Shakespeare’s Pauses, Authorship, and Early Chronology Douglas Bruster*1 Abstract. This paper explores the implications of Ants Oras’s Pause Patterns in Elizabe- than and Jacobean Drama: An Experiment in Prosody (Oras 1960) for the chronology and authorship of plays in early modern England. Oras’s brief monograph has been noticed by a relatively few scholars, mainly those interested in changes to Shake- speare’s pentameter line. Recent developments in the field, however, have rendered his data newly attractive. Compiled by hand, Oras’s figures on the punctuated pauses in pentameter verse offer computational approaches a wealth of information by which writers’ stylistic profiles and changes can be measured. Oras’s data for a large number of playwrights and poets, as well as his methodology generally, may prove instrumen- tal in constructing a portrait of the aesthetic environment for writers of pentameter verse during the sixteenth and seventeenth centuries in England. In particular, pause percentages may lend context to our attributions of texts of uncertain authorship. A hypothetical chronology is offered for Shakespeare’s earliest writing, including his contributions to Arden of Faversham, 1 Henry VI, and Edward III. Keywords: William Shakespeare, attribution, authorship, Ants Oras, prosody, met- rics, pause patterns, caesura, iambic pentameter, chronology, Arden of Faversham, 1 Henry VI, Edward III, 2 Henry VI, 3 Henry VI, Marina Tarlinskaja 1. Oras and Pause Patterns In 1960, Ants Oras published a short book in the University of Florida Monographs series (‘Humanities’) titled Pause Patterns in Elizabethan and Jacobean Drama: An Experiment in Prosody. Compiled by hand and largely forgotten by the field, Oras’s data has become newly useful owing to the rise of the digital humanities and a related interest in the statistical analysis of liter- ary style. This paper examines Oras’s pause-pattern research, and considers various possible objections to his procedures. It is my argument that Oras’s findings, as well as his methodology, offer a rewarding way of understanding the structure and development of iambic pentameter in Shakespeare and in * Author’s address: Douglas Bruster, Department of English, The University of Texas at Austin, 208 W. 21st St. Austin, Texas 78712. USA, email: bruster@austin.utexas.edu. Studia Metrica et Poetica 2.2, 2015, 25–47 doi: dx.doi.org/10.12697/smp.2015.2.2.03 http://dx.doi.org/10.12697/smp.2015.2.2.03 26 Douglas Bruster early modern drama generally. Further, Oras’s procedure can provide valuable information regarding both chronology and authorship. At only ninety pages, Oras’s book has been overlooked by many but has influenced others in a manner that belies its size. Employed in attribution studies by such scholars as Brian Vickers (Vickers 2002) and MacDonald P. Jackson (Jackson 2002), Oras’s research has also been cited approvingly in the Oxford Middleton (Taylor and Lavagnino 2007: 93–94), The Quest for Cardenio (Jackson 2012: 137–138) and, perhaps most consequentially, the influential ‘Canon and Chronology’ section of Oxford’s Textual Companion (Taylor 1987: 107–108), where it has something like pride of place among the metrical tests consulted. The title of Oras’s study is slightly misleading, as he surveys continen- tal and English writers, from Machaut and Marot through Massinger, who produced decasyllabic verse over a span of nearly three centuries. He offers discrete figures for various Canterbury Tales, the books and cantos of Spenser’s Faerie Queene, the differing versions of Jonson’s Every Man plays, and sub- divides such collaborative plays as Henry VIII, The Two Noble Kinsmen, and Pericles. His aim is to demonstrate historical and personal patterns in iambic pentameter by tabulating where punctuated pauses fall within its first nine syllables (punctuation after the tenth syllable is not counted). With almost two-thirds of its pages devoted to figures and graphs, Pause Patterns can be said to emphasise evidence over interpretation, although Oras is insightful in his concise commentary. Pauses can be counted in three different ways in Oras’s tabulation; he labels these A, B, and C pauses. A pauses are those signaled by punctuation of any kind within a pentameter line (Oras counts short lines, but not their terminal punctuation). B pauses, a sub-group of the A pause, are so-called ‘strong’ pauses within the line: pauses signaled by any punctuation mark other than a comma, including periods, question marks, colons, semi-colons, and dashes. C pauses are comprised of punctuation marks dividing ‘split-’ or ‘shared’ lines. Almost by definition, C pauses are a sub-group of the B pause: the punctuation dividing shared lines is invariably ‘strong’ in nature, rather than a comma. To summarise, then; in Oras’s methodology, A pauses are all pauses, B pauses strong pauses, and C pauses those that divide a line or lines amongst two or more speakers. All B and C pauses are also A pauses, with C pauses represent- ing the smallest (because most heavily specified) of the groups. Oras counted A, B, and C pauses for thirty-eight Shakespeare plays, adding the A and B figures as well for Venus and Adonis, The Rape of Lucrece, and the Sonnets. He also tabulated A, B, and C pauses for approximately fifty non- Shakespearean plays from the era, as well as major poems by these and other 27Shakespeare’s Pauses, Authorship, and Early Chronology authors. Additionally, he provides figures for C pauses (though only C pauses) for a large number of seventeenth-century playwrights, from Chapman and Heywood through Shirley and Davenant. All of this data is presented in the appendix to Pause Patterns (Oras 1960: 61–88). Assigning each title a row in his appendix’s tables, for example, Oras records the number of pauses in each position, followed by the total for each work. At the far right of the landscape-oriented page, he calculates the percentages for each pause