University of Virginia Library

Pericles was first published in a quarto of 1609, which attributed the play to William Shakespeare and claimed that it had been "divers and sundry times acted by his Majesty's Servants at the Globe." Scholars agree that the quarto, on which modern editions must be based, provides an imperfect text, though there has been disagreement about the extent and nature of the corruption. The majority view is that it is a "bad quarto," printed from an unauthorized manuscript, which may have been concocted largely from memory. But the quarto's textual imperfections are not so drastic as to obscure a stylistic disparity between the play's first two and last three acts. The second part of Pericles is unmistakably Shakespearian; the first part is not.[1]

Several theories have been advanced to account for this state of affairs. It has been suggested that the circumstances of the text's transmission may have produced the disjunction in style, two different "reporters" having reconstructed the two parts of the play.[2] Another hypothesis is that acts 1 and 2 of Pericles are vestiges of an early Shakespearian play that the mature dramatist thoroughly revised over acts 3-5, or that Shakespeare wrote acts 1 and 2 at the beginning of his career and failed to finish the play until around 1606-1607.[3] Other commentators have supposed that Shakespeare wrote the whole of Pericles within a single period but that he deliberately made the first two acts odd for some special artistic purpose.[4] The choruses,

in which the resurrected medieval poet John Gower furnishes narrative links in the dramatization of a story for which his Confessio Amantis served as a significant source, are composed in couplets imitative of the real Gower, and the suggestion is that the play's dialogue may have been consciously adjusted to the choruses' apt archaism. There is also a group of modern critics who seem almost oblivious to the difference between the stilted verse of Pericles 1-2 and the vigorous poetry of Pericles 3-5. But the most common view has been that the first portion of the play was written by another playwright—that Shakespeare either reworked the second half of the other man's play, or completed his unfinished play, or collaborated with him.[5]

No other play generally accepted as wholly Shakespeare's was excluded from the First Folio of 1623. So the failure of Pericles to appear in that volume tells strongly in favour of the theory of divided authorship.[6] The leading candidate for the role of non-Shakespearian contributor to Pericles has long been the minor dramatist George Wilkins, whose one independent play, The Miseries of Enforced Marriage, was published two years before Pericles, with a title page claiming that it had been acted by Shakespeare's company, the King's Men; over the last twenty years or so some quite substantial evidence has accumulated in support of Wilkins's authorship of Pericles 1-2.[7] The purpose of this paper is to add some more.

Among the most valuable contributions to the case for Wilkins's participation in Pericles is David J. Lake's investigation of the play's rhymes. There are 145 rhymes within acts 1-2 of Pericles and 128 within acts 3-5. Lake offers a mass of data demonstrating that certain features of the rhymes in Pericles 1-2 associate this portion of the play much more closely with the known verse of Wilkins than with the acknowledged verse of Shakespeare

or with work by any of the alternative candidates for the role of Shakespeare's collaborator in Pericles, namely John Day, William Rowley, and Thomas Heywood.[8] The most important of these features are the prevalence of assonantal near-rhymes, especially those involving nasal consonants; the use of a high proportion of rhymes with a short vowel plus stop-consonant; and the frequent use of an "aabcc" rhyme pattern and of sporadic outbursts of couplets surrounded by blank verse. Lake gives figures for the rates of occurrence of these features within Pericles and plays by various playwrights. He also lists a few individual rhymes shared by Miseries and Pericles 1-2, and mentions that H. Dugdale Sykes had noted several such specific rhyme-word links as long ago as 1919.[9] But the investigation of individual rhymes, rather than types of rhyme, can be carried a good deal further.[10]

