Historical evidence and the authorship of Pericles I-II

First version

Previously unpublished; © the estate of eric sams

The play of Pericles was registered by Sir George Buck and Warden Seton, together with Antony and Cleopatra, in May 1608. The latter remained unpublished until the 1623 Folio; but Pericles was duly printed in a Quarto text (henceforth Q1) dated 1609. Its story, found in John Gower's long poem Confessio Amantis (1393), had already been retold in prose by Laurence Twine as The Patterne of Painefull Adventres, published c. 1576 and reprinted in 1607. A novella by George Wilkins (henceforth PA), avowedly derived from the play but also using Twine's title and much of his text, was published in 1608.

     So Wilkins was a known plagiarist, remembered solely for his borrowings. Yet many specialists believe that he actually wrote the first two acts of Q1, because of their occasional parallels with his few other works and the priority of his PA. In this interest, the principle of economy is set aside, the obvious explanation of plagiarism is ignored, and primacy is promoted to evidence of authorship.

    But what if Q1 is indeed all by Shakespeare? Its distinguished registrars surely intended to authorise the publication of two genuine Shakespeare plays, as performed by their sovereign's theatre company and written by its leading playwright. The title-page of Q1 carries the confirmation “as it hath been diuers and sundry times acted by his Maiesties Seruants, at the Globe on the Banck-side” and the clinching affidavit “By William Shakespeare”. The same assurances recur in five more Quarto editions, dated 1609, 1611, 1619, 1630 and 1636. It is true that Q1 was in fact published by Henry Gosson, not Edward Blount, to whom it had originally been assigned; and that its inclusion among the collected works was delayed until the second issue of the Third Folio in 1664. But neither of these facts offers any evidence for the part-authorship of Wilkins; and the words of the Third Folio (“written by William Shakespeare and published in his lifetime”) tell heavily against any such notion.

    There is further corroboration in the use of John Gower himself as the play's presenter passim. Gower lies entombed and in effigy, to this day, at Southwark Cathedral near the Globe theatre. There, on the last day of 1607, Shakespeare's brother Edmund, himself a player, had been buried with rites including “a forenoone knell of ye great bell'”, all no doubt paid for by the affluent actor-playwright. There is no known connection between Gower and Wilkins, whose PA never even hints that he wrote any part of Pericles. On the contrary, PA praises the idea of John Gower's phoenix-like rebirth from his own ashes to introduce his old tale of the Prince of Tyre newly arranged for the stage. Further, such effects of resurrection from the dead became a favourite Shakespearean theme, which recurs in Pericles as in his later plays. Wilkins also confirms all six Quartos' assurances that Pericleshad been acted by the King's players, at the “Globe on the Bancke-side”. Those players may well have been colleagues of brother Edmund and fellow-mourners at his funeral. If indeed the first two acts of Pericles are inferior (a question often left open in order to admit Wilkins) the simple explanation was long ago provided by Dryden in his published introduction to a play by the son of William Davenant; Pericles  too had begun as an early effort by a young artist, who would take time to grow and develop. Dryden staked his credit on this information, about which he was well placed to know; his friend Davenant was Shakespeare's godson. Nobody knows better, though many pretend to. Other revising artists, such as Brahms and Schubert, rewrote parts of their early works; other testimony places Pericles as partly an early play. But all this is bundled aside in favour of the confident nomination of Wilkins as collaborator, a theory announced as fact by M.W.A. Smith [1] and by the Oxford editors Gary Taylor and MacD. Jackson [2] who add the hypothesis that Pericles is a “memorial reconstruction by actors”, despite the absence of any text it could have been reconstructed from.

    There the question has rested until the recent Cambridge edition of Pericles [3] publicly rejected both “Wilkins” and “memorial reconstruction”. Its reading list includes some of Smith's many articles, but his various methods and results are left unmentioned. Its editors deal with Taylor by pointing out that his “function-word” test, which allegedly rules out Shakespearean authorship of the first two acts, merely assumes that different styles must mean different writers. But Jackson's 1993 study of rhymes in Pericles is singled out for special praise as “perhaps the most persuasive case so far made for Wilkins” and its conclusions are said to be “fairly and moderately stated”. However, the Cambridge editors reject this case too, as “not beyond the bounds of coincidence”. They further reject all the “undeniably persuasive evidence so far collected that the composition ofPericles was shared or that the secret sharer was George Wilkins”.