position, rounding to one place after the decimal. Between these strings of figures, Oras also supplies an entry for ‘First Half ’ percentages (that is, the ratio of pauses falling in the first-through-fourth positions to those in the whole line minus the ‘middle’ fifth pause) and an entry for ‘Even’ percentages (the percent of pauses falling on even-numbered positions – that is, second, fourth, sixth, eighth). Finally, Oras provides line graphs (Oras 1960: 33–60) in which each title’s relevant pauses are represented as (in most cases) a slender but steep set of peaks and valleys. The mechanical nature of such scholarship – at once its strength and weakness – has been noted by at least one observer. As Vincent Leitch has described Oras’s work in another study, an inquiry into the manifestation of sound patterns in the poetry of Spenser and Milton, “It is rather like the study of subatomic particles, turning up new entities like neutrinos and quarks” (Leitch 2008: 19). Oras himself analogised his pause-pattern results to the output of a piece of medical technology that had come into general use dur- ing his lifetime: “the total patterns are likely to reveal much over which the person concerned has little or no control, almost as people are unable to control their cardiograms” (Oras 1960: 2). Extending Oras’s metaphor, we might say that his Pause Patterns monograph seeks to graph the heartbeat of iambic pentameter as it was manifested by a variety of texts written across a considerable number of years. Just as any individual’s electrocardiographic results are likely to change over time (and, as Oras posits, not be under an individual’s control), so do pause patterns develop in appreciable ways. By gathering this information, Oras put himself and others in a position to trace verse’s meaningful patterns – and, by extension, the formal careers of poets, genres, and cultures. 2. Shakespeare’s Pauses over Time Oras’s compilation of figures for Shakespeare’s work confirms something that the playwright’s readers have traditionally recognised: Shakespeare’s lines 28 Douglas Bruster become longer as his career unfolds. We could take as representative instances three lines from, respectively, the early, middle, and later parts of his career: Thou art a fool; if Echo were as fleet, (Taming of the Shrew, Ind.1.26) To be, or not to be, that is the question. (Hamlet, 3.1.55) Unto my end of stealing them. But, gracious sir, (Cymbeline, 5.5.347) 1 The lines here grow, syllabically, from an even pentameter in the early comedy to an 11-syllable line in Hamlet and a 12-syllable line in his late romance. The syntactical pause that seems built into the structure of substantial verse lines shifts, accordingly and progressively, later in each line. In these examples, as Shakespeare’s lines lengthen, the units that introduce each line (‘Thou art a fool,’ etc.) grow in both length and complexity: Thou art a fool if Echo were as fleet, To be, or not to be that is the question. Unto my end of stealing them But, gracious sir, These lines ‘break’ at increasing distances; an early play is more likely to feature a pause after the fourth syllable while a later play is more likely to have a pause in the second half of the line. In line graphs of the type that Oras constructed (Oras 1960: 33–60), pause percentages for The Taming of the Shrew, Hamlet, and Cymbeline would appear as in charts 1–3. Even a cursory examination reveals that, as represented in these plays, the dominant pause position shifts, from early to late in Shakespeare’s career, from after the fourth syllable to after the sixth. During the 1590s, most of the pauses in Shakespeare’s plays come in the first half of his line – that is, before the fifth syllable. After 1600, the majority of the pauses shift to the second half of his pentameter line. Consequently, one could nearly ‘flip’ the first chart, left to right, and produce something more at home after 1600 than before. 1 As reproduced above, the lines follow Evans, ed., in the Riverside text (Shakespeare 1997). In the First Folio of 1623 (Shakespeare 1968), the Shrew punctuation is a comma, the Hamlet marking is identical to that above, and the Cymbeline punctuation similar save that the Folio has no comma after ‘But’. Later in this essay I discuss the relation of compositors to punctuation. 29Shakespeare’s Pauses, Authorship, and Early Chronology   Charts 1–3. Pause patterns, as percentages, in three Shakespeare plays However familiar, this story is also somewhat circular. While I have selected the three lines and three plays above as representative of ‘early’, ‘middle’, and ‘later’ Shakespeare, and used them to narrate his development as a writer, our chronology of his works  – our very sense of early, middle, and later Shakespeare – is not independent of line length and prosody. To the contrary, we have long relied on these features to date his plays in the first place. This potential tautology traces to an idea articulated in a monograph from the Victorian era. Remarks on the Differences in Shakespeare’s Versification in Different Periods of his Life and on the Like Points of Differences in Poetry Generally was published anonymously in 1857, but has since been identified as the work of Charles Bathurst. There he observes: It must have been remarked by most readers of Shakespeare who are not very unobserving, that his versification, in respect of the cæsura, as it is called, or division of the pauses, differs most exceedingly in different places. This differ- ence is not as between one passage and another, or one scene and another, but generally, and in its extremes always, as between one play and another; and it depends on the time of his life (Bathurst 1857: 1). Echoing his title, Bathurst correlates chronology and caesura, with Shakespeare’s pauses revealing a veritable law of his verse: “in metre,” he continues, “Shakespeare changed very nearly regularly and gradually, always in the same direction” (Bathurst 