the second edition of A Trick to Catch the Old One (1616) each carry title-page ascriptions to 'T. Middleton'; in the latter case the name expands on the 'T.M.' of the second issue of the first quarto (1608), and these same initials also appear on the title-page of A Mad World, My Masters (1608). The precise formula 'Composed by T.M.' (to modernize the spelling) links A Trick to Catch the Old One and A Mad World, My Masters to the autograph title-page of the Rosenbach Manuscript of A Game at Chess, where it recurs. Ralph Crane's scribal transcript of The Witch names 'Tho. Middleton' on the title-page and below the author's dedicatory epistle. After Middleton's death, the quarto of A Chaste Maid in Cheapside (1630) assigns the comedy to 'Thomas Midelton Gent.' Then on 9 September 1653 Humphrey Moseley entered in the Stationers' Register More Dissemblers Besides Women, Women, Beware Women, and No Wit/Help like a Woman's as by 'Mr. Tho. Midleton', before eventually publishing all three under the same authorial name in octavos of 1657. Finally, Hengist, King of Kent; or The Mayor of Quinborough was entered on the Stationers' Register for 13 February 1661 as 'By Tho: Middleton', and published as his later that year.

Claims made in booksellers' catalogues are intrinsically far less dependable (Greg, 1938–45). Michaelmas Term was first attributed to Middleton by Edward Archer in a playlist appended to Old Law in 1656; despite Archer's notorious unreliability, that attribution has passed unquestioned, because the anonymous 1607 edition associates the play with the Children of Paul's, for whom several of Middleton's early city comedies were written, and because it is so pg 83like them in every detail of its matter and manner. Francis Kirkman's ascription to Middleton of The Phoenix in a playlist of 1661 has been accepted for the same reason: this satirical comedy had also been published in 1607 as acted by the Children of Paul's, and it too hits at familiar Middleton targets.

Although the quality of the external evidence for Middleton's authorship even of these twelve plays is variable, the complete consistency of the internal features linking them to one another and differentiating them from other dramatists' plays has ensured their all-but-universal acceptance as wholly his. More doubtful attributions must be assessed in their light. The same rule applies to attempts to apportion shares in collaborations in which Middleton is known or suspected to have taken part.

IV

Schoenbaum's comprehensive and amusing history of attribution studies in English Renaissance drama has served as a cautionary tale to later practitioners. But his mocking scepticism undervalues the achievements of nineteenth-and early twentieth-century scholars such as Fleay, Boyle, Sykes and, above all, Oliphant. Their methods were faulty, their arguments often little more than rationalizations of personal impressions, but the metrical data that they amassed is still of value and their reading in Elizabethan and Jacobean drama was wide and deep; Oliphant in particular read with rare sensitivity to the finer details of verse movement and verbal texture. From 1890 to 1930 Oliphant was unmatched in his appreciation of Middleton's stature as a dramatist, and his ability to detect his hero's hand was almost infallible. It was he who first proposed Middleton as likely author of The Revenger's Tragedy, The Lady's Tragedy, and A Yorkshire Tragedy, and as part-author of The Bloody Banquet, and who realized that Blurt, Master Constable was not Middleton's but Dekker's; Oliphant's pronouncements on other disputed and collaborative plays associated with Middleton have largely been vindicated by later studies. It was to Sykes's intuition that we owe the recognition that Webster shared with Middleton the writing of Anything for a Quiet Life and that The Spanish Gypsy is largely the work of John Ford (an ascription supplemented by Oliphant's subsequent realization that Ford had been partnered by Dekker). Sykes's support of the theory that Middleton contributed to Timon of Athens also turns out to have been well judged. Oliphant, Sykes, their contemporaries, and their predecessors identified the main canonical problems to be solved in Renaissance drama, and their books and articles are fertile sources of hypotheses to be tested by more objective criteria.

A huge advance in attribution studies in Renaissance drama was made by Cyrus Hoy in his uncovering of diverse 'linguistic patterns' among participants in the 'Beaumont and Fletcher' plays (1956–1962). The techniques had been used by a scattering of earlier scholars, but Hoy applied them with innovative thoroughness. Through meticulous counting of predetermined minutiae in the earliest printed texts, Hoy demonstrated that the Jacobean playwrights differed from one another in their habits with respect to certain word forms and colloquial contractions. For instance, in his unassisted plays Fletcher strongly preferred ye, 'em, and has, to you, them, and hath, and regularly employed such contractions as i'th', o'th', 'has, and 's for his. Massinger avoided ye, used 'em less frequently than Fletcher, made liberal use of hath, and employed i'th' and 's seldom, 'has only once, and o'th' never. Fletcher's fourteen plays contain no fewer than 4,507 examples of ye, whereas Massinger's fifteen contain only two. So in plays in which the two men collaborated their shares can readily be distinguished. The number of such alternatives possible in the Jacobean period—does or doth, I'm or I am, e'en or even, ha' or have, 'a or he, on't or of it, and so on—is very large, so that while any two dramatists will have some preferences in common, most display a distinctive overall pattern.

For the purposes of identification such trivial details are more useful than elements of greater literary import, for they are too inconsequential to encourage imitation and they are relatively unaffected by shifts in genre and stylistic register. Naturally they are subject to alteration by compositors and, more damagingly, scribes. The scribe Ralph Crane demonstrably overrode some of Middleton's habitual preferences, in his transcripts of The Witch and A Game at Chess, just as he overrode some of Shakespeare's preferences in The Tempest and other First Folio plays. Likewise, the 1639 quarto of The Bloody Banquet seems to have been set from a manuscript prepared by the same scribe who prepared copy for the 1639 quarto of Robert Davenport's A New Trick to Cheat the Devil, and who in both cases imposed some of his own habits (Taylor 2000).

But more often, a detailed investigation of the scribes and compositors who transmitted a play only demonstrates the reliability of the internal evidence. Thus, a knowledge of the stints and habits of First Folio compositors confirms that they cannot be held responsible for the evidence that Shakespeare collaborated with other authors in writing Henry the Sixth, Part One (Taylor 1995); likewise, the evidence for posthumous adaptation of Measure for Measure cannot be attributed to Ralph Crane (Taylor and Jowett). More generally, the regularity with which the plays of a particular dramatist such as Fletcher or Massinger or Ford are marked by much the same pattern, despite their diverse textual histories, is sufficient guarantee that authorial peculiarities were seldom drastically obscured by agents of transmission. A writer's habits may change over time, but in intelligible ways and without rendering his work unrecognizable.

Hoy's methods have been elaborated in connection with Middleton by Peter B. Murray in a study of The Revenger's Tragedy, by David J. Lake and MacDonald P. Jackson in independent surveys of the whole Middleton dramatic canon, and by R. V. Holdsworth in an examination of Timon of Athens. Basing their research on seventeenth-century printings or manuscripts (or photocopies of them), both Lake and Jackson checked dozens pg 84of items in Middleton's unassisted and undisputed plays and in a control corpus of over a hundred plays by all the prominent dramatists of the time and many of the lesser ones, and then applied the resultant information about Middleton's peculiarities to the doubtful and collaborative works. In forming their control groups both scholars biased the selection in favour of plays roughly contemporary with Middleton's, bearing the closest similarities to his, and written by dramatists who collaborated with him or had been nominated as alternative candidates for the authorship of disputed plays; each read every extant play written for the public stage within the period 1600–1627 (from Middleton's beginnings as playwright to his death); Jackson perused all but a handful of unobtainable plays recorded in the Schoenbaum-Harbage Annals of English Drama as written between 1590 and 1630, besides many others outside those dates. Since the two control groups differed and Hoy presented data for certain plays that were in neither of them, while Holdsworth has since compiled tables for all Shakespeare's plays, actual tallies of most pertinent features are available for over two hundred plays altogether. Moreover, Holdsworth, aware of Lake's and Jackson's findings, and seeking further indicators of Middleton's hand, checked stage directions in every available play, in print or manuscript, written between 1580 and 1642, a total of nearly 700 texts; in the course of these checks he kept an eye out for any signs of Middletonian linguistic practices. As well as the contractions, colloquialisms, word forms, and orthographical preferences investigated by Hoy, Lake or Jackson provided tallies for oaths and exclamations, synonymous connectives such as betwixt or between, among or amongst, while, whilst, or whiles, the affirmative particles ay, yes, and yea, group speech prefixes such as All or Omnes, certain rare spellings, unusual words favoured by Middleton, and collocations such as son and heir. The data leave no doubt whatsoever that the 'core' Middleton plays share a highly idiosyncratic linguistic and orthographical profile that is almost as reliable a guide to identification as actual physiognomy—or as fingerprinting. Every one of Middleton's twelve undoubted and unaided plays is more like every one of his others than is any play known to be by somebody else. And the fact that one of these Middleton plays, A Game at Chess, is preserved in his holograph offers extra reassurance that we are dealing with authorial characteristics.

