Jump to ContentJump to Main Navigation
Oxford Shakespeare Studies$

Gary Taylor and John Jowett

Print publication date: 1993

Print ISBN-13: 9780198122562

Published to Oxford Scholarship Online: October 2011

DOI: 10.1093/acprof:oso/9780198122562.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 26 February 2017

(p.260) Appendix III The Date and Authorship Of Rollo, Duke of Normandy

(p.260) Appendix III The Date and Authorship Of Rollo, Duke of Normandy

Source:
Oxford Shakespeare Studies
Publisher:
Oxford University Press

Massinger's most recent editors have described Rollo as an ‘impossible problem’.1 Difficult it may be, but the impression of insurmountable perplexity has been largely created by the fragmentation of recent scholarship. Since the Second World War there have been four major investigations of the play: John D. Jump's critical edition (Liverpool, 1948), which in its conclusions about authorship drew heavily upon material in his unpublished MA thesis (Liverpool, 1936); another unpublished dissertation, Bertha Hensman's ‘John Fletcher's The Bloody Brother; or Rollo, Duke of Normandy (Chicago, 1947); G. E. Bentley's account in The Jacobean and Caroline Stage (Oxford), iii (1956), 401–7; and Cyrus Hoy's ‘The Shares of Fletcher and his Collaborators in the Beaumont and Fletcher Canon (VI)’, Studies in Bibliography, 14 (1961), 45–67. Jump mentions in his Preface that, after his edition had gone to press, he received a ‘long letter’ from Hensman, and as a result he has inserted a few references to her conclusions in his Introduction; yet he had clearly not seen the thesis itself, and makes no attempt to assimilate or refute its evidence. Hensman, likewise, makes no reference to Jump's thesis (or, of course, his edition). Bentley, who had supervised Hensman's thesis, endorses its conclusions without reservation, yet gives only the skimpiest account of her evidence; nor had he seen Jump's unpublished thesis. None of these scholars, naturally, could have taken account of Hoy's article; Hoy himself takes no account of Hensman or Bentley. Hensman published in 1974 a two-volume study of The Shares of Fletcher, Field and Massinger in Twelve Plays of the Beaumont and Fletcher Canon (Salzburg); one chapter of this monograph summarizes the argument of her 1947 thesis, with only minor modifications to her conclusions about the authorship of parts of three scenes (ii. 239–79). But although she there makes offhanded reference to Hoy, she makes no attempt to deal with the implications of his evidence. The Oxford Massinger (1976) refers only to Hoy and Bentley. In short, no modern scholar seems to have collected or assessed all the evidence accumulated in these separate investigations.

For convenience one may distinguish three distinct problems (though in fact they are interrelated). When was the play originally written? Was it revised? Who wrote it?

Three very different dates of composition have, at various times, been proposed. The earliest of these, 1613–15, was never based on any evidence, beyond the supposition that Rollo might be the play on which Fletcher, Massinger, Daborne, and Field were collaborating at about that time, according to a letter to Henslowe.2 If one could establish, on other grounds, that the play dates from these years, and that these authors all had a hand in it, then a tentative identification of Rollo with the anonymous play might be hazarded; but one cannot begin with such an identification, and then make it the basis for (p.261) assertions about date and authorship. No modern scholar takes this early date seriously.

Jump accepts an alternative tradition, dating the play c.1625, and conjecturing that Fletcher left it unfinished at the time of his death (pp. xxx-xxxi). Three pieces of ‘evidence’ have been advanced for this date; of these, the most impor tant is a parallel between a passage in Rollo and one in Jonson's Neptune's Triumph (1624). But Jump himself elsewhere (pp. xxv, 84) disparages the significance of this parallel, and Bentley seems entirely justified in objecting that ‘Other plays had such passages, and apparently many court cooks aspired to such elaborate creations … One is, in fact, constantly amazed at the repeated assumptions by scholars that men who had enough originality to write successful plays could not observe the ordinary types about them with their own eyes but had to resort to another writer's observation of a given phenomenon’ (iii. 404). Hensman in 1974 strengthened this argument with the observation that ‘I have found no parallel instance of Jonson rehashing his own lines in a later work; neither was it Fletcher's habit to borrow other men's lines’ (ii. 270). The parallel with Neptune's Triumph thus constitutes dubious evidence for either the date or the authorship of2. 2 or Rollo. A second fact cited in support of a late date is the possible influence of a Latin play, Querolus, on 4. 2 of Rollo; an edition of Querolus was printed in 1619. But even if Querolus did influence Rollo—and Hensman shows that other sources could have contributed the same ideas—several earlier editions of the play had already appeared (Hensman thesis, 118–28). The parallels with Querolus are thus as worthless as that with Neptune's Triumph in dating Rollo. This leaves Jump's late date entirely dependent on his assertion that Fletcher's parts of the play are ‘in his latest style’ (p. xxx). No evidence for this assertion is forthcoming, in Jump's edition or his thesis. Nor are we aware of any reliable internal evidence which could distinguish Fletcher's work in 1617–20 from his work in 1623–6. Jump's own selection of the ‘more striking’ verbal parallels points to an earlier date: thirteen come from plays dated 1617–20, but only eight from plays of 1621–6.3 The Mad Lover (January 1617) and The Loyal Subject (November 1618) have three parallels each, more than any other play in the Fletcher canon. What evidence there is points to Fletcher's contribution having been written earlier than Jump imagines.