    But such unargued rejection is surely just as irrational as tame acceptance. What is truly persuasive cannot also be treated as unpersuasive. Secondly, dissenters need logical counter-evidence. Conversely, those who rely on statistics should have researched the textual data before publishing their purported proofs that the notorious plagiarist and criminal Wilkins [4] had been promoted to the position of Shakespeare's highly privileged yet uncharacteristically modest and silent collaborator. The present article therefore attempts those taxing tasks, for the first time on record. Instead of analysing individual “function-words” it relies on a conspectus of spelling, capitalisation, apostrophes, and punctuation, together with the principle of economy in attributing these differences to actual authors, not to hypothetical intervention.

    It begins by a direct comparison between the actual Wilkins, i.e. the first six chapters (henceforth A) of PA 1608 [5] and and the purported “Wilkins?'”, i.e. the first two acts (henceforth B) of Q1 1609. [6] Those two texts share much the same length (9-10,000 words); and they recount much of the same story of Pericles before he put to sea with his wife Thaysa. But are they really both from the same hand? If so, how do we explain the hundreds of differences between A and B - and the equally frequent similarities between B and the undisputedly Shakespearean last three acts of Q1 (henceforth C, also about the same length as A and B)? Each of these plain facts serves to disprove any connection of authorship between A and B.

    The following conspectus tabulates passages deliberately borrowed from B and displayed in A by the latter's author (in that order, with line-references for verification).




Citie...Seat (I.18)

accompted (14.19)

faythfulnesse…courage (I.i.63)

Viper…feed (64)

flesh which…breed: (65)

Husband…labour, (66)

in a Father; (67)

milde; (68)

I,…child (69)

they…liue (70-1)

fourtie…Sonne (116,118)

courtesie…hipocrite (121-2)

Father…Sonne…Child…Mothers (127-8,130)

Thaliard [four times] Thali [twice] (151f)

Poyson…Gold (155)

buryall (I.iv.49)

ships hitherward (61)

Ships..Corne (92-5)

Statue (II.14)

Helican...Thaliart...murdred... Tharsis (II.17,23-5)

tir'd..a shore (II. 37-8)

Fisher-men..Nets (II.i.11,13)

fenny...Subject..Sea (II.i.52)

Fish..Net (II.i.122)

law…Armour (II.i.1245)

Thaisa…Thai. (II.ii.0,8,16,18,29)

device…Shield…Ethyope (II.ii.19, 20) 

deuice..Shield..Armed (II.ii.24, 25)

habit (II.ii.57)

Clouds..Gold, that's..Touch-stone II.ii.37)

Ships..shipwracke..shore (II.iii.84)

Hellicanus…Hell...Escanes [twice] (II.iv.1-2)

Chariot..bodyes..lothing.. ador'd (8,10-11)

vnworthy..Scholemaister (II.v.40)

displeasure…traytor [thrice] (54-5)

Honour…Rebell…her state (II.v.61-2)



Cittie...seate (11,19-21)

account'd (I.30)

Faithfulnesse…Courage (16.7)

viper…feede (16.19)

Flesh, that…breede; (16.20)

husband…labour (16.21)

from a father: (16.22)

milde, (16.23)

I…Childe: (16.24)

this…liue, (16.25,26)

forty…sonne (17.29,33)

curtesie…hypocrisie (18.11,12)

father...sonne…child…mothers (18.7-9)

Thalyart [thrice] (18.25,27,36)

gold…poyson (18.28)

buriall (22.17)

shippes thither ward (22.26)

shippes…corne (23.12,14)

statue (24.11)

Helycanus...Tharsus...Thalyart...murther (24.20-5)

tyred..on the shore (26.23,26)

Fishermen…nettes (26.30,33)

finny…subiects…sea (27.10)

fish…net (28.30,31)

Lawe…armour (28.30,34)

Thaysa (25.6,29.14,24,31.10)

Device…shield…Ethiope (30.1,2)

Device..shield..armed (30.8. 11)

habite (30.35)

cloude…gold thats...touchstone (30.15-16)

shippes…shipwrecke shoare…(32.7-8)