A brief inventory of the salient discriminators will give some idea of the scope of recent work on the Middleton canon. Middleton employs the following contracted and colloquial forms (to standardize their spellings) at above-average rates: a'th', e'en, 'em, ex'lent, gi'n't (for 'given it'), 'had, 'has, I'd, I'm, I've, ne'er, on't, sh'as, 'tad, 'tas, they're, thou'rt, upo'th', y' (as in y'are and y'ave). Some of these, such as gi'n't, 'tad, and upo'th', are very rare, others more common. He is unusually fond of a as a weakened form of of, as in alate and a purpose. Like many playwrights, he makes considerable use of ha' and i'th' and some of 's for his, and he is exceptionally partial to enclitic 't, particularly after pronouns (as in for't). He prefers the modern, 'regulated' uses of the auxiliary verb do, rather than the 'unregulated' auxiliary found more often in older, more old-fashioned writers like Shakespeare (Hope). Middleton strongly prefers does and has to the formal and old-fashioned hath and doth, and tends to avoid such favourites of other dramatists as 'a as a colloquial form of he, d'ee (for do ye), of't, shannot, t'ee (for to ye), wonnot, ye, and the variants 'hem, o'the, and 'um. His most distinctive expletives—all rare and occurring at least ten times more often in his undoubted plays than in Jackson's control corpus—include a my troth, beshrewheart, cuds, cuds me, la you/why la, life, my life for yours, (a) pox, puh, push, and 'slife. Holdsworth, who cites nine examples of tak't of my word (or variations, such as a my/on my/upon my life/soul/truth) in plays associated with Middleton, was unable to find parallels elsewhere (180). Jackson listed twenty-eight comparatively common oaths, exclamations, and colloquialisms that Middleton employs more liberally than his contemporaries. At the same time, Middleton eschews a wide range of expletives, especially those with a hint of the blasphemous, that pepper the plays of his contemporaries both before and after the passing of the Act to Restrain Abuses of Players in May 1606. During the first decade of the seventeenth century Middleton persisted in preferring ay to yes, which was replacing it in most quarters, and in later years he remained partial to the old affirmative particle. In group prefixes he favoured All over Omnes. Among the idiosyncratic spellings sprinkled through his texts are -cst in such preterites and past participles as placst, do's for does, e'm (instead of 'em for them), ha's and h'as for has, and theire for they're. Middleton's practices with regard to connectives vary, though his preference for toward over towards is consistent and, like many dramatists, he avoids whiles. He several times employs some very rare words, such as the verb lin, and is unusually fond of certain collocations, such as give him his due (and the like). Holdsworth discovered that Middleton's and Heywood's texts are the only ones within the period 1602–1624 to furnish several entrance directions in the exact form 'Enter A (or A etc.) meeting B (or B etc.)'. He also compiled statistics on a stylistic trait that he calls 'interrogative repetition', showing that the rate of occurrence in Middleton plays is exceptionally high. The device, a means of maintaining momentum in dialogue, is illustrated at its most straightforward in an exchange between Allwit and Yellowhammer in A Chaste Maid in Cheapside, where Allwit says (my italics) 'I am the sorrier for't' and Yellowhammer responds, 'The sorrier? Why, cousin?' In many instances the tone of the repetition is exclamatory, but there is always a clear interrogative element as well. Holdsworth defined the device very carefully, spelling out rules for dealing with ambiguous cases.2

pg 85Such is the identikit of features from which a picture of 'Middleton' may be built up. When doubtful texts are matched against it, the likeness is sometimes so perfect that we can confidently associate them with our man. The Widow may instantly be added to the core of twelve 'undisputed and unaided' Middleton plays so as to enlarge their number to thirteen. Though published in 1652 as 'drawn by the art of Johnson, Fletcher, and Middleton', it was excluded from the Jonson and Fletcher Folios, and the catalogues of Rogers and Ley (1656) and Archer (1656) name Middleton alone; so did a seventeenth-century owner of the quarto, who crossed out Jonson's and Fletcher's names on the title-page. Since neither Jonson nor Fletcher specialists have been able to find traces of those authors, there has long been a consensus that the play is wholly, or almost wholly, Middleton's, and the thoroughly Middletonian linguistic profile confirms this orthodox view. In fact Hoy, Lake, Jackson, and Holdsworth draw upon The Widow as a source of data concerning Middleton's usages.

A few figures will suggest how compelling the indications of provenance may be. Middleton's liking for contractions with I (I'd, I'm, I've) and for e'en, ne'er, and on't is so pronounced that the totals for these six forms alone suffice to differentiate plays of the core canon from those in the control groups. According to Jackson's computations, the lowest total for an unaided Middleton play is 43 for A Trick to Catch the Old One. The range for Middleton's twelve remaining unaided plays is 63–144 and the average, to which most of the individual figures approximate, is 90; whereas the range for his one hundred control plays by thirty other dramatists is 0–55 and the average 16.3 Only Jonson's Bartholomew Fair (55) and Rowley's A Woman Never Vexed (53) afford raw totals higher than Middleton's lowest, but Bartholomew Fair is exceptionally long and A Woman Never Vexed considerably longer than A Trick to Catch the Old One. Once frequencies have been standardized (as rates per 20,000 words), separation between Middleton and non-Middleton texts is complete. Disputed full-length plays accepted into the Oxford canon as wholly Middleton's afford the following totals: The Puritan 94, The Revenger's Tragedy 104, The Lady's Tragedy 112, The Nice Valour 104. Such figures unequivocally categorize these four texts with the Middleton ones rather than with the non-Middleton controls. The eleven rare Middletonian expletives render the same decisive verdict in respect of The Puritan, The Revenger's Tragedy, and The Lady's Tragedy. The eight most distinctive of Middleton's unquestioned plays have the following numbers of instances: The Phoenix 13, A Trick to Catch the Old One 20, Your Five Gallants 30, Michaelmas Term 16, A Mad World, My Masters 27, A Chaste Maid in Cheapside 16, No Wit/Help like a Woman's 26, and The Widow 16. Sixty-seven of Jackson's one hundred control plays yield no examples at all, and only one has as many as five, one other four, and two others three. The total for The Puritan is 19, for The Revenger's Tragedy 14, and for The Lady's Tragedy 17.4

Since Lake, Jackson, and Holdsworth reached their conclusions, the Chadwyck-Healey electronic database 'Literature Online: English Drama' has made it possible to search some 4,000 plays by 700 dramatists from the late thirteenth century to the early twentieth. Use of this resource confirms that alleged Middleton markers are extremely rare outside his work (Jackson 1998). Take, for example, three unusual expletives: only six plays in the whole of 'English Drama' contain puh, my life for yours, and la you/why la, namely A Trick to Catch the Old One, Your Five Gallants, The Widow, Wit at Several Weapons, The Revenger's Tragedy, and The Puritan. The first three are undoubtedly Middleton's, and the other three have long been associated with him, Wit at Several Weapons having been identified by Lake, Jackson, and their predecessors as a collaboration between Middleton and Rowley. The checking in 'Literature Online' of other items that Lake and Jackson claim to be significant affords overwhelming support for the attribution to Middleton of The Revenger's Tragedy, The Puritan, and The Lady's Tragedy, in particular; it also strongly supports other evidence of his part-authorship of The Bloody Banquet.

In other cases the 'identikit' resemblance is not to Middleton but to somebody else. An appeal to internal evidence is no mere ploy for aggrandizing the canon. Testing for the Middletonian identifying marks leads to subtractions from the Dyce—Bullen corpus, as well as to additions. It supports the external evidence for Middleton's part-authorship of The Patient Man and the Honest Whore; on the other hand, it gives Dekker sole credit for the sequel, The Honest Whore, Part II. This example is in many ways emblematic: both Dyce (who included both parts) and Bullen (who excluded both) treated two distinct works as though they were one, and as a result one overestimated and one underestimated Middleton's work. In The Art of Thomas Middleton, David M. Holmes was anxious to demonstrate the place of Blurt, Master Constable 'in the nexus of Middleton's works' (17), but the play's linguistic profile is utterly unlike Middleton's and remarkably like Dekker's. This comedy, which was in print by 1602 and in which we might expect to find some twenty instances of Middleton's eleven most distinctive expletives, has none at pg 86all, but teems with Dekker's favourites; i'faith, which Middleton used very frequently in all his early plays but which Dekker almost completely avoided, is entirely absent from Blurt, Master Constable; and the contractions, connectives, speech prefixes for 'Omnes', and so on, exhibit a pattern that is uniquely Dekker's among dramatists writing for the private theatres around 1602. Likewise, The Family of Love, which several critics have considered not only a lively example of 'Jacobean city comedy' but also typically Middletonian in outlook and method, is so deficient in Middleton's distinguishing features that it cannot be accepted as his: a careful evaluation of the evidence strongly suggests that the play is wholly the work of Lording Barry (Taylor, Mulholland, and Jackson).

These negative findings with respect to Blurt, Master Constable and The Family of Love, far from contradicting the external evidence, help reinforce distinctions between good external evidence and bad. Blurt, Master Constable was first assigned to Middleton in Francis Kirkman's catalogue of 1661, and The Family of Love remained anonymous until Edward Archer assigned it to Middleton in his catalogue of 1656. These are very late attributions, made by booksellers whose claims are hopelessly unreliable, as W. W. Greg's careful analysis of Archer's and Kirkman's playlists, together with the 1656 list compiled by Rogers and Ley, amply demonstrated. The failure of the linguistic profiles of Blurt, Master Constable and The Family of Love to match Middleton's simply confirms Archer's and Kirkman's proven unreliability. And this confirmation lends further support, in turn, to the unequivocal verdict of the internal evidence that Middleton wrote The Revenger's Tragedy: the traditional ascription to Cyril Tourneur rests on the testimony of Archer, uncritically repeated by Kirkman. The 'Middleton identikit' thus gives warrant for an addition to the Dyce—Bullen canon that further undermines the late-seventeenth-century booksellers' authority and for two subtractions from the Dyce—Bullen canon that work to the same effect.

The same minutiae enable us to return an emphatic answer to W. D. Dunkel's question 'Did Not Rowley Merely Revise Middleton?' One telling detail will illustrate the point. In plays that he wrote alone Rowley never uses I've, to which Middleton was so partial. In the four plays and one masque that the Oxford Middleton includes as Middleton—Rowley collaborations—A Fair Quarrel, The Changeling, The World Tossed at Tennis, Wit at Several Weapons, and Old LawI've occurs twenty-eight times, always within scenes traditionally considered Middleton's handiwork, though his contribution to the actual writing of the five plays is, again according to the orthodox view, somewhat smaller than his partner's. Such a distribution of instances of the contraction must indicate actual collaboration between the two dramatists, not Rowley's mere revision of Middleton's scripts. Holmes's attempt to revive Dunkel's theory that these plays were conceived and composed entirely by Middleton alone, and that Rowley tampered with them later—a theory designed to magnify 'Middleton' at the expense of 'Rowley'—is plainly misguided. The division of labour in plays that Middleton wrote jointly with Dekker or Webster or Shakespeare can usually be deduced by the same means.