Hensman argues persuasively that the play was originally composed in or not long after the summer of 1617. She points out that the play's allusions to and attitude towards duelling almost certainly date from 1616 or after (thesis, 106–8), that the allusion to Charlemagne making ‘Three thousand Knights’ (1. 1. 94–5) looks like a dig at James I's mass-dubbing in June 1617 (pp. 109–10), that the poisoning scene (2. 2) and its aftermath (3. 2) draw upon a whole series of details from the notorious Overbury murder trial, which continued from 19 October 1615 to 25 May 1616 (pp. 114–18), and that the astrologers satirized in 4. 2 all had connec tions with that trial (pp. 126–8). She also suggests that there may have been some personal satire of Lord Haye (pp. 110–14); this is very conjectural, though given the prominence of other topical allusions from 1616–17 it can hardly be dismissed outright. Hensman might also have mentioned the metaphorical description of (p.262) ‘this nights Freedome’ as ‘a short Parliament’ (i. i. 408): Parliament was dissolved after short sittings in 1614, 1621, and 1626. Which of these dates pertains can hardly be demonstrated, but the 1614 Parliament was the shortest of the three: it lasted only two months.

Cumulatively, Hensman's evidence points very strongly to composition in or not long after mid-1617. Though some of this material could have been exploited for some time afterwards, it would be most attractive when most topical. Nothing in Hoy's or Jump's investigations in any way imperils this conclusion. No topical references to events later than June 1617 have been identified. Moreover, the absence of any reference to the play among Sir Henry Herbert's surviving memoranda supports—albeit tenuously—this early dating.

But Hensman also contends that the extant texts of Rollo derive from a substantial revision of the play, by Massinger, between 1627 and 1630. Despite Bentley's endorsement, Hensman's case for this revision is feeble. It depends upon the argument that Massinger wrote all the parts of the play which make use of Gentillet's Discours sur les moyens de bien gouverner (thesis, 74–101); in her view Massinger substantially rewrote the play under the influence of this source. This hypothesis requires Massinger to have written 4. 3, a scene with a large number of heroic couplets, which previous investigators have assigned to Field or Chapman. No one has ever assigned 4. 3 to Massinger. Moreover, Hensman admits that neither Fletcher nor Massinger can be responsible for the bulk of 3. 1 precisely because of the number of heroic couplets in that scene (p. 167); she never mentions that 4. 3, on the same evidence, cannot have been written by Massinger either. If it was not, her whole hypothesis collapses. Likewise, in 3. 1 she originally offered Jonson as Fletcher's collaborator on the original play; Jonson's hand has otherwise never been detected in that scene, and Hensman supplied no evidence of his presence beyond the statement that he used heroic couplets in his Roman tragedies (though not at all in this fashion). On its own terms, then—and even before Hoy's linguistic evidence has been examined—Hensman's identification of the shares of original author(s) and reviser is self-contradictory and implausible. Moreover, attribution to Massinger of all the material inspired by Gentillet, or which involves the characters Aubrey and Latorch, produces an unwieldy, unnecessary, and bewilderingly implausible pattern of revision.4

Hensman's case for revision is, even within the terms of Rollo alone, implausible; more generally, we have little faith in the intellectual procedures which pro duce this hypothesis. She begins with the postulate, most succinctly articulated in 1974, that ‘whenever a play is supposed to be a collaboration … when stylistic differences are commingled within speeches and episodes, the passages in which such commingling occurs may reasonably be supposed to have undergone subsequent revision’ (i. 4–5). Collaborators presumably did not sit down together taking turns writing every other speech, and the mixing of authorial characteristics in one passage or episode does suggest that one writer has gone over another's (p.263) work. But by ‘subsequent revision’ Hensman does not mean the reshaping and tidying which occur between the authors' first separate drafts and the first performance; instead, she believes that such a mingling of authorial characteristics proves that a play has been substantially revised, years later, by one of those authors. Hensman thus rules out any possibility of one collaborator modifying his colleague's first draft, in the effort to pull their separate labours together into a coherent whole. As collaborators ourselves, we find this assumption incredible. Moreover, as John Kerrigan has elsewhere demonstrated, later adaptation for the-atrical revivals generally confined itself to relatively self-contained interpolations and alterations.5 If anything, logic and the evidence would suggest that a tangled interweaving of the traits of two authors in one passage proves that the intervention of the second hand came very soon after the original composition, rather than later. Hensman's postulate seems unwarranted, and the evidence she uses in applying that postulate is equally unreliable. Hensman bases her claim for the presence of Massinger or Fletcher entirely on her account of their respective ‘verse styles’. The verse of each of these playwrights does exhibit tendencies relatively rare in the other's, but such criteria are of dubious value in determining their share of a particular collaborative play, particularly when the argument con-cerns Massinger's responsibility for a single speech, or a few lines, embedded in a scene written by Fletcher.