1857: 6). More nuanced than such a law suggests, the story of this change is also much larger than Shakespeare. Readers who make their way through the pages of any 30 Douglas Bruster anthology of Renaissance drama feel the differences in pentameter verse as the even beats and ten-syllable lines of the late 1580s begin to give way to something less predictable early in the next century. Whether referring to this transforma- tion as an ‘unscrewing’ (Saintsbury 1906, 2: 302) or an ‘evolution’ (Tarlinskaja 1987: 44), scholars agree that iambic pentameter became less iambic, and less clearly pentameter, in a process that commenced within a decade after its estab- lishment as the dominant vehicle for plays in London’s commercial theaters.2 Told many times and from various angles, the narrative of blank verse’s trans- formation depends in part on the sensitive readings of individual phrases, lines, and speeches by such scholars as Marina Tarlinskaja, George T. Wright, Coburn Freer, and Russ McDonald, to name only these. Also crucial, however, has been evidence of a more extensive nature: statistical information relating to meter and vocabulary and typically surviving in tables that grace handbooks and the appen- dices of collected works. In its nearly 100 pages of data, for example, Tarlinskaja’s detailed study of dramatic verse in the playhouses of early modern England (Tarlinskaja 2014: 287–375) provides a compendium of this kind of informa- tion. Oras’s Pause Patterns study anticipates such ‘panoramic’ (Spevack 1976) scholarship with its broad examination of the punctuated pauses of early verse. Using Oras’s figures, and data like that which he provided, we can better place works of early modern drama in chronological order. A play from the early 1590s, for example, is likely to possess a different pause profile than a play written a decade or more later. One can see a stark example of this in two plays based on the same content: the anonymous True Chronicle History of King Leir, thought to have been written in the late 1580s or early 1590s, and Shakespeare’s King Lear, commonly dated to 1605–1606. To compare their pause patterns, as in charts 4–5, is to see a difference in stylistic eras. If we lacked any other information about the two plays, these graphs would suggest that King Leir is more likely to have come from earlier in the play- houses’ tenure: as we will see, its profile most resembles Kyd’s composite in its steady decline from the fourth position. With its higher sixth-position pauses, in contrast, King Lear looks like the Jacobean text we know it to be. 2 Saintsbury characterised Beaumont and Fletcher as “the first noteworthy examples of that ‘unscrewing’ of dramatic blank verse which led, before long, to the break-up of its whole struc- ture as a dramatic medium” (Saintsbury 1906, 2: 302). Shortly thereafter Saintsbury noted that Shakespeare “in his own later plays eased the screws very freely, and rather hazardously in appearance” (Saintsbury 1906, 2: 303). Tarlinskaja observes that ‘the dramatic iambic penta meter of the Elizabethan-Jacobean epoch evolved from a more rigid to a looser form’ (Tarlinskaja 1987: 44), returning to the metaphor of evolution on 54, 82, and 218. Tarlinskaja’s 2014 study represents the state of the art on dramatic pentameter during the early modern period in England. 31Shakespeare’s Pauses, Authorship, and Early Chronology   Charts 4–5. Pauses in King Leir and King Lear Oras did not offer a chronology on the basis of his research, even though his figures suggested obvious modifications to the Chambers timeline. A number of options existed, including sorting by the ratio of ‘first half ’ pauses, the ‘even’ pauses, or some arrangement of his A, B, and C pauses. Had he made such an attempt, he would have needed to address the discrepancies between his results and the chronology he started with (wherein Titus Andronicus and The Comedy of Errors are the first two plays). It is worth noting, too, that a simple arrangement of Oras’s figures would lead to some counter-intuitive place- ments: 2 Henry IV would come before 1 Henry IV, for example, in an ordering based on both ‘first half ’ and ‘even’ pauses;3 Antony and Cleopatra would come apparently too late, and The Tempest earlier than we think it was written. Ideally, then, a chronology based on Oras’s figures would need to address whatever in them led to such counter-intuitive placements. It would also need to incorporate the most recent state of knowledge concerning Shakespeare’s collaboration with others, employing Oras’s methodology on the parts of texts that scholars believe Shakespeare wrote. Finally, such a chronology would be buttressed by a statistical test or tests to analyse both the interrelation of the pause data and the movement of Shakespeare’s style over time.4 3 As Henri Suhamy notes: “Il est inutile d’autre part de souligner longuement que les preuves internes que contient le texte de Shakespeare ne permettent pas de dater les œuvres avec une précision rigoureuse. Si l’on se fiait aveuglément aux chiffres on en arriverait à affirmer que par exemple la seconde partie de Henry IV a pu être écrite avant la première” (377). 4 For a chronological study of this kind, see Bruster and Smith 2014. 32 Douglas Bruster 3. Pause Patterns and Authorship Oras’s study shows us that although the rhythmical environment of a liter- ary marketplace changes over time, differences among writers often obtained regardless of when they wrote their works. An unremarkable observation on its face, this nonetheless has important implications for understanding authorship and even for attributing shares of various collaborative texts. How different were the practices of various playwrights? The following charts represent the pause-pattern percentages of single plays by Christopher Marlowe, Thomas Kyd, and Robert Greene, respectively. All three playwrights wrote for the pro- fessional playhouses during the late 1580s and early 1590s.   