But the traditional scholarly concern to apportion shares in co-authored Renaissance plays has come under recent attack, and the whole question of the status of collaborations requires special consideration.

V

Drama is a collaborative art. The pristine script undergoes a complex process of modification as a company of actors and their various functionaries prepare it for its destined realization upon the stage. Without performers, a theatrical space, and an audience a playscript is mere 'words, words, words'. Even in the Jacobean theatre, costumers, carpenters, designers of props, bookkeepers, and musicians, as well as actors, were essential to the cooperative enterprise of turning handwritten signs into live entertainment. And the revival of an old play almost inevitably entailed some alteration of the promptbook that had guided earlier stagings. Scripts evolved to meet changes in public taste and the requirements of different venues. Songs might be interpolated, speeches and whole scenes cut or added or reordered. An originating author is but one of several agents in the creation of the theatrical event for which spectators pay their admission fees.

Moreover, the generation of those initial playscripts themselves was often enough a cooperative venture during the decades in which Middleton was active (Vickers 2002, 18–43). The diary of theatre manager Philip Henslowe records the transactions of a playwriting industry analogous to modern television scriptwriters' joint efforts to meet market demands. Bentley calculates that there are 1,100 or so plays of the period 1590–1642 about which we have evidence of authorship, and that close to 20 per cent 'contained the work of more than one man as either collaborator, reviser, or provider of additional matter'. Since amateur plays were seldom written in collaboration and title-page statements may simplify the true circumstances of composition, he guesses that 'as many as half of the plays by professional dramatists in the period incorporated the writing at some date of more than one man' (199). Nearly two-thirds of the 282 plays mentioned by Henslowe have more than one author. It is not unusual for Henslowe to pay five playwrights for a particular script.

In the frequency with which he collaborated, Middleton is more representative of his age than Marlowe, Jonson, or Shakespeare. Middleton appears to have been the sole author of nineteen extant dramatic texts: the twelve-play 'core' described in Section III above, plus A Yorkshire Tragedy, The Puritan, The Revenger's Tragedy, The Lady's Tragedy, The Nice Valour, The Widow, and Masque of Heroes. His joint dramatic works include Wit at Several Weapons, A Fair Quarrel, The World Tossed at Tennis and The Changeling with Rowley; Old Law with Rowley and Heywood; The Patient Man and the Honest Whore, The Bloody Banquet pg 87and The Roaring Girl with Dekker; The Spanish Gypsy with Dekker, Ford, and Rowley; Anything for a Quiet Life with Webster; and Timon of Athens with Shakespeare. Further, he seems to have adapted Shakespeare's Macbeth and Measure for Measure. In addition to these twelve extant collaborative works, Middleton combined with Dekker, Drayton, Munday, and Webster to write the lost Caesar's Fall, or Two Shapes, and Dekker explicitly credited him with a speech in The Magnificent Entertainment.

The prevalence of collaborative writing for the London stage during the early decades of the seventeenth century has lent extra force to modern critical campaigns for a shifting of emphasis away from 'post-Enlightenment' author-based approaches to the text onto 'the discourses, figures, locations, and cultural practices participating in its emergence', to quote Jeffrey A. Masten (1992, 352). According to the new paradigm, jointly scripted plays epitomize 'the social production of texts'. The apportioning of authorial shares in such works is problematic and ultimately futile. After all, 'the collaborative project in the theater was predicated on erasing the perception of any differences that might have existed, for whatever reason, between collaborated parts' (Masten 1992, 342). 'What does it matter who is speaking', asks Masten (337), quoting Foucault quoting Beckett. Even Nicholas Brooke, a critic of an older generation, asserts, in regard to The Changeling, that 'the labours of scholarship which have gone into trying to distinguish individual contributions are largely wasted'. Most of the castle plot has been assigned to Middleton, the madhouse plot to Rowley, but (Brooke objects) 'the stylistic tests adduced can only really demonstrate the difference of language between tragedy and comedy, which any single author, especially one brought up on Elizabethan theories, would make' (70).

However, Brooke is wrong and Masten exaggerates. Understanding of 'the social production' of dramatic texts in Renaissance England is surely increased by any information we can glean about the ways in which collaborating playwrights divided their labours. The obstacles to such gleanings are by no means as insurmountable as theorists imply. The fact that in some composite plays the shares of the contributing playwrights cannot be disentangled is itself of interest, since in others the divisions are sharp and clean. Obviously there are various routines by which two or more authors might have contributed to the play texts that survive. The more we can discover about these routines the better. It is true that co-authors, no less than single authors, strive for coherence and unity, but contrast is among the chief principles of Jacobean dramatic construction, so that two playwrights might fully indulge their talents for different kinds of dialogue and nevertheless create a unified aesthetic object. And the linguistic markers that enable us to differentiate Middleton's writing from Rowley's in The Changeling are largely impervious to generic shifts between tragedy and comedy: Middleton employs I've regardless of stylistic register, whereas Rowley employs it neither in his tragedy All's Lost by Lust, nor in his comedies A Shoemaker, a Gentleman and A Woman Never Vexed, nor in any of the material, comic or otherwise, in the other plays in which he had a share. Middleton favours the exclamations puh, push, and pish, whereas Rowley favours tush, regardless of context.

In the introduction to her New Mermaid edition of The Roaring Girl, Elizabeth Cook claims that 'Attempts to allocate lines or scenes of the play to either Middleton or Dekker beg as many questions as they answer and have tended to be based on pre-existing valuations of each author' (xxxvi). But she too is wrong. Critical 'valuations' of Dekker and Middleton are irrelevant to the techniques used by Lake and Jackson—whose books Cook completely ignores—to determine the probable shares of the two dramatists in the writing of The Roaring Girl. Among colloquial contractions and exclamations that Middleton, in his undoubted plays, favours much more than Dekker are for't, heart, hum, I've, life, on't, pist, puh, push, 'tad, and 'tas, and among forms much more commonly found in Dekker's plays are doth, Gods so, hath, humh, 'sdeath, 'sheart, tush, uds, umh, and zounds. Scenes 3, 4, 5, 8, and 11 (2.1, 2.2, 3.1, 4.1, and 5.2) contain forty-eight examples of the Middleton favourites and only three of the Dekker forms, whereas scenes 1, 2, 6, 7, 9, and 10 (1.1, 1.2, 3.2, 3.3, 4.2, and 5.1) contain thirty-two examples of the Dekker markers and only two of the Middleton ones (Jackson 1979, 95–101). Since neither playwright's avoidance of the other's preferred linguistic forms is absolute, the contrast in The Roaring Girl (48:3; 2:32) strongly supports Richard Hindry Barker's sense that in this play 'the shares of the two authors can be determined with some approach to exactness' (270). The same division of scenes leaves seventy-five parentheses in the 1611 quarto's printing of Dekker's share of the comedy, but only four in Middleton's (Jackson 1979, 97). This difference, which cuts across the compositorial stints, is also in line with the two men's contrasting practices, as these may be inferred from the plays they wrote alone. Metrical data and figures for the proportion of polysyllabic words yield similar allocations (Jackson 1979, 100–1, 208–9).

That Middleton's and Dekker's shares of The Roaring Girl can so readily be distinguished lends extra significance to the heterogeneous mix of markers in The Patient Man and The Honest Whore. In that play the absence of any definite clustering of Middleton's or Dekker's favoured forms— which both, however, occur—indicates a different mode of collaboration from that employed in the later comedy. It would be foolish to ignore such contrasts.

W. H. Auden once reported that in his dramatic collaborations with Christopher Isherwood the accommodation of each man's individual style to the other's allowed a third authorial personality to materialize, distinct from either writer's alone (260). Something akin to this phenomenon can be discerned in several Jacobean plays. The Middleton-and-Rowley pairing is not simply the sum of the two participants. 'Middleton, working alone, did nothing pg 88so tragically true as the finest scenes in "The Changeling" or so full of the expression of an extremely poetic sensitiveness as certain passages of "The Fair Quarrel"', declared Oliphant. Whether or not we agree with this judgement, Rowley—a strangely uneven writer, more celebrated in his own day as comic actor—certainly 'seems to have had the ability to draw out Middleton's best' (Oliphant 1929, II, 13). T. S. Eliot considered Middleton's greatest achievement in comedy to be his creation, with Dekker, of the 'perpetually real … and human figure of Moll the Roaring Girl' (99). Nevertheless, although Middleton-and-Rowley cannot be straightforwardly equated with Middleton and Rowley or Middleton-and-Dekker with Middleton and Dekker, authors with their personal creative gifts remain crucial agents in the generation of collaborative texts. Middleton and Rowley combined to write plays unlike those of Beaumont and Fletcher or even of Middleton and Dekker. Beaumont and Fletcher could never have created The Changeling. When partnered by Rowley, rather than Dekker, Middleton wrote nothing akin to The Roaring Girl.

Thus, the compound of two authors may produce an effect different than the work of either author in isolation. Failure to recognize this fact not only leads to critical simplifications, reducing dialogic complexity to monologic banality; it can also hamper attempts to determine the authorship of a play in the first place. For much of this century, The Bloody Banquet was excluded from the Dekker canon because Dekker scholars felt that many of its features were unDekkerian, and was excluded from the Middleton canon because Middleton scholars felt that many of its features were unMiddletonian. This conundrum could only be solved by examining the play not as a whole, but scene by scene, in the process demonstrating that unDekkerian features concentrate in scenes which show strong signs of Middleton's authorship, and that unMiddletonian features concentrate in other scenes which show strong signs of Dekker's authorship (Taylor 2000). The resulting play does not belong to either author, but to both.