To such unreliable evidence of authorship Hensman adds, finally, the evidence of source material. Hensman deserves considerable credit for identifying the sources which contributed to Rollo, but she conspicuously fails to prove a direct relationship between authorship and sources. Gentillet was available when Rollo was originally written; Massinger was at that time already collaborating with Fletcher; consequently, even if Gentillet's influence were only discernible in passages clearly written by Massinger, such an association could not establish that Massinger's contribution was written after Fletcher's death. Nor does the use of Gentillet in certain passages written by Massinger create a presumption in favour of his authorship of other passages influenced by Gentillet: after all, Fletcher and Shakespeare both drew upon Chaucer's ‘Knight's Tale’ in writing their respective shares of The Two Noble Kinsmen, and an examination of Cyrus Hoy's introduc-tions to the plays of the Dekker canon confirms the normality of such overlap-ping. By assuming a constant correlation between the use of certain sources and the presence of Massinger, Hensman is forced to construct similarly elaborate hypotheses for Massinger's revision of many other plays (The Queen of Corinth, Thierry and Theodoret, The Laws of Candy, A Very Woman, etc.). One's confidence in Hensman's conclusions about Rollo diminishes with every additional play forced into the same Procrustean mould.

Hoy's linguistic evidence gives this whole scenario the scholarly kiss of death. Hensman must assign 4. 1 to Massinger, since it contains some of the most striking parallels with Gentillet; yet it contains one ơth' and one ơthe, forms which Massinger never uses in any of his acknowledged work. She must also, for similar reasons, assign 3. 1. 388–420 to Massinger; yet it contains two examples of ye, which Massinger likewise never uses. Hensman must also assign to Fletcher certain passages in Act 1, which all previous investigators have given, in its entirety, (p.264) to Massinger. But there is nothing to suggest Fletcher in any of these passages; the total absence of ye argues against him, as does the presence of the contraction V (1. 1. 54), a form which occurs only once elsewhere in Fletcher's acknowledged work. The same contraction occurs three times in 4. 2, which Hensman would assign to Fletcher; hath also occurs three times in that scene, though Fletcher elsewhere never used it more than six times in any single play. Hoy's linguistic evidence, in all of these scenes, merely confirms standard attributions based on parallel passages, metrical practice, and stylistic mannerisms. Hensman's account of the text's evolution must be rejected, as much for its own weakness as for the strength of the authorship evidence arrayed against it.

Of course it remains possible that the extant text has suffered some degree of revision or adaptation, even if not in the drastic form conjectured by Hensman. But Hensman offers no evidence for such a redaction. The play was certainly performed at Court in 1630, and this performance must surely have been prompted by a public revival; thereafter the play was extremely popular. However, records of performance survive only by chance. A play with such provocative allusions to the Overbury murder trial may well have been kept out of the royal repertoire for a decade or more, even if it were popular in the London theatres.6 No prologue advertises that the play has been in any way recast; as Jump notes, this suggests (though it cannot prove) that no substantial revision has taken place (p. xxix). Hensman's conjectural dating of the revision (1627–35) depends upon the twin assumptions that Aubrey's part was played by and written for John Lowin, and that Lowin must have been the same age as Aubrey when he first played it (pp. 20–1). Neither contention seems reliable evidence for the date of a revision, even if we could be sure that one took place.

Moreover, even if the extant texts do represent a Caroline redaction, we have no particular reason to suspect that any of the songs have been interpolated. Like most other investigators, Hensman assigns to Fletcher—and hence to the original period of composition—all three songs; she also includes 4. 3, which prepares for the seduction in the final scene, among the unrevised material. Everyone assigns to Fletcher the central portion of 3. 1, which initiates the Edith plot. Therefore, beyond the presence of a similar stanza in 5. 2 of Rollo and 4. 1 of Measure, we have no extrinsic or intrinsic reason to suspect that the song in Rollo was interpolated at a date later than the original composition of the play, c. 1617–20. Whatever our view of the textual situation in either play, no one doubts that the original composition of Measure preceded the original composition of Rollo; therefore, anyone who wants to defend the integrity of the song in Measure can always claim that it was lifted from Shakespeare's play c. 1617–20. Speculations about a late revision of Rollo do little to strengthen that defence.