Charts 6–8. Pause patterns in plays by Marlowe, Kyd, and Greene Even with this chronological proximity, one can note differences among the syntactic patterns of their lines. No play of Marlowe’s, for instance, features a fourth position percentage higher than Kyd’s lowest figure (from his transla- tion, Cornelia); likewise, The Spanish Tragedy’s figure of 39.1% for its fourth position is higher than only one of Greene’s plays – Friar Bacon and Friar Bungay, at 38.5%. Indeed, charts of Greene’s pause patterns reveal an extremely high ‘peak’, betraying a regular punctuated pause after his fourth syllables, and, not surprisingly a high percentage of pauses after even syllables. High figures for even-syllable pauses are the sign of a thoroughly iambic poet. Greene’s composite for even-syllabled pauses is in the 70s. Kyd’s percentage is in the 60s (as is Marston’s), Marlowe’s in the upper 50s, and Shakespeare’s – owing to the length of his career – from the upper 60s in Titus Andronicus through the upper 50s in such late plays as Henry VIII and The Two Noble Kinsmen. Jonson 33Shakespeare’s Pauses, Authorship, and Early Chronology is the least concerned with metrical fluidity; unsurprisingly, his percentages range from the mid- to lower 50s (Oras 1960: 68). While far from a fingerprint, in certain instances of synchronic collabora- tion – texts written virtually simultaneously by two or more hands – pause patterns may help to indicate which parts of a text came from which author. To give one example: scholars have long considered Pericles, a play published in 1608, as partly by Shakespeare, and partly by George Wilkins (see Jackson 2003). Usually acts 1 and 2 are given to Wilkins, and acts 3–5 to Shakespeare. The following charts set out the pauses, from Oras’s count, for this traditional division of Pericles. Charts 9–10. Pause patterns of two Pericles dramatists Writing at approximately the same time, Wilkins and Shakespeare nonethe- less display markedly divergent syntactical profiles. Wilkins, born in 1576, was Shakespeare’s junior by 12 years; significantly, he was not as experienced a playwright. Perhaps unsurprisingly, then, the pause profile of his iambic pentameter in Pericles is closer to something that Shakespeare would have done almost a decade earlier in his career, with Hamlet (see chart 2, above). If the Wilkins share of Pericles looks backward, stylistically, toward the late 1590s, Shakespeare’s leans forward into the emerging, more flexible style of such contemporaries as Thomas Middleton and John Fletcher. Pauses percentages can enrich our conversations about attribution even when they do not resolve specific questions. The chart below, for example, sets out the pause profiles of sections of the first part of Henry VI (1H6) that have been ascribed to various playwrights by Gary Taylor (Taylor 1995). 34 Douglas Bruster   Chart 11. Hypothesised shares of 1 Henry VI (after Taylor 1995) Taylor attributes the ‘X’ portion to Shakespeare and the ‘Z’ to Thomas Nashe, who – Taylor suggests – wrote the first act of this history play. Although these two sections feature a visible uptick after the sixth syllable, the ‘X’ section (Shakespeare’s, in the Taylor attribution) does not feature the highest (which belongs to author ‘Y’) or even second highest ‘even’ count (author ‘Z’, Nashe). Nevertheless, a pause analysis beginning with Taylor’s attributions would lin- ger over the sixth-position differences between, on one hand, X and Z and, on the other, W and Y. While such variation is not unprecedented in the data that Oras gathered, it nonetheless seems especially meaningful given Taylor’s identification of other patterns of difference amongst these sections. If we believe that 1H6 was written at approximately a single moment – rather than as a process, over a number of years – the discrepancies in these pause patterns might be attributable to multiple authorship. The longstanding arguments concerning collaborative authorship in 1 Henry VI have recently been joined by similar observations about 2 Henry VI and 3 Henry VI. Hugh Craig (Craig 2009: 40–77) has argued that vocabulary – in particular, divergences in the percentages of function words and lexical words – indicates the likelihood of multiple hands in 2 Henry VI. Together with John Burrows, Craig has explored a similar divergence of vocabulary in 3 Henry VI (Craig and Burrows 2012). Without subscribing to Craig’s identi- fication of Marlowe as the 2 Henry VI coauthor, Tarlinskaja confirms Craig’s sense of linguistic variety in that play, as well as in 3 Henry VI (Tarlinskaja 2014: 112–116). What might we learn about these plays, and the possibility 35Shakespeare’s Pauses, Authorship, and Early Chronology of multiple authorship, by looking at their pause patterns? Particularly if we divide them along the lines that Craig, along with Burrows, has offered? The following charts, charts 12 and 13, present the pause patterns of 2 Henry VI and 3 Henry VI. Charts 12–13. Hypothesised shares of 2 Henry VI (after Craig 2009) and 3 Henry VI (after Craig and Burrows 2012). While the uncontestably Shakespearean sections have a (just) slightly more pronounced development following the fifth and sixth syllables, both charts present pause figures so closely aligned that it is difficult, on the basis of pause data above, to ascribe their respective sections to multiple authors. Such leaves us with various possibilities, including but not limited to the following: (1) both plays are the product of a single hand; (2) any collaborators had the same pause patterns as Shakespeare; or (3) the pause patterns here are the product not of the author(s) but of scribe(s) or compositor(s) who homogenised the texts. Given the evidence, the first of these possibilities seems worth strong consideration. 