Twentieth-century writers, collaborating on a film or television script, often sit in one room developing plot and dialogue, scene by scene and speech by speech, through mutual give and take. Presumably English Renaissance dramatists sometimes operated in this way, or reworked, pruned, and expanded each other's drafts in a complicated layering of individual contributions; this may have happened in The Patient Man and the Honest Whore, and again in 2.1 of The Bloody Banquet, where features of both Dekker and Middleton seem inextricably mixed. But the method of collaboration best attested in the early modern theatre involved the allocation to the participating authors of responsibility for structural units—whole acts or scenes. This kind of clear-cut division would have facilitated the meeting of Henslowe's deadlines. Other procedures are theoretically possible, and appear to have been not uncommon in practice—division by 'main plot' and 'sub-plot', by serious and farcical episodes, or by some other categorization of subject matter. An author such as Anthony Munday, whom Francis Meres commended as 'our best plotter', might have sketched an outline for others to fill in (Chambers, 1930, II, 195). And once a script had been completed, one or other of the original collaborators might have given it a last polishing, or somebody else entirely might have revised it long afterwards.

Middleton and Rowley's success as a team was founded on a modus operandi that allowed each to follow his true bent—Middleton's for a skilled probing of the motives and behaviour of psychologically complex characters and for satirical comedy marked by pungent wit and multiple ironies, and Rowley's for a vigorous melodrama of moral blacks and whites and for boisterous clowning. The nature of Middleton's involvement in his collaborations with other playwrights doubtless varied, but, without aspiring to know 'who is speaking' each and every word of these plays, we may legitimately search for clues to the modes of authorial interaction that created them. Attempts to apportion 'shares' in the writing of collaborative plays lack the urgency of attempts to identify the participating authors, since the inclusion or exclusion of a play within The Collected Works of Thomas Middleton depends on Middleton's presence, rather than on the precise extent of that presence. But to take no interest in 'who wrote what' within plays of dual or multiple authorship would be to lack historical curiosity and to hamstring exploration of the art of Middleton, his collaborators, and the Renaissance theatre in general.

VI

But Middleton did not limit himself to the theatre. Unlike Shakespeare, Jonson, and Fletcher, Middleton wrote many civic pageants and a number of topical and popular prose pamphlets. Unfortunately, we cannot assume that such works share the same authorial characteristics as the plays; it is always possible that genre may have affected style. The diversity of Middleton's output adds to the difficulty of determining a canon. And although investigators of canonical problems in the dramatic canon can draw upon more than a century of scholarship devoted to such issues in the plays of Middleton and his contemporaries, much less attention has been paid to authorship problems in non-dramatic genres.

Middleton's civic pageants and entertainments were the most collaborative of all his works. Like a director of Hollywood epics, he organized extravagant spectacles requiring many helpers. Fortunately, few uncertainties of authorship surround Middleton's narrative descriptions of these shows, to which his name is unequivocally attached.5 His long poems likewise have remained unquestioned since pg 89the discovery of his true date of birth revealed The Wisdom of Solomon Paraphrased, Microcynicon and The Ghost of Lucrece to be juvenilia.

The true extent of his prose writing, by contrast, is much more difficult to determine, in part because such pamphlets were routinely published with little or no indication of authorship. Only three extant non-dramatic works have Middleton's full name on the title-page (The Wisdom of Solomon Paraphrased) or appended to a preface (The Ghost of Lucrece, Sir Robert Sherley). Though The Peacemaker was published anonymously, or rather implicitly pseudonymously attributed to King James, an extant manuscript licence specifically attributes it to Thomas Middleton. Three other pamphlets (Microcynicon, The Black Book, Father Hubburd's Tales) attach the ambiguous initials 'T.M.' to some prefatory material; in The Two Gates of Salvation, a second issue adds an initialled dedication, linking it to Middleton's post as Chronologer of London. Thus, although there is varying external evidence for Middleton's authorship of eight non-dramatic pamphlets, only one of those—the very first, published when he was seventeen—names him on the title-page. Clearly, no one expected Middleton's name to help sell a pamphlet, and after his trumpeted adolescent entrance into print Middleton himself abandoned the self-importance which would have insisted on conspicuous advertisement of his authorship. It would therefore hardly be surprising if some of his pamphlets, like some of his plays, were published anonymously.

But we can identify such anonymous pamphlets only if we first have a database of undisputed pamphlets. That database has been slow to form. Middleton's non-dramatic canon has only gradually become apparent to modern scholars. Dyce was ignorant of the manuscript evidence for Middleton's authorship of The Peacemaker; likewise, The Ghost of Lucrece does not appear in the nineteenth-century Works because its sole surviving copy did not resurface until the twentieth century; Dyce and Bullen did not include Two Gates because they were unaware of the attribution to Middleton in the second issue.

This gradually expanded core canon of non-dramatic works, once established, makes it possible to investigate the relationship between the style of Middleton's plays and the style of his pamphlets. Moreover, although Middleton created characters and wrote speeches in verse for the Lord Mayor's shows, the published pamphlets issued in conjunction with those shows chiefly consist of narrative and descriptive prose; they therefore expand the reliable database of Middleton's non-dramatic prose.

Much of the internal evidence available for plays is not available, or seldom available, for pamphlets. Expletives and colloquial contractions, for example, are far less prevalent in expository or narrative prose than in dialogue for the stage; verse style is easier to describe and quantify than prose style. Moreover, there are a few significant differences between Middleton's habits in plays and his habits elsewhere: in writing dialogue he strongly preferred the colloquial 'has' and 'does', but in non-dramatic works he normally uses the older 'hath' and 'doth'. Nevertheless, even after all these difficulties have been taken into account, Lake showed that internal evidence could distinguish between pamphlets by Middleton and pamphlets not by him (270–3). The choice of connectives may have confirmatory value, and in some cases a sprinkling of rare Middletonian spellings affords further corroboration. Middleton's hand shows itself most clearly in the sporadic appearance of words and collocations to which he was peculiarly partial and in a heavy accumulation of parallels in thought and expression to his acknowledged works.

The citing of 'parallels' in attempts to identify the authors of Renaissance texts has fallen into disfavour, in reaction to the abuse of such evidence by some earlier investigators, who were too apt to bolster a weak case with mention of a smattering of trivial verbal similarities of a kind that might link almost any two contemporary works, whoever had written them. Both the quality and the quantity of the parallels are crucial in the interpretation of what they imply. Middleton is demonstrably self-repetitive. He echoes other writers infrequently, but often recycles his own ideas, images, and phrases, so that plays, masques, poems, and prose writings are interconnected in an intricate network of striking reminiscences and anticipations. Holdsworth's methodical and persistent exploration of these interconnections has gone far towards re-establishing parallels in the investigator's armoury of usable evidence. The absence from Blurt, Master Constable, 2 Honest Whore, and The Family of Love, of significant Middleton parallels strongly confirms the linguistic indications that he had nothing to do with those disputed plays, and the paucity of such parallels in The Spanish Gypsy proves that others must be responsible for the bulk of that play. In The Revenger's Tragedy, The Puritan, A Yorkshire Tragedy, The Lady's Tragedy, The Nice Valour, and certain scenes of The Bloody Banquet, by contrast, passages that can be closely paralleled in indubitably canonical plays are reassuringly plentiful. In the collaborative plays Middleton parallels show a marked tendency to cluster within the scenes and sections attributable to him on other grounds. Holdsworth's exhaustive analysis of Timon of Athens shows that the shares of Middleton and Shakespeare are, in each case, distinguished by the abundance of parallels with the appropriate collaborating author's corpus and the scarceness of parallels with his partner's.

'Literature Online' allows even greater systematization of such searches for rare phrases and collocations that are shared between an anonymous, collaborative, or disputed, work and rival candidates for its authorship. Comprehensive lists of links with writers belonging to some chronological or other pertinent group can be compiled, and the results analysed. The method, which can identify, from a lineup of writers, the true author of a passage of known provenance, has been fully described by Jackson (1998; 2003, 190–217). Taylor (2002) has employed it to help pg 90show that Thomas Heywood contributed 345 lines to Old Law.

Early in the twentieth century, parallels were used by Sykes (1925) and Adams to establish that the author responsible for The Wisdom of Solomon Paraphrased and The Ghost of Lucrece (by 'Thomas Middleton') also wrote Microcynicon, The Black Book, and Father Hubburd's Tales (by 'T.M.'). Using similar evidence, the Oxford Middleton has included five pamphlets published anonymously. Three of these—The Penniless Parliament of Threadbare Poets, Plato's Cap, and The Owl's Almanac—are mock almanacs, clearly linked to one another, generically and textually, and also clearly linked to Middleton's lifelong interest in almanacs. Two other pamphlets—News from Gravesend and The Meeting of Gallants—seem to be collaborations with Thomas Dekker, with whom Middleton also collaborated in the theatre; both were written in 1603–4, when both men were writing pamphlets in order to earn a living while the theatres were closed by plague.

In all five anonymous pamphlets where Middleton's hand has been detected, the attribution is strengthened by circumstantial evidence. Such circumstantial plausibility also reinforces his claim to the handful of individual poems attributed to him in the seventeenth century: a manuscript epitaph on Richard Burbage, a commendatory poem printed in the first edition of The Duchess of Malfi, and a poem about his imprisonment for A Game at Chess. Statistical tests are helpless in analysing texts of such brevity, but there is no reason to doubt these attributions: unlike Donne, Middleton was not famous enough to attract casual attributions in manuscript miscellanies. Indeed, his name is much more likely to have been omitted from poems he wrote, than to have been added to poems he did not. The Oxford edition attributes to him three anonymous poems: a verse description of the engraving on the title-page of the first edition of A Game at Chess, an epitaph on Sir George Bolles, and a poem celebrating the dedication of the new parish church of St James. Each of these texts is circumstantially linked to Middleton, and each strongly resembles the verbal texture of his acknowledged work.