Who wrote the play?7 That Fletcher and Massinger were the main contributors (p.265) has never been seriously doubted; nor (with the exception of Hensman's Byzantine theory of revision) has there been much dispute about their shares of the play: Massinger wrote Act 1 and 5. 1. 1–89; Fletcher Act 2, 3. 1. 263–330, 3. 2, the closing soliloquy in 5. 1, and most if not all of5. 2. Hoy's linguistic evi-dence amply confirms these attributions. But these tests also confirm that the rest of the play—Act 4, and all but 67 lines of 3. 1—cannot have been written by either Massinger or Fletcher; moreover, as previous investigators have almost universally agreed, the linguistic evidence suggests that two different authors may have contributed to this central section of the play. The probable date of composition almost certainly rules out Daborne and Wilkins; the linguistic evidence argues against ascription to Middleton, Rowley, or Daborne, and does little to encourage Wilkins's candidacy.8 Only three plausible candidates remain among those often identified in this portion of the play: Chapman, Field, and Jonson.

Hoy demonstrates that Jonson could, on the linguistic evidence, have written 4. 1 and 4. 2, the scenes assigned to him by Jump; Chapman, likewise, could well have written 4. 3 and the bulk of 3. 1—which Jump, following William Wells, assigned to him. Field's linguistic pattern could in fact fit either set of passages; in some ways it fits them better than Chapman's or Jonson's do. However, the total absence of prose, or of rhyming couplets, from 4. 2 makes Field's presence unlikely, as does the learnedly arcane vocabulary, and no one has ever been tempted to see Field's hand there. Herford and Simpson concede that there are some striking parallels between 4. 1 and Jonson's work, particularly in Discoveries, and their only argument against his authorship is the assertion that The verse is too fluid for Jonson'.9

Hensman demonstrates that most of the political theory in 4. 1 derives from Gentillet, a major source for the play's plot; Jonson's expression of similar sentiments in Discoveries thus constitutes the weakest of evidence on which to attribute to him the authorship of this scene. Likewise, no significant verbal paral-lels between 4. 2 and the Jonson canon have ever been offered. The mockery of astrological jargon might be Jonson's; but Hensman shows that such satire was frequent enough in the early seventeenth century, and by no means a Jonsonian monopoly. As for metrical statistics, Jump in his own thesis was forced to admit that they ‘not only fail to give striking confirmation of the presence of Chapman and Jonson … in Rollo but even fail to bring out at all clearly the quite unmistakable fact that the author of IV.iii and the riming parts of IILi did not write IV.i and ii’ (p. 54). Metrical evidence does not rule out Jonson; but nor does it compel us to rule him in. As for the linguistic evidence, Hoy himself conceded that, in itself, it hardly constitutes convincing proof of Jonson's presence: it merely ‘confirmed’ an attribution made primarily upon other evidence. Such (p.266) ‘other evidence’ turns out, upon closer examination, to be illusory. Hoy might, indeed, have turned the linguistic evidence against Jonson's authorship: 4. 1 contains one example of w'yee, SL form which appears nowhere else in the Jonson canon, but which does occur four times in Field's two unaided plays. The most probable date for Rollo's composition also tells against Jonson's participation. Jonson's last known collaborations were Sejanus (1603) and Eastward Ho (1605)—both, for different reasons, disastrous. Rollo would not only constitute a puzzling exception to the artistic independence of Jonson's subsequent career; it would also interrupt his period of silence, from 1616 to 1624, when he appears to have written nothing else for the public stage. Such considerations cannot be regarded as decisive: if Jonson did contribute to Rollo then we would have to revise our picture of his career, and that picture cannot itself be allowed to arbitrate on his presence in Rollo. However, the oddity of Jonson's participation, in terms of his personal circumstances, does cast further doubt upon an attribution based on such flimsy internal evidence. Moreover, the very same considerations make Field an attractive candidate. Field was, in 1617–20, at the height of his career; all of his known work from 1611 on was written in collaboration; in particular, he collaborated with Massinger in The Fatal Dowry (1616–19), with Fletcher in Four Plays in One (1614–16?), and with both in The Honest Man's Fortune (1613), The Jeweller of Amsterdam (1616), The Queen of Corinth (1617), and The Knight of Malta (161 8). No other playwright is known to have collaborated with both men between 1616 and 1626; only Middleton, Rowley, and Dekker collaborated with either, and all three seem ruled out of Rollo on other grounds. Field may not have contributed to Rollo, but on present evidence he seems likelier than Jonson to have done so.

The fourth playwright is also difficult to identify. Chapman in his other tragedies never uses ‘em; it occurs twice in the first part of 3. i.10 He only uses i'th' four times, over the course of four entire plays; it, too, occurs twice in the first part of 3. 1. On this basis one can hardly be confident about assigning that scene to Chapman; yet both anomalies, along with the liberal sprinkling of couplets and the rest of the linguistic pattern, would be acceptable in Field. Likewise, the value of the verbal parallels with Chapman's work is perhaps somewhat diminished by the fact that Field, whom Chapman called his ‘son’, was especially prone to imitate Chapman in scenes of high tragedy, like 3. 1 and 4. 3 of Rollo.11 A. R. Braunmuller raises a number of plausible objections to the attribution to Chapman: \d) little of Rollo is dense enough to be unrevised Chapman and (b) there's rarely the wealth of (often obscure) classical allusion one would expect and (c) it's surprising to find Chapman writing a scene with all female characters (4. 3).’12 Moreover, the concentration of couplets, which so remarkably distinguishes 4. 3 and most of 3. 1 from the rest of the play, if anything tells against the attribution to Chapman. Braunmuller, like Jump in his thesis, compares Rollo to Caesar and Pompey, which many scholars regard as Chapman's last play; but (p.267) Braunmuller analyzes the distribution of couplets, where Jump merely counts them:

Excluding pairs of lines that end (or happen to end?) on the same word, often a proper noun, I would judge that Chapman uses couplets for the same purposes most of the older Elizabethan survivors did: as scene endings, to mark sententiae, and (less frequently) to close a speech, especially a Nuntius or semi-narrative speech. With those exclusions, I see only eight, widely dispersed, couplets in Caesar, and nothing to match the stretch at Rollo, 3. 1. 22–29, or even the four lines at 4. 3. 54–57. In Rollo, the same exclusions leave (in addition to the long passage): 3. 1. 35–6, 84–85, 149–50, 341–42 (?), 34M7, 379–80 (?), 4– 3– 13–M, 34–35 (? a flabby sentence), 59–60, 74–75. Moreover, the proportion of speech-ending couplets is much higher than in any comparable stretch of lines in Caesar.

Field, by contrast, used couplets much more generously, in the manner of Rollo 4. 3 and 3. 1. His share of The Fatal Dowry (three-and-a-half scenes) contains fifteen examples, to Caesar's eight; they are extraordinarily frequent in his two early unassisted plays. Finally, Braunmuller notes that there is little real evidence for attributing to Chapman the end of 3. 1, after Fletcher's episode; we may add that there is no real evidence for attributing 4. 3 to him. This leaves us with only a handful of Chapman parallels, in the first section of 3. 1, which could claim to be of any positive value in identifying Chapman rather than Chapman's ‘son’ Field.

All in all, Chapman's claim cannot be considered at all secure. Like Jonson, Chapman had a markedly idiosyncratic style, particularly in his tragedies, and although certain scenes of Rollo gesture in the direction of that style, none actually achieves it. Imitation is, in the circumstances, a strong possibility, and Field's candidacy is if anything buttressed by the sense that we are dealing with writing ‘like but not by’ Chapman. On the other hand neither scene contains the exclamation ‘Pish’, or a speech spoken by ‘Omnes*—mannerisms of which Field was fond.13 The slight discrepancy between Rollo and Chapman's linguistic pattern in his earlier tragedies could be explained by the fact that we have no certain examples of his unadapted late dramatic work. Furthermore, Brinkley's rejection of Field's claim to a part of this play carries some weight: it depends, not on any subjective judgement that the scenes in question fall below Field's standard, but on the alleged absence—not only from these scenes but from the whole play—of certain mannerisms present elsewhere in Field's work.14 However, her statement that metrical tests ‘preclude the possibility of his collaboration’ should be treated with some scepticism. The most important of such tests involves Field's consistent preference for ‘strong stops’ (full stop, exclamation, query, dash, colon, semicolon) at line-ends rather than in mid-line. For 4. 3 and non-Fletcherian 3. 1 the totals are, by our count, 15 line-end: 7 medial. Such a distribution contrasts (p.268) strongly with 4. 1 and 4. 2 (24 line-end: 64 medial), and could actually be Field's. Since this metrical test is crucial to Brinkley's rejection of Field, her argument can hardly be considered compelling. In fact, the two scenes do contain some of Field's favourite tricks, such as the tree image at 3. 1. 15–16 or the classical allu-sion at 3. 1. 353. The use of prater in the stage direction (3. 1. 387) would also be characteristic of Field (or Chapman, Massinger, or Jonson). In the absence of a more detailed investigation of the differences between Field's work and Chapman's, we can hardly rule out either; but one or the other probably contributed to the play, and at the moment Field looks more attractive than Chapman.

We had reached this conclusion even before seeing Hensman's 1974 book, where she abandons Jonson and instead advocates Field's authorship of the non- Fletcherian parts of 3. 1. She does not mention the linguistic evidence, but does draw attention to the ‘invariable regularity of iambic pentameter rhythm’, the ‘inverted and awkward grammatical structure’ of several lines (e.g. 3. 1. 8–9, 15–16, 26), and ‘numerous examples of Field's characteristic use of words in their archaic or root sense’ (ii. 273–4). Elsewhere she mentions Field's fondness for metaphors based on minerals or precious stones, his (sometimes incongruous) appeals to the five senses, and his frequent use of mild oaths and Anglo-Saxon verbs (i. 57–60)—all applicable to the language of 3. 1 and 4. 3. Despite the vagueness of some of these criteria, they do usefully demonstrate that the writing of 3. 1 and 4. 3 is within the range of Field's habitual manner, and that Brinkley's rejection of his claims cannot be given any credence. Unfortunately, though, Hensman makes no attempt to distinguish Field's manner from Chapman's, and bases her rejection of the Wells/Jump/Hoy attribution entirely on what she herself called an 6 ad hominerrf argument (ii. 274 n. 38). She alleges that Chapman had given up play writing by this time, and last collaborated in 1605—two false statements, which therefore hardly constitute convincing arguments against Chapman's participation.