4. Pauses and Early Shakespeare Pause patterns may shed light on two other plays from early in Shakespeare’s career, each of disputed authorship. The following charts represent the pause patterns in Arden of Faversham and Edward III, respectively. Each of these plays 36 Douglas Bruster has been attributed, at least in part, to Shakespeare. For Arden of Faversham, I have used Macdonald P. Jackson’s suggestions in Determining the Shakespeare Canon (Jackson 2014) to divide the play into ‘Shakespeare’ and ‘non-Shake- speare’ scenes; for Edward III I have employed the division suggested by the research of Ward Elliott and Robert Valenza (Elliott, Valenza 2010). Charts 14–15. Hypothesised shares of Arden of Faversham (after Jackson 2014) and Edward III (after Elliott and Valenza) In certain aspects, the pause patterns that these graphs represent may diverge more than one would expect from chance. For instance, the Arden chart, chart 14, reveals a higher fourth-position figure in the ostensibly Shakespearean sec- tion, and the Edward III chart, chart 15, a lower, less-developed sixth-position figure for the ostensibly non-Shakespearean section. But, taking into account the real divergences we have seen in the various shares of 1 Henry VI and Pericles, can we say, on the basis of the data above, that either Arden or Edward III is likely to have been written by multiple playwrights? Perhaps both plays were written by the same author, and some other factor is responsible for the slight if perceptible differences in pause patterns visible in the charts here. It is possible for instance, that these differences trace to diachronic composi- tion – to having been composed by the same hand over an interval during which, for whatever reason or reasons, the writer’s style developed in appre- ciable ways. It is also possible that these differences are owing to the nature of the materials used for printing: judging from the Sir Thomas More pages, Shakespeare’s manuscripts (to name only these) were probably punctuated very differently than a professional scribe’s. (In this regard, it seems poten- tially significant that the Shakespearean segment of Arden features a smaller 37Shakespeare’s Pauses, Authorship, and Early Chronology proportion of punctuated pauses, per verse line, than the non-Shakespearean section – what we might expect if copy for that segment of the quarto were closer to his pen.) Whatever the reason, we could note that, in relation to the more iambic second-half syntax we have seen in various playwrights, the ostensibly non-Shakespearean material of Edward III is the least developed of the four sections examined. It is worth pointing out that Eliot Slater believed Edward III to feature distinct parts (‘A’ and ‘B’) which nonetheless both trace to Shakespeare; Slater argued that much of what is defined above as non- Shakespearean writing (part ‘B’) was composed by Shakespeare earlier in his career (Slater 1988: 124–125, 134–135). The possibility of Shakespeare generating same-text material at various times may confirm and qualify a suggestion that Tarlinskaja makes concern- ing various sections in Arden of Faversham, 2 Henry VI, and 3 Henry VI. Noting similar prosodic textures in the ostensibly non-Shakespearean sections of these works, she offers that the dramatist – whom she calls ‘Y’ – “seems to be a poet of an older generation, and [...] is the collaborator in Arden, and 2, 3 Henry VI” (Tarlinskaja 2014: 116). Could the pause patterns in these texts indicate that author ‘Y’ is Shakespeare himself, writing at an earlier phase of his development as a stylist? Before we consider this possibility, it seems right to acknowledge other playwrights who could be considered as potential authors of this material. Pause-pattern evidence for the few texts we have of undisputed authorship suggests that Kyd is unlikely to have written any of the Arden material: his fifth-position percentages are always higher than his sixth, producing a kind of ski-slope pattern to graphs of his verse. Nothing we believe that Kyd wrote has the fifth-position notch or the sixth-position uptick. Marlowe remains a possibility, at least in terms of pause patterns; the graph for the non-Shakespearean sections of Arden is not unlike his profile in Edward II and The Massacre at Paris. If Tarlinskaja’s author ‘Y’ is indeed Shakespeare, such would imply that the pause profiles of Shakespeare’s pentameter lines went through at least one stage before reaching the familiar sawtooth formation we see in the great majority of his verse. It is necessary to say ‘great majority’ here because the graphs of A Midsummer Night’s Dream and King John (Oras 1960: 46) have an idiosyncratic shape; they demonstrate a transitional phase during which Shakespeare slid his pauses across the middle of the line, through the 5th position and toward the 6th. With an eye toward such variation, and in an entirely speculative vein, the three charts below group the pause profiles of some Shakespearean and disputed texts, and parts of these texts, along with dates that would imply an even development. All these profiles have been featured earlier in this essay. The chart on the left, chart 16, features pause profiles for the ostensibly 38 Douglas Bruster non-Shakespearean section of Edward III, and Taylor’s ‘W’ and ‘Y’ sections for 1 Henry VI (Taylor 1995). Chart 17, in the middle, has profiles for 1 Henry VI’s author ‘X’, The Taming of the Shrew, the Shakespearean section of Edward III, and 3 Henry VI (complete). The chart on the right, chart 18, displays data for both parts of Arden (the Shakespearean and ostensibly non-Shakespearean) as well as the undisputed section of 2 Henry VI. Charts 16–18. Three phases in Shakespeare’s development? This three-stage arrangement hinges, obviously, on the appearance of the notch at the fifth position and the related rise of pauses after the sixth. The sequence imagined here – and it must be stressed, again, that this illustration is purely conjectural – implies a poet developing longer and longer phrases in his pen- tameter line, so that ultimately (as in his career generally) he is pausing further into his sentences – from ‘Thou art a fool’ (fourth) to ‘To be, or not to be’ (sixth as well as second). While it is a difference of only a few percentage points across a text, the long duration of Shakespeare’s career allows us to see the kind of development imagined here as both measurable and chronologically significant. If these charts bear any relation to historical reality, Shakespeare’s first writing may have included the ostensibly non-Shakespearean portion of Edward III and perhaps the ‘W’ and ‘Y’ sections of 1 Henry VI (chart 16); he may have next written Shrew, the ‘X’ section of 1 Henry VI, the Shakespearean portion of Edward III, and 3 Henry VI (chart 17, where the texts mentioned last are responsible for the two lines flagging a fifth position notch); Titus Andronicus may have followed (see chart 19, below); and then he may have written (in whatever sequence before the spring of 1592) both parts of Arden and revised the text of 2 Henry VI (which had almost certainly existed, in some 39Shakespeare’s Pauses, Authorship, and Early Chronology form, prior to the 3 Henry VI material) – thus lending the unrevised portion a comparatively earlier pause profile as well as, perhaps, divergent vocabulary. While conjecture, and dependent on the idea of an even stylistic development to Shakespeare’s practice, the sequence sketched here would go some ways toward accounting for these otherwise puzzling materials. It is thus offered as a possible narrative for further investigation. 5. Collaborative Plays Either accommodation or divergence in the style of collaborators could qualify the value of pause data. Is it possible that the mere fact of writing together somehow blends two writers’ styles? Or perhaps that, when prepared for print, various contributions were made more uniform by a scribe or scribes, com- positor or compositors? To address this question, the series of charts below traces, first, the differences between Shakespeare and his collaborators in two plays – Titus Andronicus and Timon of Athens – then considers whether the shares accorded these collaborators in the Shakespeare plays square with their practice in two plays of their own composition.   Charts 19–20. Peele and Shakespeare in Titus Andronicus and Middleton and Shakespeare in Timon of Athens (after Jowett) These charts represent the pause patterns in two collaborative plays of Shakespeare according to our best sense of how authorship was divided. What is interesting is how close the Peele section of Titus seems to be to Shakespeare’s practice. We could also note the general similarity between Middleton’s section 40 Douglas Bruster of Timon and Shakespeare’s. Although the Timon graph presents a greater divergence than the Titus figures, a comparison of Peele’s and Middleton’s style in those texts with their style in sole-authored plays suggests that, on the basis of pause profiles, our Timon attribution may be more questionable than that for Titus. Charts 21–22. Peele in Titus Andronicus, The Arraignment of Paris, and The Battle of Alcazar; Middleton in Timon of Athens and The Lady’s Tragedy. Chart 21 implies that the Peele profile in Titus (1590–1591?) is a decent fit for his practice in The Arraignment of Paris (early 1580s?) and The Battle of Alcazar (1589?). In contrast, the Middleton chart, chart 22, shows significant diver- gence between the profile from the Timon-attributed writing and that from The Lady’s Tragedy (also known as The Second Maiden’s Tragedy, from around 1611). Middleton’s syntactical pattern in the latter is closer to Fletcher than to Shakespeare, and asks us to revisit the broad similarity in chart 20, which sets out the Shakespeare and Middleton parts of Timon (1606?) side-by-side. The similarity we see in chart 20 is not there in chart 22. In fact, the patterns in chart 22 diverge substantially. How to account for this? At least four pos- sibilities advance themselves: (1) Middleton’s style is inconsistent, or changed considerably between 1606 and 1611; (2) his style accommodated itself to Shakespeare’s as they collaborated; (3) a scribe or compositor, or multiple agents of this kind, were responsible for imposing a uniformity of punctu- ated pauses in Timon; or (4) our current attributions overestimate Middleton’s share of Timon of Athens. Without more information concerning Middleton’s prosody, it is difficult to say which of these possibilities is most likely. 41Shakespeare’s Pauses, Authorship, and Early Chronology Two late plays by Shakespeare, Henry VIII and The Two Noble Kinsmen, provide additional information about differences between collaborators. Both of these plays are known to have been written with John Fletcher, an expe- rienced playwright 15 years Shakespeare’s junior. Charts 23 and 24 provide the pause patterns for the two playwrights’ respective shares of these dramas. Charts 23–24. Fletcher and Shakespeare in Henry VIII and The Two Noble Kinsmen Both charts reveal a tight similarity in the patterns for these two plays up to and including the fifth position. The sixth and seventh positions, however, show a sharp divergence in the percentage of punctuated pauses. Shakespeare’s more iambic practice gives his lines a higher percentage after the sixth syllable, whereas Fletcher displays an idiosyncratic affection for the seventh position, frequently to underwrite line splits (Oras 1960: 25). If Shakespeare and Fletcher are writing like one another in terms of syntax, that similarity is tempered by an individuality that becomes more apparent later in the pentameter line. 6. Objections and Qualifications The clear exposition of his methodology and “breathtakingly exhaustive” (Wright 1988: 317 n.2) analysis of hundreds of thousands of lines of verse lend Oras’s study the air of the empirical. Yet, as Oras was himself aware, his research – perhaps any such prosodic analysis – depends on a number of assumptions that qualify the