My procedure was first to compile a list of all the rhymes in Pericles and in The Miseries of Enforced Marriage.[11] The texts were scanned carefully from beginning to end and rhyme words were keyed into a computer in alphabetically arranged pairs: breath|death, ill|will, life|wife, and so on. Those from Miseries were tagged with an extra final symbol so that they could be identified as coming from that source play, not from Pericles. The order in which a rhyme's components appeared within the text—ill followed by will, or will followed by ill—was regarded as irrelevant. This holds for every phase of the investigation reported in this paper. The majority of rhymes were in couplets. When three consecutive lines rhymed (as in "away . . . stay . . . day"), three rhyme pairs were recorded: away|day, away|stay, and day|stay. Similarly a sequence of four lines rhyming on a single sound would produce six rhyme pairs. Thus "die . . . testify . . . eye . . . justify" yields die|eye, die|justify, die|testify, eye|justify, eye|testify, justify|testify. There were very few cases in which there could be doubt over whether a rhyme was intentional or accidental. Imperfect rhymes were included when they occurred within a rhymed context or in places, such as gnomic lines or ends of speeches, where rhyme might be expected. The Micro-OCP concordance programme was then used to generate an alphabetical list of all the rhyme pairs and to tally the number of occurrences of each in Pericles and in Miseries. As rhymes from both plays were printed in the one list, with the total of ill|will, for example, in Pericles being followed by the total of ill|will

in Miseries, shared rhymes were visible almost at a glance. A list of rhymes shared by Pericles and Miseries is given in Table 1, which also shows the number of occurrences of these rhymes in each play and provides the Pericles line references. Where links are recorded the rhyme words appear in exactly the same form in each play, except that in Pericles the plural or verbal -s ending occurs in beds|heads, debts/gets, and swears|tears and the last rhyme listed is in the form thrived|wived. My intention was to treat -s endings, whether plurals or verbal inflexions, as equivalent to the root form, but to regard prefixes, suffixes, and more complex inflexions as establishing separate rhyme words. The reason for the slight relaxation of this rule in the case of thrives|wives or thrived|wived will be explained shortly. There were no other cases where relaxation of the rule would have produced additional links.

It is immediately obvious that although the total number of rhymes in Pericles 1-2 is not much larger than the total number of rhymes in Pericles 3-5, it is within acts 1-2 that the great majority of rhyme links with Miseries fall. Twenty-six of the Pericles line references are to the first two acts, only six to the last three. Also noticeable is that the two rhymes that occur most frequently in Miseries—and they are easily the most frequent not only in the list of shared rhymes but in Miseries as a whole—are also the only two shared rhymes that occur as often as three times in Pericles. But in order to determine the significance of the links with Miseries we need data concerning rhymes in all Shakespeare's works.

The necessary information is available in Helge Kökeritz's Shakespeare's Pronunciation (1953), which ends with an index to all the rhymes in Shakespeare's collected works, including The Two Noble Kinsmen and the largely spurious miscellany The Passionate Pilgrim. From Kökeritz's index it is clear that three of the rhyme pairs common to Miseries and Pericles occur nowhere else in the Shakespeare canon. These are consist|resist, him|sin, and impudence|offence. Fears|years is paralleled only by fear|half year in I Henry IV, 4.1.136, tears/years only by Venus and Adonis, 1091, both|oath only by The Rape of Lucrece, 569, and awe|law only by Richard III, 5.3.312; and the index records only two further instances of bed|wed, bring|sing, and do|you. Even examples of ill|will and life|wife are far from numerous: outside Pericles there are eleven of each.

However, a just assessment of the presence in Pericles 1-2 of so many rhymes that are also found in Miseries requires comparisons between Miseries and all Shakespeare's works. I therefore consulted Kökeritz's comprehensive inventory for citations of each individual rhyme in my alphabetical list of rhymes in Miseries so as to obtain counts of the links with the various Shakespearian plays and poems. "Links" were counted not merely in terms of the presence of a rhyme in both Miseries and the Shakespearian work but in terms of the number of occurrences of the rhyme in each work. Thus, to illustrate from the data in Table 1, since all|call appears once in Pericles and

twice in Miseries, two links are recorded, and since die|eye appears twice in Pericles and once in Miseries two links are recorded. The rhyme him|sin appears twice in each play, and so yields four links, while ill|will, which occurs three times in Pericles and six times in Miseries, yields as many as eighteen links. The same principles were used in calculating links with each play and poem.