What conclusions can we draw from all of this? Most of Rollo was undoubtedly written by Fletcher and Massinger; just as clearly, the linguistic, stylistic, and metrical evidence all testify that one or more other hands was involved. Though they radically disagree on what he may have written, Jump, Hensman (originally), Bentley, and Hoy all identify Jonson's hand in the play; but of the three candidates usually advanced, Jonson's claim is far and away the weakest. Certainly, the Hensman-Bentley attribution to him of most of 3. 1 has no discernible basis in either logic or evidence, and the detectable signs of his presence in Act 4 are manifestly unreliable.

We are disposed to regard Jonson's participation as unproven and relatively unlikely. By contrast, the presence of either Field or Chapman, or both, is highly probable, but intrinsically difficult to prove. Rollo might well have been Field's last play, and for almost a decade before its most probable date of composition, he had been working only in collaboration. These two facts in concert make it difficult to evaluate apparent discrepancies between Rollo and Field's other work. Likewise, Rollo would have been Chapman's first play for many years; only one Chapman play undoubtedly later than Rollo survives (Chabot, Admiral of France), and its extant text has suffered an indeterminable amount of posthumous revision. These circumstances render the evaluation of internal evidence more than (p.269) usually hazardous. The personal and artistic relationship between the two men does not make it any easier, particularly since (for reasons explained below) Chapman's influence on Field may have been forcefully renewed in 1619. Chapman clearly prefers ‘while’ and Field clearly prefers ‘whil(e)st’, but the word does not appear at all in the relevant parts of Rollo. Field's claims to 4. 3 and most of 3. 1 seem to us, on present evidence, better than Chapman's. But one or the other almost certainly contributed to the play, possibly in collaboration with a fourth unidentified figure.

Chapman's presence, or Field's, would reinforce the evidence for dating the play c. 1617–20. If Field contributed to the play, it must have been written between his move to the King's Men and his death in early 1620.15 Chapman's complete translation of Homer was published in 1616; his translation of Hesiod appeared in 1618, prefixed with a commendatory poem by Jonson. After the Hesiod, Chapman's next known work was Chabot, Admiral of France, which Bentley assigns to 1621 or January 1622; his lost collaboration with Brome, Christianetta, may well date from the following years, and in any case can hardly be much earlier, since Brome first surfaces as a dramatist, in another collaborative play, in October 1623. So far as the fragmentary record allows us to judge, after completing his translations of Homer and Hesiod, Chapman again, after a long absence, turned to writing plays—in the same period that Rollo must have been written.

There is thus nothing intrinsically implausible about Chapman collaborating in a play written for the King's Men c. 1617–20. However, neither Christianetta nor Chabot was written for the King's Men; nor, for that matter, were any of Chapman's other plays. Chapman only ever had one connection with the King's Men: Nathan Field. Early in his career Field had performed—certainly or probably—in Chapman's Gentlemen Usher, May Day, Sir Giles Goosecap, All Fools, Bussy D'Ambois, Monsieur D'Olive, The Widow's Tears, Eastward Ho, 1 and 2 Byron, and The Revenge of Bussy; the only commendatory verses attached to Field's first play were written by Chapman, and addressed ‘To his Loved Sonne, Nat. Field’; Field's most recent editor regards Chapman as the major influence on his work.16 Moreover, although Bussy D'Ambois originally belonged to a children's company, Field obviously took it with him when he joined the King's Men; it is the only one of Chapman's plays which ever formed part of their repertoire. Some of Chapman's editors go so far as to claim that Field had a main hand in the much-altered redaction of Bussy, published in 1641; whether or not this is true, there can be little doubt that Field, in some sense, ‘took Chapman with him’ when he joined the King's Men. Field's brief tenure with that company therefore provides the only obvious opportunity for Chapman's collaboration on a Fletcher/Massinger play for the King's Men.

Since Wells first proposed him, in 1928, every investigation but Hensman's has accepted Chapman's presence in the play; Hensman completely ignored Wells, and made no attempt to contest his conclusions. If, notwithstanding Braunmuller's doubts, Chapman did contribute to Rollo, then his personal (p.270) circumstances would allow us to date the play with greater precision than Hensman attempted. A series of documents from a lawsuit, discovered by C. J. Sisson and Robert Butman, make it clear that, for several years after 1612, Chapman was in financial straits, living in obscurity in Hinckley; he returned to London in the autumn of 1619, but before that he seems clearly to have stayed out of the capital, except for an appearance in June 1617, of indeterminate but probably brief duration.17 Chapman's participation would therefore limit composition of Rollo to either the summer of 1617, or to some time in or after autumn 1619. Field's death, some time in1619 or early1620, provides a probable terminus for any association Chapman may have had with the King's Men; a date later than 1620 would also presuppose the use of ‘topical’ material the topicality of which had already waned. In terms of Chapman's own circumstances, the later of these two dates—1619/20—seems the more probable; but the topical allusions would favour mid-1617.