Helycanus..Eschines (33.31,34.1)

Charriot..bodies..adored..loathd (33.4,17,20-1)

vnwoorthy…schoolemaister (38.12-13)

displeasure...traitour...traytour [twice] 38.30,32-3)

honour…rebell…his State (39.5-6)


    This selection could easily be extended. But how can it be explained? Not, surely (pace Academia) by ignoring orthography altogether, or by assuming that the many necessary changes were obligingly effected by some unknown person. No; the sex-change and demotion of the goddess Honour, together with all the other discrepancies, such as the quite different spellings of character-names (including Dyonysa and Symonides, as consistently preferred in A, as against Dioniza/Dyoniza or Simonides/Simonydes in B) point much more plainly at two different people. The same applies to differences of plot or action. Thus when Pericles arrives to sue for the hand of Antiochus's daughter, A avers that “Antiochus then first beganne to perswade him from the enterprise and to discourage him from his proceedings by showing him the frightful heads of the former Princes”, whereas in B Antiochus begins by praising his daughter's perfection. Again, A makes the pageant at the court of King Symonides begin with “a prince of Macedon” and continue with “a Prince of Corinth”, although the sequence adopted by B is first “a knight of Sparta” and then “a Prince of Macedon” Further, A bestows the motto of B's second knight on his own third contestant, and conversely; similarly with the fourth and fifth. In B, one grave counsellor confides the incest story to the other before, but in A after, the account of heaven's dust punishment. Again the list could be prolonged. All such textual data plainly imply that Wilkins had seen and sometimes remembered Pericles but had not written any part of it.

    This is further confirmed by A's visual recollections; thus it describes (22.22f) describes the entry of “a fainting messenger” who “came slowly in to them, his fearefull lookes described that he brought sorrow, and in slowe wordes hoe deliuered this”, which clearly recalls a stage performance, as indeed A says, in terms. It follows that A may contain accurate recollections of B as spoken and presented on the stage; so the Oxford editors are right in accepting this possibility and the Cambridge editors are mistaken in rejecting it. But memory, however reliable, remains a world away from authorship.

    The table above provides another pointer to two different authors. Abstract nouns, such as Faithfulness and Courage, are capitalised in A, whereas B has faythfulnes, faythfulnesse and (three times) courage. The same applies to other shared abstractions and generalities; thus A has Absence, Affection, Charity, Desire, Iudgement, Perfection, Presence, Rule, Vengeance, as compared with B's absence, affections, charity, desire, iudgement, perfections, presence, rule, vengeance. Predictably, A also has many more such words than B (e.g. Agents, Argument, Chalenges, Communication, Coniunction, Enigma, Grauity, Impression, Incumbrance, Interiections, Ioy, Madness, Maturity, Modesty, Nobilitie, Preferment, Zeale). The inference is that A and B were set up from authorial copy in two different hands, one of which was much more given than the other to the use and capitalisation of abstract concepts. This can be confirmed from Wilkins's extant play The Miseries of Inforst Mariage; like A, but unlike B and C, it teems with just such capitalised words.

    Shakespeare's C of course follows his practice of capitalising the names and products of flora and fauna. The same system appears throughout B. But such words are far rarer in A; and the examples that A shares with B (corne, flesh, horse, serpents, viper) remain uncapitalised. So again B and C are akin, and A is not.

    Again, the table above, which shows four apostrophes in B and none in A, calls for a complete comparison. The result is that B has over 100 examples, and A has just one, in the word “lou'd”. Not even the most convinced Wilkinsite, surely, will postulate an unevidenced agent who removed a hundred apostrophes from A or introduced the same amount into B; the simple explanation is that A always avoided them, unlike B, which again makes their authors two different people. As it happens, we can check this hypothesis too; Wilkins also avoids apostrophes in his play The Miseries of  Inforst Mariage (1607), with such spellings as callt (=call it), doot (=do it), dunt (=done it), hees (=he is), sheele (=she will) and so forth. So Wilkins cannot have written B, which reads (for example) “hee's” twelve times.

     Apostrophes are often used in B and C (though never in A) to form past tenses, for example when -de is used instead of modern -ed. Diligent search discloses only three examples of this usage throughout A: bribde, subdude, tride. But B has sixteen of them, and C has eighteen. Similarly A has only ten past tenses ending in Shakespeare's favourite -t, compared with B's twenty-two and C's nineteen. In both instances, A is the odd text out.