Kökeritz normally groups minor inflexional and derivative forms with their basic words, agree(d,-s,-ing) rhyming not only with be|me, etc., but also with bleed, sees, seeing, and abhor (thee) rhyming with more, for thee, and adore thee. Thus a single head word thrive(d,-th) includes rhymes with thrive, such as thrive|wive in Twelfth Night, and rhymes with thriveth and thrived, such as thrived|wived in Pericles. However, when a rhyme word is exceedingly common, as are die and eye, Kökeritz gives a separate head word to even the -s plural or verbal inflexion, dies and eyes. In such cases I differentiated between die|eye rhymes and dies|eyes rhymes. In practice, it matters little what rules-of-thumb one adopts. The word kiss, for example, is given the head kiss(ed,-es,-ing, you), but the inflexions and the feminine rhyme with you are irrelevant to kiss|this and to most of the other examples. In the large majority of cases, "links" implied exact equivalence of the rhyme words, and most of the few exceptions involved the simple -s ending. It was easy enough to preserve Kökeritz's separate categorization of such homophones as knight and night, somewhat more difficult to be sure of consistently maintaining his distinctions between such different words as the verb and the noun tear. Again, multiple rhymes were broken down into their components, so that triple rhyme yielded three rhyme pairs, a series of four rhyming lines (whether consecutive or part of a more complex pattern) yielded six rhyme pairs, a string of five rhyming lines yielded ten rhyme pairs, and so on. When a particular rhyme word appeared twice or thrice within such a sequence calculations were made as though it appeared only once. Thus "holly . . . folly . . . holly . . . jolly" was taken as creating three rhyme pairs, not six. Two attempts to extract the relevant information, made over a year apart, resulted in sets of figures that were identical for most works and so closely similar for others that errors or inconsistencies in the application of my criteria must have been few and of no practical consequence.

Once we have figures for the number of rhyme links between George Wilkins's The Miseries of Enforced Marriage and each one of Shakespeare's plays and poems, we still need to calculate how frequent these links are in proportion to the number of possible links, that is to the total number of rhymes in each Shakespearian work. Unfortunately, it is not possible to use ready-made tallies of Shakespeare's rhymed lines, such as those presented in the metrical tables appended to the second volume of E. K. Chambers's William Shakespeare (1930) or those given by Frederic William Ness in The Use of Rhyme in Shakespeare's Plays (1941). Ness's totals are often seriously at variance with Chambers's, which themselves result from Chambers's efforts

to correct the tables of F. G. Fleay. Chambers's main table excludes "external" verses—prologues, epilogues, choruses, interludes, masques, Pistol's bombast, and the Mousetrap play within Hamlet, and it is not perfectly clear whether his supplementary figures are intended to make good all these omissions. More importantly, a simple halving of the number of rhymed lines in a play will not give the number of rhyme pairs, for the reasons indicated above. The Fool's jingle in King Lear, 1.4.131-136—"goest . . . showest . . . knowest . . . owest . . . trowest . . . throwest"—creates fifteen rhyme pairs. The three "b" lines in The Rape of Lucrece's ababbcc stanzas create three rhyme pairs, so that each seven-line stanza yields five rhymes.

It was therefore necessary to produce new figures for the number of rhymes in each Shakespearian play and poem by working systematically through Kökeritz's one hundred page inventory, and marking each example of a rhyme on graph paper in which the various Shakespearian titles were allotted separate columns. This proved a tedious and tricky task. It takes a while to familiarize oneself with the organization of Kökeritz's index. References to all instances of above|love, for example, are given under the head word above; later, under the head word love, the rhyme above is listed, but the references are not repeated. However, references for multiple rhymes ("fight . . . sprite . . . night") and for especially unusual rhymes (am|Tom) are sometimes given under the later alphabetical head or heads as well. So it requires vigilance to avoid counting some instances of a rhyme twice. Kökeritz, who comments on the inaccuracy of Ness's statistics, seems to have been careful himself, and he is much more cautious about admitting rhymes as intentional, though he includes some doubtful examples, marking them with an asterisk. I included these few doubtful examples in my counts. But I excluded rhymes peculiar to bad quarto texts, ignoring all those that Kökeritz marks with a Q, except those from King Lear and the single example from Richard II.