The later date might also explain an anomaly in the sources which Hensman left unresolved. The plot of Rollo is clearly based on Roman, not French, history, and the change of date and locale has never been satisfactorily explained; yet all but one of Chapman's tragedies have French settings. More specifically, Hensman pointed out that the portrayal of Rollo as a tyrant was highly unusual, and that this treatment would seem to derive from Duchesne's Gesta Normannorum in Franciae Ante Rollorem Ducem (1619); the authors ‘either drew upon Duchesne's volumes, or upon early Norman chronicles similar to them’ (Hensman thesis, 136). Her own extensive survey uncovered no such ‘similar’ chronicles, and in terms of the historical literature—and popular consciousness—Duchesne's volumes represent a clear departure, the first polemical representation of Rollo as a tyrant. Until some other source for this point of view can be found, Rollo must be dated in 1619–20 rather than 1617.

Two other facts might also lend support to this date. One is the already-mentioned publication of a reprint of Querolus in 1619. The other is Jonson's reference, in January 1618/19, to a number of contemporary playwrights, in his private conversations with Drummond (Fletcher, Chapman, Field, Day, Middleton, Shakespeare). Massinger is not mentioned at all, which would perhaps be rather surprising if Jonson had collaborated with him during the last eighteen months. But neither of these considerations carries much force: earlier editions of Querolus were available, and Jonson probably didn't contribute to Rollo at all. The extant source materials favour 1619–20, the topical allusions favour late 1617 or 1618, and for our purposes no decision need be taken about which of these dates should be favoured.

The perceptive reader will perhaps have noticed that we have as yet said nothing about the authorship of the beginning of 5. 2, in which the song ‘Take oh take’ occurs. Jump, like virtually every earlier investigator, assigns the entirety of 5. 2 to Fletcher, yet in the first 134 lines he cites only two verbal parallels: ‘spring of beauty’ at 5. 2. 35 (which occurs in The Elder Brother and The Prophetess) and ‘Angell eyes’ at 5. 2. 87 (which also occurs in The Humorous Lieutenant, Monsieur Thomas, and Women Pleased). Yet these 134 lines contain eighteen examples of (p.271) you, and not a single ye; Hoy on this basis concludes that they can hardly be Fletcher's, and he assigns them to Massinger. Hoy cites Massinger parallels for two phrases in this section of the scene. For Massinger's ‘my sacrifice | Of love, and service’ (Roman Actor 3. 2. 174–5) Rollo has The gentle sacrifice of love and service' (5. 2. 26), and we have noticed no parallels for this phrase in Fletcher's undoubted work. Hoy also attributes to Massinger references to the ‘wind’ in ‘Arabia’, associated with ‘perfumes or … spices’ (Duke of Florence 2. 3. 62–4, New Way to Pay Old Debts 3. 1. 73–4, Bashful Lover 1. 1. 234; Rollo 5. 2. 39–41). But Fletcher also refers to the ‘sweet & Arabian gums’ (Noble Gentleman 1. 4. 12¬ 13). The evidence of verbal parallels in the first ninety lines of 5. 2 strongly suggests either that Fletcher and Massinger both worked on this passage, or that verbal parallels are of little use here in distinguishing Fletcher from Massinger. Given the close working relationship of the two playwrights and Massinger's tendency to echo other men's plays, it would be unwise to be dogmatic about their shares here; but shared composition is obviously possible. Hoy assumes instead that Massinger later revised something of Fletcher's, but it is difficult to see why or how this should have happened. The play must always have included something like this scene, with the death of Rollo and the culmination of the Edith subplot; anyone wanting to interpolate the song could have done so without altering much of the preceding or following dialogue. The first part of the scene (11. 1–92) in fact uses you only twice, and contains one Has (for ‘He has’), a form Massinger only uses once elsewhere. This section of the scene also contains both of the striking verbal parallels with Fletcher's work (Jump, p. 106), and many of Fletcher's stylistic mannerisms (Hensman thesis, 187). Fletcher, moreover, had been given the ‘Edith’ section of 3. 1, and the dramatic material at the beginning of 5. 2 seems entirely to his taste. By contrast, the middle of the scene, 5. 2. 93¬ 134, contains sixteen examples of you, no examples of ye, no clear parallels of style or vocabulary with Fletcher's work, and nothing which would contradict ascription to Massinger. Massinger, by universal agreement, wrote the beginning of 5. 1, which prepares for Hammon's attack on Rollo, and this section of 5. 2 begins with Hammon's entrance and ends just after he has killed Rollo and been fatally wounded himself; Fletcher, likewise by consent, wrote most or all of Aubrey's long closing soliloquy in 5. 1, and Fletcher's hand clearly resumes with Aubrey's entrance at 5. 2. 136 (‘I charge yee let us passe’). Assuming that Massinger wrote 5. 2. 93–134 will thus account for all the available evidence, and produce a pattern of collaboration no more implausible than Fletcher's intervention to write sixty-two lines in the middle of 3. 1 (which everyone acknowledges).