    We can also compare the first Quarto of King Lear, which like A was printed for Nathaniel Butter in 1608. This edition of Lear was miscalled a “memorial reconstruction by actors” for most of last century. It shares with Pericles Q1 the imprimatur of Sir George Buck, a title-page attribution to Shakespeare and his company, and evidence (such as copious apostrophes and blank verse misprinted as prose) of being set up from authorial foul papers. Who could have penned such copy but Shakespeare himself, and why should not the same apply to the contemporary Pericles Q1? If so, its B and C would both be by Shakespeare, not Wilkins.

    Many other resemblances or differences of spellings contradict the contrary hypothesis. Thus several common abbreviations are found in both B and C but not in A, such as ath' or a' th (on the, of the), doo't (=do it), heer's (here is), in't (in it), on't (on it, of it), shee'le (she will), there's (there is), 'you're' (you are). Of course it might be maintained that such abbreviated forms are unsuitable for the prose of A; but they are rare or non-existent in the play of Miseries also. Yet again the conclusion is that its author, Wilkins, wrote A but had no hand in B.

    Further, we know from Sir Thomas More that Shakespeare could write “together” as “togeather”, and this spelling recurs in C, the Shakespearean portion of Pericles. So it does in B. But A has “together”. Other spellings and locutions shared by both B and C include anie, daunger, doo, fayth, goe, hie (=high), manie, melancholie, scholler, stop, Syr, vsde (=used), woondrous; but these are not found in A, which has any, danger, do, faith, go, high, many, melancholie, Scholler, stoppe, Sir, vsed, wonder. So A and B had different authors.

    Here is a selection of other equally surprising differences between B and A (in that order, separated by a slash): approue/approoue, behoulding/beholding, Branch/braunch, buylding/building, buylt/built, choyce/choice, denyed/denied, deuowre/devour, dryuing/driuing, Ethyope/Ethiope, fayth/faith, fourtie/forty, furnisht/furnished, gard/guard, harmonie/harmony, ill/Ill, maryage/mariage, nightie/nighty, miserie/misery, mooude/mooued, pouertie/pouerty, ritches/riches, sayle/saile, Scholemaister/Schoolemaister, secrecie/secresie, shold/should, solicit/sollicited, Sommer/summer, speach/speech, stayde/stayed, tombe/intoombed, Tryumph/Triumph, vnworthy/vnwoorthy, vsde/vsed, watry/watery, wayte/waiteth, worth/woorth. Such evidence again suggests that B was not written by the author of A; furthermore, these two different people are prima facie of different ages, because B (which has tch for modern ch, oul for modern ol, y for modern i, and so forth) is generally the more old-fashioned of the two. But C resembles B in this respect as in so many others; yet again, A is the odd text out.

    In sum: if A and B had the same author he had decided, for unknown and unexplained reasons, to rewrite his own text in a wholly different style. And indeed the Oxford editors claim that Wilkins, though allegedly a part-author of Pericles, need never have possessed his own copy of his own contribution. However, there is no evidence for any such hypothesis, and no reason for believing it; nor does it in any way support the theory of collaboration. This article argues that the textual facts are in need of economical and evidenced explanation, not speculation.



[1] such as “Counting Wilkins In: Stylometry Reveals Who Wrote Acts I and II of Pericles”, The Shakespeare Newsletter ccviii (Winter, 1990) 60.

[2] G. Taylor and MacD. Jackson, Pericles, in William Shakespeare: A Textual  Companion, Clarendon Press, Oxford, '1987' (recte 1988), 556-92.

[3] Pericles, ed. D. DelVecchio and A. Hammond, The New Cambridge  Shakespeare, Cambridge University Press, 1998.

[4] see Roger Prior, “The Life of George Wilkins”, Shakespeare Survey 25 (1972), 137-52.

[5] Wilkins, G., The Painfull Adventures of Pericles Prime of Tyre, 1608, reprinted ed. T. Mommsen, Oldenburg 1857 (also ed. K. Muir, Liverpool 1953).

[6] Pericles 1609, facsimile ed. W. Greg, London 1940.