findings. As even a casual student of metrics realises, prosody is in certain respects as much an art as a science: experienced 42 Douglas Bruster readers can and do disagree about what counts as syllable, not to mention a stress. Thus an eleven-syllable line to one scholar can seem perfectly decasyl- labic to another. Such variability has obvious implications for the repeatability of a tabulation like Oras’s. Even more uncertain is the status of the text. Oras is careful to cite the editions he uses for his research; for the most part, these are facsimiles of the texts as they were first published. Yet we have learned too much about the conditions of early modern print culture to suppose that these texts represent anything like a direct recording of an author’s practice in iambic pentameter: scribes, compositors, printers, and publishers can all be imagined as potential influences on, and even collaborators in the produc- tion of, the appearance of pentameter in these printed works. Anyone who opens the First Folio, for example, can soon find punctuation marks that seem misplaced, or superfluous; still other lines will seem to lack punctuation that grammar, logic, or clarity calls for. Other influences are internal to the texts themselves: genre, rhyme, prose, characterological idiolect, and dramatic subgenres such as wit combats, dec- larations, and plays-within-the-play (including masques) all have a potential claim on the style of a work’s verse.5 To be considered, too, is the fact that a text may not represent a single ‘event’ of composition, but rather reveal layers of continuation and revision. Last in what is admittedly an abbreviated list of qualifying factors is the dubious assumption that Bathurst made (Bathurst 1857): namely, that patterns move inexorably in one direction. Authors can and do choose to modify the style of their writing, as we are reminded by Julius Caesar – apparently written in the midst of plays very unlike it in many ele- ments of style. If we are tempted to read pause-pattern data as a cardiogram, then, we should be aware of its limitations. Rather than an unmediated record of a single writer’s single event of writing, a text may be a composite of many hands, and many times. It may also be shaped by unknown contingencies of form or occasion. To interpret Oras’s findings without acknowledging their limitations is to ignore his own cautions, relayed in the study. What evidence is there, then, concerning the potential influence of various external factors, literary and industrial, on the pause profiles of early modern drama? The following chart, chart 22, plots the pause percentages for five edi- tions of A Midsummer Night’s Dream. These editions, it should be pointed out, were not generated independently from some Platonic idea of the play; by and large they represent tradition rather than innovation, as editors and printers use previous texts for their models. 5 On the metrical differentiation of character, see Tarlinskaja 1987: 135–176. 43Shakespeare’s Pauses, Authorship, and Early Chronology Charts 25–26. A Midsummer Night’s Dream in five editions; Dream’s three compositors in F1 1623 Even granting such interwoven practice, however, the consistency of percent- ages across the five texts, as shown in chart 25, is remarkable – if partly in suggesting the inertia of punctuation in received materials. A further observa- tion could be made on the closeness of chart 25’s profiles. The first quarto of A Midsummer Night’s Dream seems to be one of the more heavily punctuated playbooks of the English Renaissance. Commas – particularly mid-line com- mas – abound, lending it 43% more punctuation marks than the First Folio Dream of 1623. Yet even with this extreme overabundance of punctuation, the percentages for Q1 Dream are not radically different from the texts that came in its wake. Glancing across at chart 26, we see the percentages for three composi- tors – B, C, and D – commonly ascribed with setting the type of A Midsummer Night’s Dream for the First Folio of 1623. Of the three figures, compositor D diverges the most, yet, even so, the pause profile of his material matches that of the text generally. While further research might confirm or deny an indi- vidualised pattern in other Folio plays, the profile here suggests that, at least in this text, the material has a pause profile regardless of who sets it into type. 7. Conclusions As we have seen, pause pattern analysis in the style of Oras offers useful vantages on questions of authorship and chronology. Yoked with studies of vocabulary, meter, and imagery, and as an ancillary to these genres of analysis, pause profiles can strengthen cases for attribution. A virtue of the procedure involves the binary nature of its evidence: in a text, a space either has or does 44 Douglas Bruster not have a punctuation mark. As such, Oras’s pause procedure is more repro- ducible than more sensitive tests of metrical emphasis and stress. While the latter have the apparent advantage of tapping directly into an author’s linguistic rhythm at some distance from the determining agency of a scribe, composi- tor, or other editorial agent, such tests often seem difficult to recreate in an objective way. As with all such procedures, Oras’s methodology depends on the existence of large numbers of verse lines. Shorter texts – such as, say, a single sonnet or even A Lover’s Complaint – lack sufficient pauses for confident pronouncements to be made about their date or authorship. (The shorter the text, the more tests are needed). At the same time, pause profiles add to our understanding of the rhythmical climate of early modern literature, as well as the personal styles of various writers. They thus can complicate, in the best sense of that word, arguments concerning authorship and writing.6 References [Bathurst, Charles] 1857. Remarks on the Differences in Shakespeare’s Versification in Different Periods of his Life and on the Like