Rollo thus appears to have been written in mid-1617 or between autumn 1619 and summer 1620; the extant text shows no signs of later revision or adaptation. Massinger wrote Act 1; Fletcher Act 2, 3. 1. 263–330, and 3. 2; both contributed to both scenes in Act 5. The authorship of Act 4, and of the remainder of 3. 1, remains uncertain, but Jonson's participation seems relatively unlikely. Either Field or Chapman very probably contributed to the play, and an unidentified fourth dramatist may also have been involved.

Notes:

(1) The Plays and Poems of Philip Massinger, ed. Philip Edwards and Colin Gibson, 5 vols. (Oxford, 1976), i, p. xx n. 5.

(2) Henslowe Papers, ed. W. W. Greg (London, 1907), 65–6.

(3) We have checked Jump's parallels against recent scholarship on the authorship of the plays in the Fletcher canon, and only included plays or parts of plays currently attributed to Fletcher himself.

(4) Massinger seems to have undertaken a revision of Love's Cure similar to that which Hensman conjectures for Rollo: see George Walton Williams's introduction, in Beaumont and Fletcher Canon, iii. 4–7. However, in this case Massinger's Spanish source was not pub-lished until 1625, so its influence palpably derives from late adaptation of Beaumont and Fletcher's original play. Moreover, in Love's Cure Hoy's linguistic evidence clearly links Massinger with the material influenced by this source.

(5) See Ch. 3 n. 90

(6) For an exact parallel see The Chances: Bentley plausibly dates this c.1617 (iii. 318–23), but the first recorded performance was on 30 Dec. 1630. As Walton Williams remarks, ‘Though there are no records of performances of the play during Fletcher's lifetime, it is probable that a play that was to become most popular on the stage would have been so from the first’ (Beaumont and Fletcher Canon, iv. 543).

(7) The Stationers' Register entry (4 Oct. 1639) assigns the play to ‘J:B’, and Qi (1639) to ‘BJ.F.\ These initials have been interpreted as a reference to Fletcher and/or Beaumont and Jonson; no one has ever suspected Beaumont's hand (he was almost certainly dead before the play was written), and although Jonson has been suspected the initials are poor external evidence. As Jump concludes, after a droll ummary of the various interpretations of the crytogram, ‘It seems that a study of these initials will not solve the authorship problem’ (p. xxvi). The more reliable, independent Q2 (1640) identifies only ‘JOHN FLETCHER’ as the author.

(8) Compare Hoy's tables with Lake's (bands 1 and 4). Middleton, Rowley, and Daborne all prefer has and does; only Rowley and Wilkins offer any parallels for i'the and o'the (one each). The date of Wilkin's death is not known, but his last known writing dates from 1608.

(9) Ben Jonson, Works, ed. C.H. Herford and Percy and Evelyn Simpson, 11 vols. (Oxford 1925–5), x (1950), 295.

(10) We have here been able to supplement Hoy by means of Jackson, Attribution,13.

(11) For Chapman parallels elsewhere in Field see The Plays of Nathan Field, ed. William Perry (Austin, Tex., 1950), 30; on Field's straining after high style, see Hoy.

(12) Private communication, 10 Nov. 1983. (Braunmuller kindly commented for us on the plausibility of attributing anything in Rollo to Chapman; he did not discuss the merits or demerits of Field's claim.)

(13) Field overwhelmingly preferred ‘oh’ (see App. II, above), and Chapman, in Bussy D'Ambois at least, preferred ‘o’; 4. 3 lacks the exclamation altogether, but in the non-Fletcherian lines of 3. 1 it occurs 10 times, always as ‘o’. While the presence of ‘oh’ would argue strongly for Field, the use of ‘o’ (also overwhelmingly preferred elsewhere in Q2) is of little value, since it may merely represent the preference of the scribe who prepared the printer's copy.

(14) R. F. Brinkley, Nathan Field: The Actor-Playwright, Yale Studies in English, 77 (New Haven, Conn., 1928), 141. The % of feminine endings in the unFletcherian part of 3. 1 is 34, in 4. 3 34.5; Field's highest % elsewhere is 19 (Fatal Dowry). However, Field's % rises from his early to his late work, and Rollo might be his last play.

(15) On the date of Field's death, between May 1619 and Aug. 1620, see Bentley, Jacobean and Caroline Stage, ii. 435 and iii. 301.

(16) Perry edn. (see n. 10), 14, 29–32.

(17) ‘George Chapman, 1612–22: Some New Facts’, Modern Language Review, 46 (1951), 185–90.