As a check on the accuracy of my counts, I calculated directly the number of rhymes in the Sonnets, A Lover's Complaint, Venus and Adonis, and The Rape of Lucrece. The Sonnets consist of 154 sonnets, each with seven rhymes, except that Sonnet 126 is a tail-piece with twelve lines and six rhymes and Sonnet 99 has an extra line connected with the opening quatrain and so yields nine rhymes. This gives 1079 rhymes. The total I compiled from Kökeritz's inventory was 1075. Kökeritz is responsible for at least part of this discrepancy. He has omitted the headword truth and so has no reference for the truth|youth rhyme in Sonnet 138, though he includes the rhyme (without reference) under the headword youth. A Lover's Complaint has 47 stanzas rhyming ababbcc, giving five rhymes in each and 235 overall. My total, compiled from Kökeritz's inventory, was 242. Venus and Adonis has 199 stanzas rhyming ababcc. This gives 597 rhymes, compared to my total of 603. The Rape of Lucrece has 265 stanzas, each with five rhymes. This gives 1325 rhymes altogether, compared to my total of 1338. Presumably my

counts for the three narrative poems accidentally included some rhymes twice, though Kökeritz's index may itself contain pertinent minor mistakes. I corrected my original tallies for the poems in accord with the later calculations, but the main point of these checks was to provide assurance that the processing of the data available in Kökeritz's inventory had been tolerably reliable. We can reckon on a margin of error of not much more than 1.0 per cent. My total for the much shorter The Phoenix and Turtle proved exactly right. The poem contains thirteen quatrains (abab), followed by five rhymed tercets, but in one of these the lines end "be . . . she . . . be", giving one rhyme rather than three. The total of 39 rhymes agrees with the figure extracted from Kökeritz's data.

An additional step in the investigation was to compile, with the help of Micro-OCP, an alphabetical list of all the rhymes in those sections of The Travels of the Three English Brothers which have been assigned to George Wilkins.[12] The play, another belonging to the King's Men, was printed in 1607 with a dedicatory preface to which were appended the names of John Day, William Rowley, and George Wilkins, and there has been a fair measure of agreement among scholars upon which scenes were composed by Wilkins. The list of Travels rhymes was then compared with the list of Miseries rhymes, so that the number of links could be ascertained. Travels was thus treated like each of Shakespeare's works, in that the proportion of its rhymes that were shared with Miseries was calculated.

Table 2 presents for each Shakespearian work and for Wilkins's putative share of Travels (a) the total number of rhymes, (b) the number of rhyme links with The Miseries of Enforced Marriage, and (c) the number of rhyme links with Miseries expressed as a percentage of the total number of rhymes (that is, total-b divided by total-a and multiplied by 100). The plays are ordered according to the size of this percentage. Pericles 1-2 is treated separately from Pericles 3-5.

Pericles 1-2 heads the list with a percentage of rhyme links with Miseries that is almost double that of the highest authentic work by Shakespeare. This means that Pericles 1-2 is far more "Wilkinsian" in its rhymes—when "Wilkinsian" is defined in terms of the one play of George Wilkins's undoubted sole authorship—than any of the other plays or poems. This is a remarkable result considering the rather small number of rhymes in some of the plays, which would produce considerable chance variation in the percentages:

in a play without many rhymes, a few instances of Miseries favourites such as ill|will and life|wife would have given a high percentage of links.