It is of course entirely possible that Middleton wrote other anonymous poems or anonymous pamphlets. Indeed, if The Two Gates of Salvation (mostly an anthology of Biblical passages) and Sir Robert Sherley (mostly a translation of a Latin description of a pageant) were not explicitly attributed to him in what appear to be reliable documents, it is unlikely that any literary critic would have suspected or accepted his authorship of either. If he translated other continental works, or composed other religious pamphlets, they would probably escape our detection. Barring the unpredictable rediscovery of lost plays, Middleton's dramatic canon seems relatively fixed; his non-dramatic canon may not be.

VII

Fingerprints, though invaluable for identifying people, tell us nothing about their qualities as human beings. Similarly, the linguistic details that help us decide whether a play or pamphlet is by Middleton are of little literary importance, though an orthography that marks colloquial elisions and weakened pronunciations does tend to accompany a poetic style heavily indebted to everyday speech. But to the ear attuned to its rhythms and tones, Middleton's verse is highly distinctive, especially in plays written from 1611 onwards. An attempt at a brief description of Middleton's style, both before and after 1611, inevitably raises the subject of chronology.

In thinking about chronology we need to distinguish between the date at which a work was first exposed to the public, whether in performance or in print, and the date at which it was composed. Performance or publication would usually have followed closely upon completion of the writing of a playscript or pamphlet, but composition is sometimes a protracted business: works may be started and laid aside, to receive intermittent attention over periods of many years. We cannot hope to chart all the details of the evolution of Middleton's writings before their release to the public. Our main ambition must be to determine, if only approximately, the date at which a play was first ready for the stage or a pamphlet or poem for the printing-house. To try to establish dates of this kind is not to privilege study of the author's 'mind and art'. The new historicist or cultural materialist must inevitably share the traditional scholar's concern with chronology. The more precisely a work can be located in time, the more confidently it may be referred to its specific cultural and political context. Critics interested in 'intertextualities' also need to know the order in which texts appeared.

For the dating of Middleton's writings we are reliant mainly on documentation from outside them, on topical allusions within them, and on the ways in which they have influenced, or been influenced by, works whose dates can be fixed with some certainty. The notoriety of A Game at Chess ensures that we know precisely when this box-office success was the nine days' wonder of the London stage. The mayoral pageants were designed for specific occasions. Plato's Cap, 'Cast at this year 1604, being Leape-yeere', and The Owl's Almanac in 1618, were, like modern calendars and appointment diaries, obviously written to be marketed by the New Year. Henslowe's record of payment to Dekker and Middleton, some time before 14 March 1604, of five pounds 'in earnest' of Patient Man and Honest Whore, and the entry of Part 1 in the Stationers' Register on 9 November 1604 set clear limits between which it must have been completed. For several works there is similarly conclusive evidence. For others the indications are few and ambiguous, so that the range of dates for which different commentators have argued span more than a decade. Nineteenth-century scholars regarded Hengist, King of Kent; or, The Mayor of Queenborough as one of Middleton's very first plays; it is now generally assigned to the period 1616–20, and this edition presents new evidence for 1620. Such major readjustments to a chronological ordering of Middleton's works obviously affect our sense of their interconnections, pg 91and although in this case we can be reasonably confident that modern scholarship has corrected an earlier mistake, scholars continue to disagree about the dating of some plays. Nevertheless, an account of Middleton's dramatic verse and the changes it underwent may be attempted, if only as a stimulus to further research. The necessary work has barely been begun. Understanding of Middleton's development has yet to profit from the kinds of analysis that have been devoted to the Shakespeare canon.

An immense amount of scholarly activity has gradually built up a clear picture of Shakespeare's stylistic development. During the nineteenth century Fleay, his New Shakspere Society associates, and several German scholars subjected his prosody to minute analysis, calculating the percentages of double endings, run-on lines, trochaic substitutions, anapaestic feet, light and weak endings, and so on, in the blank verse of each play. Twentieth-century scholars have extended and refined this research. When metrical figures are related to chronology, in so far as it can be determined by other means, Shakespeare's verse style can be seen to evolve in intelligible ways. Broadly speaking, the trend is away from a succession of strictly iambic, decasyllabic, end-stopped lines towards greater freedom, variety, and flow. Where external evidence for a Shakespeare play's date of composition is lacking, the metrical figures allow an approximate placing within the overall evolution. Similarly, the formal similes of the early plays give way to a densely metaphoric language in which images blend into one another in kaleidoscopic profusion. Statistical studies of Shakespeare's vocabulary reveal a tendency for plays written at about the same time to share more of their rare-word vocabulary than plays written further apart. An index of 'colloquialism in verse'—in which various contracted and colloquial forms are totalled and related to the number of words within each play's verse—reveals a marked increase in the plays written after 1600. With the help of such data, the approximate order in which Shakespeare's plays were composed has by now been determined to the satisfaction of most scholars (Taylor 1987).

Leeds Barroll and others (Bristol, de Grazia) have decried the use of internal evidence in attempts to establish a Shakespearean chronology. Some of these complaints are specific to the particulars of Shakespeare, and do not apply to Middleton or other authors. For instance, Shakespeare's plays are characteristically set in some other place and/or some other time; so-called 'topical allusions' to current events in contemporary London would therefore be dramatically out of place, and they are correspondingly rare, and sometimes disputed. Middleton, by contrast, characteristically works in a temporal and spatial present, and his texts teem with references to current events, most of them quite indisputable. Intrinsically, as indications of the date of composition, such allusions are much more reliable in Middleton's canon than in Shakespeare's.

But some of the objections raised by sceptics are not specific to Shakespeare, but claim instead to be more generally applicable theoretical strictures. Barroll, for instance, stigmatizes the very term 'internal evidence' as 'most often an oxymoron' (238). But his criticisms rest on a misconception. He takes percentages for a selection of metrical features that rise or fall over Shakespeare's career, shows that they all order the plays differently, and concludes that they are therefore unsatisfactory, because 'If one uses tests whose results are expressed in numbers, quantitative measures must be rigidly adhered to within the scientific universe created' (238–9).

No professional statistician would agree with him. Barroll misrepresents the principle that underlies application of internal tests to matters of dating—the familiar and pertinent statistical concept of correlation. The blank verse of Shakespeare's late plays contains more double endings than the blank verse of his early ones, but nobody supposes that each successive play employed a higher proportion of double endings than the one before in a consistent and uniform development, nor that a play for which the percentage is 25 was necessarily composed later than a play for which it is 21. Artistic change is not so steady and predictable. Rather, there is a relation between chronological sequence, in so far as it may be established by external evidence, and an inverse rank order based on the magnitude of the percentage of double endings in the plays' blank verse. Mathematical measures of correlation can gauge the degree and significance of this relation. There need not be perfect coincidence between the two orderings for a statistically significant relationship to obtain. Several metrical features correlate with chronology in this way, and when the most reliable are combined into a single index the correlation becomes especially strong.

Indices based on colloquial contractions, on vocabulary links, or on rates of occurrence of function words yield similar correlations with a skeleton chronology based on references to the plays in historical documents, their allusions to contemporary events, and their intertextual connections with datable non-Shakespearean works. Since the different kinds of internal evidence—metrical, orthographical and linguistic, lexical—are independent of one another, when they agree in indicating a particular period for a play's likely date of composition, they come close to settling the matter, even though 'external' confirmation may be lacking. Such considerations cannot tell us whether Coriolanus preceded or followed Antony and Cleopatra, but they leave us in no reasonable doubt that both plays were written later than, for example, Othello. Cautiously interpreted, the evidence of quantifiable stylistic variables is an invaluable aid towards dating a writer's works.

By comparison with Shakespeare's, Middleton's stylistic development has been only roughly charted.6 Few metrical pg 92data have been accumulated, and no concordance has existed to serve as basis for statistical analysis of his vocabulary. With the possible exception of Ants Oras's analysis of 'pause patterns', to be described shortly, nothing remotely comparable to the intensive and wide-ranging work of Hoy, Lake, Jackson, and Holdsworth on questions of authorship relating to Middleton has been devoted to uncovering metrical and other stylistic indicators for a Middleton chronology.

As a dramatic poet Middleton has in the past usually been considered 'workmanlike', but less gifted than Webster, for example. The poetic vitality of The Revenger's Tragedy has been amply acknowledged, but the credit has gone to Tourneur, formerly thought to have been the play's author. The lukewarm judgement on Middleton as poet owes much to late-Romantic notions of poetry that predate Modernism, let alone more recent movements. In fact, Middleton is a truly remarkable dramatic poet. His achievement as a dramatist is inseparable from his achievement as a poet, for the poetic medium he devised is a subtle vehicle for conveying his characters' anxieties, pretences, and compulsions. His verse is seldom showy. It is less rhetorical than Shakespeare's, mimicking more closely the rhythms and syntax of colloquial speech. But it moves with a peculiar nervous intensity, partly due to the lively interplay between units of sense and line divisions. Middleton's imagery appeals to the intellect, rather than to the eye: it is not flamboyant, but relies on personifications that remain unelaborated and crowd one upon another. At its best Middleton's verse is quick with metaphorical activity. And, as Oliphant noted, he is the one Renaissance dramatist who vies with Shakespeare for 'sovereign mastery of words' (1927, 75). He has an acute awareness of multiple meanings, so that he is not only addicted to the pun and quibble but uses a word less as a single note than as a chord (Ricks 1960, 1961).

As a metrist Middleton was strikingly original, but mainly in his post-1611 plays. His earlier style owed more to rhyme, tolerated more inversions of normal word order, and was somewhat less apt to depart from the decasyllabic norm; but it also mixed prose and verse easily and readily, sometimes moving from one to the other and back again within the same speech. The verse of The Revenger's Tragedy, in which the earlier style culminates, is a medium that can accommodate an extraordinary range of tones. The court characters in The Revenger's Tragedy may be caricatures of lust and ambition, but their dialogue is eminently speakable. It sounds idiomatic, registers nuances, and invites changes of pace. It conveys a sense of real life that offsets the elements of morality-play abstraction.