Points of Differences in Poetry Generally. London: John W. Parker and Son. Bruster, Douglas; Smith, Geneviève 2014. A New Chronology for Shakespeare’s Plays. Digital Scholarship in the Humanities. Published 8 December 2014. http://dsh. oxfordjournals.org/content/early/2014/12/08/llc.fqu068 accessed 30 May 2015. Carnegie, David; Taylor, Gary (eds.) 2012. The Quest for Cardenio: Shakespeare, Fletcher, Cervantes, and the Lost Play. Oxford: Oxford University Press. Craig, Hugh 2009. The three parts of Henry VI. In: Craig, Hugh; Kinney, Arthur F. (eds.), Shakespeare, Computers, and the Mystery of Authorship. Cambridge: Cambridge University Press, 40–77. Craig, Hugh; Burrows, John 2012. A Collaboration about a Collaboration: The Authorship of King Henry VI, Part Three. In: Deegan, Marilyn; McCarty, Willard (eds.), Collaborative Research in the Digital Humanities: A Volume in Honour of Harold Short, on the Occasion of His 65th Birthday and His Retirement September 2010. Farnham, Surrey and Burlington, VT: Ashgate, 27–65. 6 For their generous assistance with this research, the author would like to acknowledge and thank Yasmine Jassal, Anand Jayanti, and Kelsi Tyler. 45Shakespeare’s Pauses, Authorship, and Early Chronology Elliott, Ward; Robert, Valenza 2010. Two Tough Nuts to Crack: Did Shakespeare Write the ‘Shakespeare’ Portions of Sir Thomas More and Edward III? In: Linguistic and Literary Computing 25(1–2): 67–83, 165–177. Freer, Coburn 1981. The Poetics of Jacobean Drama. Baltimore: Johns Hopkins University Press. Jackson, MacDonald P. 2002. Pause Patterns in Shakespeare’s Verse: Canon and Chronology. In: Literary and Linguistic Computing 17(1): 37–46. Jackson, MacDonald P. 2003. Defining Shakespeare: Pericles as Test Case. Oxford: Oxford University Press. Jackson, MacDonald P. 2012. Looking for Shakespeare in Double Falsehood: Stylistic Evidence. In: Carnegie, David; Taylor, Gary (eds.), The Quest for Cardenio: Shakespeare, Fletcher, Cervantes, and the Lost Play. Oxford: Oxford University Press, 133–161. Jackson, MacDonald P. 2014. Determining the Shakespeare Canon: “Arden of Faversham” and “A Lover’s Complaint”. Oxford: Oxford University Press. Jowett, John (ed.) 2004. The Life of Timon of Athens. Oxford: Oxford University Press. Leitch, Vincent B. 2008. Widening Circles: The Postwar Critical Work of Ants Oras. In: Aunin, Tiina; Lange, Anne (eds.), Widening Circles: The Critical Heritage of Ants Oras. Tallinn: TLÜ Kirjastus, 15–32. McDonald, Russ 2006. Shakespeare’s Late Style. Cambridge: Cambridge University Press. Oras, Ants 1960. Pause Patterns in Elizabethan and Jacobean Drama: An Experiment in Prosody. Gainesville: University of Florida Press. Saintsbury, George 1906. A History of English Prosody: From the Twelfth Century to the Present Day, 3 vols. London: Macmillan and Co. Shakespeare, William 1968. The First Folio of Shakespeare. New York: W.W. Norton. Shakespeare, William 1997. The Riverside Shakespeare, 2nd ed., Evans, G. Blakemore et al. (eds.). Boston: Houghton Mifflin. Slater, Eliot 1988. The Problem of “The Reign of King Edward III”: A Statistical Approach. Cambridge: Cambridge University Press. Spevack, Marvin 1976. Shakespeare Microscopic and Panoramic. In: Mosaic 10(3): 117–127. Suhamy, Henri 1984. Le Vers de Shakespeare. Paris: Didier Erudition. 46 Douglas Bruster Tarlinskaja, Marina 1976. English Verse: Theory and History. The Hague and Paris: Mouton. Tarlinskaja, Marina 1987. Shakespeare’s Verse: Iambic Pentameter and the Poet’s Idiosyncrasies. New York: Peter Lang. Tarlinskaja, Marina 2014. Shakespeare and the Versification of English Drama 1561– 1642. Surrey, England, and Vermont, USA: Ashgate. Taylor, Gary 1987. The Canon and Chronology of Shakespeare’s Plays. In: Wells, Stanley; Taylor, Gary; Jowett, John; Montgomery, William. William Shakespeare: A Textual Companion. Oxford: Clarendon Press, 69–144. Taylor, Gary 1995. Shakespeare and Others: The Authorship of Henry the Sixth, Part One. In: Medieval and Renaissance Drama in England 7: 145–205. Taylor, Gary; Lavagnino, John (eds.) 2007. Thomas Middleton and Early Modern Textual Culture. Oxford: Oxford University Press. Vickers, Brian 2002. Shakespeare Co-Author: A Historical Study of Five Collaborative Plays. Oxford: Clarendon Press. Wright, George T. 1988. Shakespeare’s Metrical Art. Berkeley: University of California Press. 47Shakespeare’s Pauses, Authorship, and Early Chronology Appendix Table 1. Raw counts for pauses from texts cited. Figures from Oras (Oras 1960) unless indicated; tabulations from asterisked titles are by author. Shrew 63 117 52 422 228 206 58 33 8 Hamlet 43 121 65 461 254 470 195 71 19 Cymbeline 61 159 104 493 389 796 473 445 192 1H6 W* 16 55 18 109 63 43 24 17 2 1H6 X* 20 25 7 82 54 50 27 4 3 1H6 Y* 37 61 32 165 114 70 46 27 3 1H6 Z* 17 46 21 107 64 65 24 9 2 Arden non-Sh* 19 33 38 219 112 128 31 10 1 Arden Sh* 4 6 8 73 27 34 6 3 0 King Leir 178 184 57 490 187 117 68 53 5 King Lear 40 104 56 405 160 547 223 131 40 E3 non-Sh* 3 20 22 139 79 36 27 3 0 E3 Sh* 8 24 13 133 63 58 25 4 3 2H6 non-Sh* 6 11 10 39 19 22 8 3 2 2H6 Sh* 78 156 80 445 229 295 106 32 8 3H6 non-Sh* 44 86 55 232 127 125 58 30 1 3H6 Sh* 100 128 87 465 211 223 91 52 7 Titus Peele* 17 51 23 134 73 72 22 10 1 Titus Sh* 45 89 53 282 149 166 45 11 3 Timon Midd* 21 38 25 95 62 97 70 33 15 Timon Sh* 16 45 45 204 170 285 170 86 24 H8 Fletcher 23 92 57 233 213 275 347 178 35 H8 Sh 25 34 26 177 157 328 246 131 71 TNK Fletcher 29 52 60 241 216 306 303 159 43 TNK Sh 10 34 13 127 122 232 163 97 38 Lady’s Tragedy 21 48 32 166 206 424 340 84 21 1 Tamburlaine 26 59 40 194 119 80 26 19 4 Span Tragedy 60 127 48 332 129 97 36 16 3 James IV 14 61 41 434 97 187 30 15 1 Pericles Wilkins 13 23 13 141 55 130 35 16 4 Pericles Sh 5 26 17 97 60 168 79 47 23 Arraignment 17 53 40 117 43 63 17 13 1 Battle of Alcazar 7 45 30 137 47 64 9 3 0 MND Comp B* 7 14 7 60 49 33 12 9 4 MND Comp C* 13 40 16 148 110 80 40 19 3 MND Comp D* 16 37 19 97 78 70 34 11 7