The results are very difficult to explain without resort to a theory of dual authorship. Pericles 3-5, in striking contrast to Pericles 1-2, falls about midway on the list, exhibiting roughly the same proportion of rhyme links with Wilkins's The Miseries of Enforced Marriage as is exhibited by such other late romances as The Winter's Tale and Cymbeline. The notion that acts 1 and 2 of Pericles are early Shakespearian work receives no support from Table 2. There is no tendency for Shakespeare's earliest plays to display, in their use of rhyme, the same association with Miseries that is evident in Pericles 1-2. Nor does the theory that the style of Pericles 1-2 results from Shakespeare's conscious attempts at archaism provide a satisfactory explanation of the data. If attempts at archaism were responsible for the prevalence of rhyme links with Miseries, the links should be predominantly with the Gower choruses, which deliberately imitate Gower's medieval rhymes. But of the 58 rhyme links with Miseries in Pericles 1-2, 46 occur within the dialogue of Pericles 1-2 with its 100 rhymes, and only 12 within the Gower choruses to Pericles 1-2 with their 45 rhymes: the percentage of Miseries links with Gower choruses 1 and 2, namely 26.7, is higher than for any of the Shakespearian works, but the percentage of Miseries links with the dialogue of Pericles 1-2 rises as high as 46. In Pericles 3-5, in contrast, neither the Gower choruses nor the dialogue are significantly linked with Miseries in their use of rhyme. And of course the author of The Miseries of Enforced Marriage had no reason for aiming at archaism within that particular play, whose very success was due to its topical subject matter, the abuse of the powers of a guardian over his ward in the making of a marriage settlement.

An indication that the rhyme links between Pericles 1-2 and Miseries are due to common authorship is afforded by the results for Wilkins's share of Travels, for which the percentage of links with Miseries is higher than for any of the Shakespeare works. It can hardly be coincidental that those sections of Pericles and Travels which have often been attributed to Wilkins are both more "Wilkinsian" in their rhymes than the sections of Pericles generally accepted as Shakespearian or than any other of Shakespeare's works.

In themselves the results are perhaps less decisive for Wilkins's authorship of Pericles 1-2 than for the dual authorship theory generally. Dramatists other than Wilkins may have had an unShakespearian liking for Wilkinsian rhymes. Shakespeare appears to have favoured variety in his rhymes, and the high proportion of links between Pericles 1-2 and Miseries may have arisen because the rhyming in both these pieces of dramatic writing tends towards the commonplace. There are, however, less predictable links, such as consist|resist, him|sin, and impudence|offence, and this new evidence has to be considered in connection with all the other internal evidence that has been offered in support of the case for Wilkins's authorship of Pericles 1-2.

David J. Lake showed that in its rates of use of certain categories of

rhyme and of rhyming "rifts" and "rafts" Pericles 1-2 was more like Wilkins's verse than the undisputed verse of Shakespeare or the verse of Day, Heywood, or Rowley. M.W.A. Smith, using a computer programme to compare different plays in terms of rates of occurrence of the most common words, and strictly adhering to preset rules, has demonstrated that Pericles 1-2 is more like the acknowledged work of George Wilkins than that of George Chapman, Thomas Middleton, Ben Jonson, John Webster, Cyril Tourneur, or William Shakespeare. Smith validates his methods by showing that they can pick the correct authors of plays by Webster and Jonson, for example, when these are treated as anonymous. For the purposes of his tests, "Shakespeare" is defined as the author of Cymbeline and The Winter's Tale. Smith has not yet compared Pericles 1-2 with early Shakespeare plays. But my own statistical analysis of selected function words, which compares Pericles 1-2 and Pericles 3-5 with samples from 112 plays by twenty playwrights and represents early, middle, and late Shakespeare, shows that whereas samples by Shakespeare most closely match Pericles 3-5, samples by Wilkins most closely match Pericles 1-2. In another article I have shown that certain uses of the word "which" occur in Pericles 1-2 at rates more closely paralleled in Wilkins's work than in any play by Shakespeare or in thirty plays by nineteen other dramatists of the period. And as early as 1919 H. Dugdale Sykes had made out a case for Wilkins's responsibility for Pericles 1-2 which, though methodologically defective, was suggestive enough to persuade the Arden editor, F. D. Hoeniger, in 1963 that Wilkins probably contributed to the play. Sykes pointed especially to conspicuous elision of the relative pronoun as a feature of Wilkins's style, and it is notable that all sixteen examples in The Travels of the Three English Brothers fall within the three-fifths of the play assigned to Wilkins, as do nine of the ten examples of unusual "which".[13]