Seventeenth-century dramatists rung more and more changes on the basic, end-stopped iambic pentameter line first brought to vigorous life in Marlowe's Tamburlaine, but freedom from the tyranny of the norm was accomplished in two main ways, involving two different treatments of line ends. The first was the addition of a final unstressed syllable to create a so-called 'double' or 'feminine' ending. An occasional variation in the verse of Shakespeare and his early contemporaries, in Fletcher's it became normative: a majority of Fletcher's blank verse lines have feminine endings, and many of these are monosyllables that carry considerable weight. The second was the use of enjambment, or the running-on of the sense from the end of one line into the beginning of the next, with a consequent increase in mid-line pauses and a tendency for these to fall at later points in the pentameter. This feature, too, was at first employed sparingly, but became almost ubiquitous in the verse of such Caroline playwrights as Shirley and Davenant. Extreme forms of enjambment, whereby lines end on prepositions, conjunctions, auxiliary verbs, and the like, can so predominate in a passage of Davenant's verse as to make it seem like chopped-up prose. Shakespeare was among those who adopted a judicious mix of variations, while never allowing the essential structure to collapse. Throughout his career Middleton favoured end-stopping, eschewing the 'weak' and 'light' endings created by extreme enjambment, preferring to gain flexibility through an almost Fletcherian use of feminine endings. Yet Middleton's verse, with its colloquial contractions and natural syntax, has no trace of the lullaby of sentiment that can mark Fletcher's cadences; it has a tense and vibrant music of its own.

Besides double endings, including those in which the extra syllable is emphatic, Middleton had an unusual fondness for 'triple' and even 'quadruple' endings, in which the stressed first syllable of a word such as 'gentlemen' or 'gentlewomen' completes the pentameter line and two or three relatively unstressed syllables follow. Like Fletcher, he often creates double or triple endings by adding vocatives, such as 'sir' or (more idiosyncratically) the trochaic 'madam', to an otherwise complete line. By means of anapaests and other trisyllabic feet he tends to crowd his lines with extra syllables, which in pronunciation may be slurred or elided to fit the measure. Particularly characteristic is his habit of making double endings out of 'on't', 'for't', or some similar contraction, or out of 'else', as in 'You must keep counsel, mother, I am undone else' (Women Beware, 1.1.46).

Certain turns of syntax facilitate Middleton's metrical licences and add variety to the strength and position of pausation within lines. Oliphant noted the most conspicuous of these. Endemic is his fondness for making subordinate what in other writers would be a sentence's principal clause, as in 'Will make the rest show nothing, 'tis so glorious' (No Wit, 8.195), where the more common construction would be ''Tis so glorious that it will make the rest show nothing'. 'I thank' is frequently used in this manner, as in 'He came in a good time, I thank him for't' (No Wit, 6.172). Also characteristic are the employment of emphasizing or confirmatory phrases beginning with 'That' or 'There', as in 'And know it to be mine—there lies the blessing' (Women Beware, 1.1.15); conditional approval expressed by 'I'll say that for'; the use of the superlative followed by 'that ever', as in 'Now say if't be not the best piece of theft | That ever was committed' pg 93

Play

Double

Rhymed

Slurred

The Phoenix

30

27

29

Michaelmas Term

20

33

32

A Mad World, My Masters

31

29

23

A Trick to Catch the Old One

27

24

14

Your Five Gallants

21

19

23

No Wit/Help like a Woman's

33

15

25

A Chaste Maid in Cheapside

31

9

19

The Witch

32

10

18

More Dissemblers Besides Women

33

11

20

Hengist, King of Kent

24

9

21

A Game at Chess

42

12

Table 1. Percentages of verse lines with double endings, rhyme, and slurred syllables in eleven Middleton plays, according to Matthew Baird, 1928.

(Women Beware, 1.1.43–4); and the habit of attaching ability to inanimate things, as in 'Now for a welcome | Able to draw men's envies upon man' (Women Beware, 3.1.102–3).

These idiosyncrasies become much more prominent after 1610, although the stylistic shift may seem more sudden and radical than it was, simply because we apparently possess only a single collaborative play written between early 1607 and early 1611 (The Bloody Banquet). If more of his dramatic work from that period were extant, the shift might appear more gradual. In any case, the changes in his dramatic verse are potentially quantifiable, but full and accurate figures have not yet been computed. In a thesis of 1928 Matthew Baird presented his calculations of percentages of double endings, rhymed lines, and lines with slurred syllables in eleven Middleton plays. His results are given in Table 1.

The data show a diminution in rhyme after the city comedies, with the last of those having the least rhyme. The figures for 'slurred syllables' fall into no definite pattern, and there appears to be no appreciable increase in double endings until A Game at Chess. However, the complete accuracy of Baird's counts may be doubted, and the precise basis of his calculations was not made clear. Figures compiled by Barker (1945), for a small selection of plays and from samples only, agree quite well with Baird's, but show a significant increase in feminine endings in plays written after 1606/7 (Table 2). The probable reason for Barker's percentages being higher than Baird's is that Barker's samples were exclusively of blank verse, in which feminine endings are normally more common than in rhymed lines, and that his 'feminine' endings included 'triple' endings, whereas Baird's perhaps did not.

Data for some collaborative plays that have been tested show Middleton's shares fitting well with his unaided plays of about the same date and differentiate these shares from Dekker's in The Roaring Girl and from Rowley's in A Fair Quarrel, The Changeling, and Wit at Several Weapons. According to Baird, in Middleton's scenes of The Roaring Girl 34 per cent of the verse lines have

Play

Lines tested

Feminine

(per cent)

The Phoenix

250

27

Michaelmas Term

150

24

The Revenger's Tragedy

400

28

The Lady's Tragedy

500

42

No Wit/Help like a Woman's

500

45

A Chaste Maid in Cheapside

400

46

Table 2. Percentages of sampled verse lines with feminine endings according to Barker, 1945.

double endings, 14 per cent have rhyme, and 21 per cent have 'slurred syllables', whereas the figures for Dekker's contribution are 14 per cent double endings, 41 per cent rhyme, and 14 per cent 'slurred syllables', in accord with his practices elsewhere. Barker's counts of feminine endings in verse scenes of The Changeling show an overall figure of about 54 per cent for Middleton, and 31 per cent for Rowley (Schoenbaum 1955, 211). These figures confirm Pauline G. Wiggin's account of the differences between Middleton and Rowley as metrists. She also noted the relative frequency of triple endings in Middleton's share of The Changeling, the prevalence of end-stopped lines, and the orthodox disposition of stresses; in Rowley's share she found almost no triple endings, about twice Middleton's proportion of run-on lines (including those producing weak endings), and frequent inversions of stress in the second and last foot. Wiggin also gives tallies (30–1) that enable us to reckon the percentages of feminine endings in the two dramatists' shares of A Fair Quarrel: for Middleton the figure is 46, and for Rowley 26. Over a century ago, Robert Boyle's metrical data for the 'Beaumont and Fletcher' plays included some figures for Wit at Several Weapons; when these are converted to percentages and related to the probable division of labour between Middleton and Rowley, the total for Middleton is 42 per cent feminine endings and 9 per cent run-on lines, and for Rowley 24 per cent feminine endings and 19 per cent run-on lines (Jackson 1979, 210). In order to be strictly comparable, counts must be made by one person at one time. The fifty lines that Barker tested in 5.3 of The Changeling have inflated his overall percentage for Rowley in that play: Rowley's other scenes are much lower, and 5.3 is probably of mixed authorship. However, it does seem that Middleton used feminine endings more liberally after 1611.

The fullest investigation of any aspect of Middleton's prosody (and the only chronological study which competes in thoroughness with the authorship studies by Hoy, Lake, and Jackson) was undertaken by Ants Oras in the course of his research on pause patterns in the verse of English Renaissance drama. Basing his computations on the earliest printed texts, Oras counted the numbers of pauses falling in the various possible positions —after the first, second, third, fourth, fifth, sixth, seventh, eighth, or ninth syllable within the line. Study of pg 94

Table 3. Ants Oras's figures for percentage of pauses in first half of the pentameter line in Middleton's plays.

Play

Percentage of pauses in first half of line

The Phoenix

33.0

Michaelmas Term

30.4

A Mad World, My Masters

22.5

A Trick to Catch the Old One

30.4

The Revenger's Tragedy

33.8

Your Five Gallants

40.0

No Wit/Help like a Woman's

23.4

The Lady's Tragedy

27.7

A Chaste Maid in Cheapside

19.0

The Witch

17.4

More Dissemblers Besides Women

20.9

The Widow

22.7

Hengist, King of Kent

26.5

Women, Beware Women

22.5

The Changeling (Middleton's share)

19.8

A Game at Chess

18.4

a wide range of dramatists revealed a changing rhythmic climate, to which individuals responded in different ways. Oras presented his findings in a series of graphs, all drawn to the same scale and showing for each play the percentages of pauses that fall in the nine positions within the blank verse or rhymed pentameter lines. His least vulnerable data—unaffected by scribes or compositors—are for pauses corresponding to divisions of a line between two or more speakers. The graphs for Middleton illustrate the evolution of his pause patterns, each graph tending to be most like that for other plays written at about the same time. Graphs for most early plays show a peak at the sixth syllable, preceded by a lesser but still significant peak at the fourth. In graphs for plays from 1611 onwards the secondary peak recedes or disappears and the main peak at the sixth syllable rises even higher; pauses after the seventh syllable also tend to increase, until in Middleton's share of The Changeling and in A Game at Chess they outnumber even those after the sixth. As listed in Table 3, Oras's raw figures show pauses in the first half of the pentameter line (after syllables 1–4) decreasing, when these are calculated as a percentage of all pauses except those after the fifth syllable.

Although Oras's data reveal a far from steady chronological development, they are decisive in associating The Revenger's Tragedy with the early Middleton, rather than with Tourneur. A mathematical measure of the degree to which Oras's various graphs match one another— the Pearson product moment correlation, using the raw figures—reveals that each of the thirteen core Middleton plays (adding The Widow to the unquestioned dozen) affords a closer fit to The Revenger's Tragedy than does The Atheist's Tragedy, and that closest of all is A Trick to Catch the Old One, the very play that was coupled with The Revenger's Tragedy in a Stationers' Register entry of 1607.

Table 4. Totals for early and late forms in unaided, full-length Middleton plays. Early forms consist of the expletives a/by my troth, by the mass, foh, la you/why la, pist, puh, 'sfoot, 'slid/'slud, and tut; late forms are hold my life, hoyda, pish, and o'th'. Some counts include slight variations, as defined by Jackson 1979, 70–1 and 79, n. 10. Data are drawn from Jackson and Lake.

Play

Early/late forms

The Phoenix

38/1

Michaelmas Term

31/1

A Mad World, My Masters

63/2

A Trick to Catch the Old One

37/1

The Puritan

51/1

The Revenger's Tragedy

26/0

Your Five Gallants

64/0

No Wit/Help like a Woman's

16/25

The Lady's Tragedy

5/4

A Chaste Maid in Cheapside

7/12

The Witch

0/6

The Nice Valour

4/13

More Dissemblers Besides Women

1/9

The Widow

4/20

Hengist, King of Kent

0/14

Women, Beware Women

3/15

A Game at Chess

2/0

Oras furnished line-split data for two hundred and fifteen plays beside The Revenger's Tragedy: when each of these is in turn matched against it, four of the best dozen matchings prove to be with Middleton plays, namely A Trick to Catch the Old One, The Phoenix, Hengist, King of Kent, and No Wit/Help like a Woman's. James Shirley is the sole other dramatist with more than a single play among the top dozen matchings, and he has only two, The Traitor and Hyde Park, both of which belong to the 1630s—a quarter of a century after The Revenger's Tragedy was composed— and share none of its other metrical characteristics. Even if we had no other evidence of Middleton's authorship of The Revenger's Tragedy, Oras's counts of pause patterns would establish him as the best candidate (Jackson 1983, 28–9). They also buttress the case for adding The Lady's Tragedy to the canon: Oras remarks on 'the extraordinary likeness of its line-split design to the mature patterns of Middleton, quite particularly in those plays closest to it in time, above all that of The Witch' (Oras, 32).

Another chronological indicator deserves consideration. Some expletives are more prevalent in Middleton's early plays: a my troth, by my troth, by the mass, foh, la you/why la, pist, puh, 'sfoot, 'slid/'slud, and tut. Three expletives and one contraction appear more frequently in later plays: hold my life, hoyda, pish and o'th'. The totals of instances of early and late forms in full-length plays wholly or almost wholly by Middleton are recorded in Table 4.

The most notable feature of Table 4 is the clear division between the first seven plays, composed up to 1607, and the others, datable from 1611 onwards. While the Act to Restrain Abuses of May 1606 might have inhibited Middleton's partiality for by the mass, 'sfoot, 'slid, and 'slud, pg 95it is unlikely to have influenced the other shifts in usage. The Bloody Banquet, for which the count of early/late forms is 10/1, sits comfortably on the early side of the chronological divide.

VIII

Towards the beginning of this essay I maintained the value of reading plays, poems, entertainments, and prose works within their 'authorial' context. The new Middleton canon and chronology results in interesting conjunctions, as one small illustration will show. Working independently on separate plays and separate canons, modern scholars have generally agreed that six plays attributed to Middleton and Shakespeare were written during the period July 1605–July 1606 (probably in this order): Middleton's A Yorkshire Tragedy, Shakespeare's King Lear, Shakespeare and Middleton's Timon of Athens, Middleton's The Revenger's Tragedy, Middleton's The Puritan, Shakespeare's Macbeth (later adapted by Middleton). The Revenger's Tragedy was published anonymously in 1607/8, A Yorkshire Tragedy was published as Shakespeare's in a quarto of 1608, and The Puritan was published in 1608 with the claim that, though acted by the Children of Paul's, it had been written by 'W.S.' These three plays are now included for the first time in an edition of Middleton's collected works. Recognition that Middleton collaborated with Shakespeare on Timon of Athens is also recent. We cannot be absolutely sure of the order of composition of the six works, but there is no doubt that they belong to a single period of little more than a year, at the outside.

The strong association between Middleton and the King's Men at this time is in itself an item of theatrical history that has for decades been denied by a long succession of commentators on The Revenger's Tragedy. But even more significant are the qualities connecting the plays now gathered together as Middleton's, and as his at a particular time. Consider A Yorkshire Tragedy and The Revenger's Tragedy. Both have repeatedly been described, by critics from Swinburne onwards, in terms of intensity, fieriness, rapidity, ferocity, vividness, and force.7 The action of each has been seen as driven by the manic energy of its chief character. The persons in each play have been said to lack rounded humanity but to belong to a dramatic world stripped to essentials and with its own dynamics. The scholar Oliphant (1929) thought A Yorkshire Tragedy 'written at white heat'; the director Shelton (1965) thought The Revenger's Tragedy 'written in a ferment of excitement' (Scott, 40). It is doubtful whether any other play in the whole of English Renaissance drama could be linked to these two so closely as they have been linked to each other through critical commentary that employs the same adjectives and images. It seems likely that Vindice's savagery in The Revenger's Tragedy owes much to the Husband's paranoid violence in A Yorkshire Tragedy, with its basis in fact. In any case, the qualities shared by the two plays clearly have their origins in a single, and singular, authorial imagination.

Timon of Athens also has its points of contact with The Revenger's Tragedy. As long ago as 1876 F. G. Fleay, who had analysed the metrical features of over two hundred Renaissance plays, reported that results for the 'non-Shakespearean' portions of Timon of Athens most closely agreed with those for The Revenger's Tragedy (195). In 1909 E. H. C. Oliphant, who also considered Timon of Athens to be 'beyond reasonable doubt only partly [Shakespeare's]', proclaimed that 'The tone of III, 1 is that of the Revenger's Tragedy' (195–6). In 1968 two critics independently made similar connections. Philip Edwards remarked that if Timon of Athens had missed publication in the First Folio, 'We should not have known of Shakespeare's power to write satirical merchant comedy in a style which only Middleton can equal'; he remarked that the dunning scenes in Act 3 of Timon of Athens 'have no parallel elsewhere in Shakespeare' (134). Brian Gibbons, finding in Middleton's city comedies a 'tension between the satiric-comic form and the tragic themes of savagery, disease and evil', judged that 'the dramatic manner of Tourneur in The Revenger's Tragedy and of Shakespeare in Act III of Timon of Athens may be indebted to Middleton's art of comedy' (205). It is now clear that most of those parts of Timon of Athens which Fleay regarded as non-Shakespearean, including Act 3, were in fact written by Middleton, who was also the sole author of The Revenger's Tragedy. Even in his collaboration with Shakespeare, Middleton's authorial personality has manifested itself to four commentators whose dates of writing span close on a century, whose critical concerns are diverse in the extreme, and who were unaware that they were dealing with a single playwright. Unwittingly, they have conducted an experiment for us—a kind of blind matching. There could scarcely be a neater demonstration that authors, however circumscribed their autonomy, are agents to be reckoned with.

works cited

  • Adams, Joseph Quincy, ed., The Ghost of Lucrece (1937)

  • Auden, W. H., interviewed by Michael Newman, in Writers at Work: The 'Paris Review' Interviews: Fourth Series, ed. George Plimpton (1976), 243–69

  • pg 96 Baird, Matthew, 'Collaboration of Thomas Dekker and Thomas Middleton'. B.Litt. thesis, University of Oxford, 1928

  • Barker, Richard Hindry, 'The Authorship of The Second Maiden's Tragedy and The Revenger's Tragedy', Shakespeare Association Bulletin 20 (1945), 51–62

  • —— Thomas Middleton (1958)

  • Barroll, Leeds, Politics, Plague, and Shakespeare's Theater: The Stuart Years (1991)

  • Bentley, Gerald Eades, The Profession of Dramatist in Shakespeare's Time 15901642 (1971)

  • Boyle, Robert, 'Beaumont, Fletcher and Massinger', Englische Studien 5 (1882), 74–96

  • Brady, Jennifer, and W. H. Herendeen, eds., Ben Jonson's 1616 Folio (1991)

  • Bristol, Michael, Shakespeare's America, America's Shakespeare (1990)

  • Brooke, Nicholas, Horrid Laughter in Jacobean Tragedy (1979)

  • Burke, Seán, Authorship: From Plato to the Postmodern: A Reader (1992)

  • —— The Death and Return of the Author (1992)

  • Butler, Martin, 'Jonson's Folio and the Politics of Patronage', Criticism 35 (1993), 377–90

  • Chambers, E. K., William Shakespeare: A Study of Facts and Problems, 2 vols (1930)

  • Cook, Elizabeth, ed., The Roaring Girl, New Mermaid (1997)

  • De Grazia, Margreta, Shakespeare Verbatim: The Reproduction of Authenticity and the 1790 Apparatus (1991)

  • Dryden, John, John Dryden, ed. Keith Walker (1987)

  • Dunkel, W. D., 'Did Not Rowley Merely Revise Middleton?', PMLA 58 (1933), 799–805

  • Dutton, Richard, Ben Jonson: Authority: Criticism (1996)

  • Edwards, Philip, Shakespeare and the Confines of Art (1968)

  • Eliot, T. S., Elizabethan Essays (1934)