Eric Sams has argued that Wilkins, an inveterate plagiarist, must have imitated Shakespeare,[14] and certainly Wilkins's prose narrative The Painful Adventures of Pericles Prince of Tyre (1608) follows the plot of the extant Pericles play and often echoes its wording. But a theory that Wilkins in Miseries and Travels imitated the style of Pericles 1-2 does not even begin to explain why Wilkins should have modelled his verse on those particular acts of the play or why they should be so notably unlike any other section of the Shakespeare canon. In Pericles we have a play whose omission from the First Folio encourages the suspicion that it is not wholly Shakespearian, and a play, moreover, which divides stylistically into two parts. The natural conclusion is that the part which sounds like Shakespeare's is his, and that

somebody else wrote the rest. At present the evidence favours the identification of that second "somebody" as George Wilkins.[15]

Table 1 Rhymes shared by Pericles and The Miseries of Enforced Marriage.

                                                       

Rhyme	Pericles Reference	Frequency:
		Pericles/Miseries
above/love	2.3.21	1/1
all/call	5.1.246	1/2
awe/law	1.P.35	1/1
bed/head	2.3.97	1/1
bed/wed	2.5.92	1/2
begin/sin	1.P.30	1/1
be/me	1.2.108	1/2
both/oath	1.2.119	1/1
breath/death	2.1.7	1/1
bring/sing	1.P.13	1/1
buried/dead	2.1.80	1/1
consist/resist	1.4.84	1/1
debt/get	4.P.33	1/1
die/eye	1.P.39, 1.1.32	2/1
do/you	1.1.50	1/1
fear/here	1.4.79	1/1
fears/years	1.2.84	1/1
go/woe	3.P.42	1/1
him/sin	1.2.39, 2.P.23	2/2
ill/will	1.1.103, 2.1.140, 2.1.171	3/6
impudence/offence	2.3.68	1/1
kiss/this	1.2.77	1/4
life/wife	1.P.37, 1.4.45, 5.1.245	3/6
swear/tear	4.4.27	1/2
tears/years	1.4.18	1/1
thrives/wives	5.2.10	1/1

Table 2 Shakespearian plays and poems and George Wilkins's share of The Travels of the Three English Brothers listed in order of the number of rhyme links with Wilkins's The Miseries of Enforced Marriage, when links are expressed as a percentage of the total number of rhymes in the play or poem.

                                                                                             

Title	Rhymes	Links	Percentage
Pericles 1-2	145	58	40.0
Travels of . . . English Brothers	232	62	26.7
Measure for Measure	46	11	23.9
Romeo and Juliet	254	55	21.7
All's Well That Ends Well	144	31	21.5
The Comedy of Errors	210	45	21.4
2 Henry VI	61	12	19.7
Twelfth Night	83	16	19.3
Julius Caesar	21	4	19.0
I Henry VI	176	28	16.0
Othello	57	9	15.8
Henry V	53	8	15.1
Richard II	283	41	14.5
Timon of Athens	80	11	13.8
The Two Gentlemen of Verona	74	10	13.5
Sonnets	1079	140	13.0
Much Ado About Nothing	39	5	12.8
Troilus and Cressida	109	13	11.9
The Winter's Tale	53	6	11.3
The Rape of Lucrece	1325	139	10.5
Macbeth	144	15	10.4
Pericles 3-5	128	13	10.2
As You Like It	143	14	9.8
Titus Andronicus	73	7	9.6
Hamlet	117	11	9.4
Love's Labour's Lost	624	57	9.1
The Passionate Pilgrim	181	16	8.8
Cymbeline	102	9	8.8
2 Henry IV	46	4	8.7
The Merchant of Venice	162	14	8.6
Richard III	94	8	8.5
King John	85	7	8.2
Venus and Adonis	597	49	8.2
A Midsummer Night's Dream	567	39	7.4
The Tempest	98	7	7.1
3 Henry VI	86	6	7.0
The Two Noble Kinsmen	89	5	5.6
A Lover's Complaint	235	13	5.5
King Lear	130	7	5.4
Coriolanus	19	1	5.3
The Phoenix and Turtle	39	2	5.1
The Merry Wives of Windsor	65	3	4.6
I Henry IV	31	1	3.2
The Taming of the Shrew	97	3	3.1
Henry VIII	43	1	2.3
Antony and Cleopatra	28	0	0.0