  • Fleay, Frederick Gard, Shakspeare Manual (1876)

  • Foakes, R. A., and R. T. Rickert, eds., Henslowe's Diary (1961)

  • Gibbons, Brian, Jacobean City Comedy (1968)

  • Greg, W. W., 'Authorship Attributions in the Early Play-lists, 1656–1671', Edinburgh Bibliographical Society Transactions 2 (1946), 305–29

  • Harbage, Alfred, Annals of English Drama 9751700, revised S. Schoenbaum, revised Sylvia Wagonheim (1989)

  • Holdsworth, R. V., 'Middleton and Shakespeare: The Case for Middleton's Hand in Timon of Athens', PhD thesis, University of Manchester, 1982

  • —— ed., Three Jacobean Revenge Tragedies (1990)

  • Holmes, David M., The Art of Thomas Middleton (1970)

  • Hope, Jonathan, The Authorship of Shakespeare's Plays (1994)

  • Howard-Hill, T. H., ed., A Game at Chess, Revels (1993)

  • Hoy, Cyrus, 'The Shares of Fletcher and his Collaborators in the Beaumont and Fletcher Canon', Studies in Bibliography 8 (1956), 129–46; 9 (1957), 143–62; 11 (1958), 85–106; 12 (1959), 91–116; 13 (1960), 77–108; 14 (1961), 45–67; 15 (1962), 71–90

  • Jackson, MacDonald P., Studies in Attribution: Middleton and Shakespeare (1979)

  • —— ed., The Revenger's Tragedy, Attributed to Thomas Middleton: A Facsimile of the 1607/8 Quarto (1983)

  • —— 'Editing, Attribution Studies, and Literature Online: A New Resource for Research in Renaissance Drama', Research Opportunities in Renaissance Drama 37 (1998), 1–15

  • —— Defining Shakespeare: 'Pericles' as Test Case (2003)

  • Knight, Charles, ed., The Pictorial Edition of the Works of Shakspere. Second edn. revised, Doubtful Plays. New York: Routledge, n.d.

  • Lake, David J., The Canon of Thomas Middleton's Plays (1975)

  • Masten, Jeffrey A., 'Beaumont and/or Fletcher: Collaboration and the Interpretation of Renaissance Drama', ELH 59 (1992), 337–56

  • —— Textual Intercourse: Collaboration, Authorship, and Sexualities in Renaissance Drama (1997)

  • Mincoff, Marco, 'The Authorship of The Revenger's Tragedy', Studia Historico-Philologica Serdicensia 2 (1940), 1–87

  • Murray, Peter B., A Study of Cyril Tourneur (1964)

  • Oliphant, E. H. C., 'Shakspere's Plays: An Examination', Modern Language Review 3 (1908), 337–55; 4 (1909), 190–9, 342–51

  • —— 'Problems of Authorship in Elizabethan Dramatic Literature', Modern Philology 8 (1911), 411–59

  • —— 'A Dekker—Middleton Play, The Bloodie Banquet', Times Literary Supplement, 17 December 1925, 882

  • —— 'The Authorship of The Revenger's Tragedy', Studies in Philology 23 (1926), 157–68

  • —— The Plays of Beaumont and Fletcher: An Attempt to Determine Their Respective Shares and the Shares of Others (1927)

  • —— ed., Shakespeare and His Fellow Dramatists (1929)

  • Oras, Ants, Pause Patterns in Elizabethan and Jacobean Drama: An experiment in Prosody (1960)

  • Ricks, Christopher, 'The Moral and Poetic Structure of The Changeling', Essays in Criticism 10 (1960), 290–306

  • —— 'Word-Play in Women Beware Women', Review of English Studies NS 12 (1961), 238–50

  • Schoenbaum, S., Middleton's Tragedies: A Critical Study (1955)

  • —— Internal Evidence and Elizabethan Dramatic Authorship (1966)

  • Scott, Michael, Renaissance Drama and a Modern Audience (1982)

  • Smith, M. W. A., 'The Authorship of The Revenger's Tragedy', Notes and Queries 236 (1991), 508–13

  • —— 'The Authorship of Timon of Athens', TEXT 5 (1991), 195–240

  • Stenger, Harold L., 'The Second Maiden's Tragedy: A Modernized Edition with an Introduction'. PhD dissertation. University of Pennsylvania, 1954

  • Swinburne, A. C., The Age of Shakespeare (1908)

  • —— A Study of Shakespeare. London: Heinemann, 1918; Golden Pine Edition; first printed Chatto & Windus, 1879

  • Sykes, H. Dugdale, Sidelights on Shakespeare (1919)

  • —— Sidelights on Elizabethan Drama (1924)

  • —— 'Thomas Middleton's Early Non-Dramatic Works', Notes and Queries 148 (20 June 1925), 435–8

  • Taylor, Gary, 'The Canon and Chronology of Shakespeare's Plays', in Stanley Wells and Gary Taylor (with John Jowett and William Montgomery), William Shakespeare: A Textual Companion (1987), 69–144

  • —— 'Shakespeare and Others: The Authorship of Henry the Sixth, Part One', Medieval and Renaissance Drama in England 7 (1995), 145–205

  • —— 'Thomas Middleton, Thomas Dekker, and The Bloody Banquet', Papers of the Bibliographical Society of America 94 (2000), 197–233

  • —— 'Middleton and Rowley—and Heywood: The Old Law and New Attribution Technologies', Papers of the Bibliographical Society of America 96 (2002), 165–217

  • —— 'Thomas Middleton, The Spanish Gypsy, and Collaborative Authorship', in Words That Count, ed. Brian Boyd (2004), 241–73

  • Taylor, Gary, and John Jowett, Shakespeare Reshaped, 16061623 (1993)

  • Taylor, Gary, Paul Mulholland, and MacDonald P. Jackson, 'Thomas Middleton, Lording Barry, and The Family of Love', Papers of the Bibliographical Society of America 93 (1999), 213–41

  • pg 97 Vickers, Brian, Appropriating Shakespeare: Contemporary Critical Quarrels (1993)

  • —— Shakespeare, Co-Author: A Historical Study of Five Collaborative Plays (2002)

  • Watson, Robert N., 'Tragedy', in The Cambridge Companion to English Renaissance Drama, ed. A. R. Braunmuller and Michael Hattaway (1990), 301–51

  • Wiggin, Pauline G., An Inquiry into the Authorship of the MiddletonRowley Plays (1897)

Notes

1 Recent discussions of the concept of 'authorship' are too numerous to specify. Sane guidance is provided by Seán Burke, The Death and Return of the Author and Authorship: From Plato to the Postmodern: A Reader. See also Brian Vickers, Appropriating Shakespeare, 101–15.

2 Jackson attempted a preliminary analysis of the rates of some high-frequency 'function words'—and, it, of, that, to, and the like—which form the framework upon which any piece of verse or prose is constructed. The findings supported the ascription to Middleton of The Revenger's Tragedy, A Yorkshire Tragedy, The Puritan, and The Lady's Tragedy and the rejection from the Middleton canon of Blurt, Master Constable and, less emphatically, The Family of Love. Smith has since developed sound statistical techniques for processing and evaluating such data, applying them to The Revenger's Tragedy and Timon of Athens: Smith's results support Middleton's claim to Revenger, but, while consistent with dual authorship of Timon, raise doubts over the nature and extent of Middleton's participation in that play. However, it should be said that Smith's work, although mathematically more sophisticated, is based upon smaller databases than the work of Jackson, Lake, and Holdsworth.

3 This is the arithmetic mean; the median is 15, and the mode (the most frequently occurring total) is as low as 7.

4 The Nice Valour's two examples are matched or surpassed by only ten per cent of the non-Middleton plays and are similar to totals for plays of Middleton's later period.

5 There are minor questions associated with Civitatis Amor and The Sun in Aries. In Civitatis, Middleton's signature occurs at the end of 'The Entertainment at Whitehall', and the following section on 'Prince Charles His Creation' is probably not his. Munday has often been claimed as co-author of Aries.

6 The fullest analyses of Middleton's style in dramatic poetry are those by Stenger (124–87) and Mincoff. Oliphant offers excellent thumbnail sketches in Shakespeare and his Fellow Dramatists, I, 10–12, 21, and The Plays of Beaumont and Fletcher, 83–5.

7 Holdsworth, Three Jacobean Revenge Tragedies, reprints a range of reactions to The Revenger's Tragedy. John Churton Collins called it a 'bleared, rapid and uneven work … fierce and vivid … brilliant and powerful', and wrote of the 'appalling and unrelieved intensity' of Vindice's character, his 'savage and devilish energy' (Holdsworth, Revenge Tragedies, 29). Swinburne, The Age of Shakespeare, had alluded to the 'fiery jet' of the play's 'molten verse' and to its sustained intensity (277). Oliphant used such phrases as 'dramatic intensity' and 'glowing fury' (1929, II, 94). Eliot applied the adjectives 'fierce' and 'intense', alluded to 'explosiveness', 'rapidity', and 'top speed', and judged that The Atheist's Tragedy showed a few 'flashes of the old fire' that had ignited the earlier play, which, like Collins, he also attributed to Tourneur (121–30). Schoenbaum, Middleton's Tragedies, also employed the 'flame' figure. Charles Knight described A Yorkshire Tragedy in terms of 'ferocity … intensity' (254). John Addington Symonds, Shakespeare's Predecessors, likened its effect to a 'flash of lightning' by which we are 'seared and blinded' (435). Swinburne strung together 'fierce … hot … breathless heat … raging rate of speed … fire' (1918, 142). Oliphant (1929) called it 'tense', 'gripping', 'forceful', 'vivid', 'breathless', and 'fiercely naturalistic', thought it 'written at white heat', and remarked on the 'insane fury' of the chief character (II, 25–9).

logo-footer Copyright © 2018. All rights reserved.
Access is brought to you by Log out