The Authorship of Shakespeare's All's Well that Ends Well

Laurie Maguire and Emma Smith

'Many Hands - A New Shakespeare Collaboration?' [PDF]

TLS April 19th 2012


[Full text]

The Folio of 1623 is the only authority for Shakespeare’s comedy All’s Well that Ends Well (whose composition is traditionally dated 1602-03). The textual problems in this Folio play are well documented. Although editors uniformly posit authorial ‘foul papers’ as the source for F All’s Well, they have to acknowledge (as Gary Taylor reminds us) the ‘paucity of Shakespearian spellings’ and an ‘unusually high number of errors and cruces’.[1]The variable speech-prefixes, cited since McKerrow as proof of underlying authorial copy, behave oddly in this play, often changing mid-scene: in Act 1, scene 3, for example, the Countess of Rossillion is Cou. or Coun., then Old. Cou. before reverting to Cou.; in 3.2 she is Count. then Lad.  (with one variant Old La.).  These within-scene fluctuations are unlike apparent parallels in  Romeo and Juliet or Love’s Labour’s Lost where different speech-prefixes in different scenes  make it possible to argue for Shakespeare’s scene-specific view of a character’s identity (Lady Capulet as maternal when identified as Mother, Armado as boastful when identified as Braggart).

The naming of characters in All’s Well is bumpy throughout.  Two French lords go to Rossillion as a pair to deliver news of Bertram’s military destination (Act 3, scene 2), appear with Bertram in the military camp as a duo (Act 3, scene 6) but then separate, one to gull Paroles, the other to escort Bertram on his assignation with Diana. They are variously designated 1 Lo. G. and 2 Lo. E., Fren. E. and Fren. G., Cap. E. and Cap. G., eventually being named in the dialogue in Act 4, scene 3 as the brothers Dumaine. (Harold Brooks saw in this an ontological evolution, with Shakespeare developing named characters out of types.[2]) The ‘1' and ‘2' speech-prefixes do not correspond consistently with ‘E.’ and ‘G.’, nor do they correspond consistently with the plot,  for both 1 and 2 are identified in different scenes as the Lord who accompanies Bertram to Diana. Elsewhere Steward becomes, briefly and locally, Rynaldo (TLN 1575, 1587) and a character called Violenta is given an entrance (TLN 1604); she may be a ghost character or an alternative name for Diana, whose subsequent stage direction Enter Bertram, and the Maide called Diana. (TLN 2017-18)is oddly emphatic.

Elsewhere in All’s Well stage directions are curiously narrative, explanatory, and summary. The most distinctive is that at TLN 1089-90 (Paroles and Lafew stay behind, commenting of this wedding) but analogous directions abound: Enter the King with divers yong Lords, taking leave for the Florentine warre (TLN 594-5); Enter Count Rossillion and the Frenchmen, as at first (TLN 1730-31); Enter one of the Frenchmen, with five or sixe other souldiers in ambush (TLN 1911-12).  (These may indicate the playwright’s use of his ‘plot’ in scripting his scenes.)  One highly specific stage direction is misplaced. She addresses her to a Lord (957) prefaces Helen’s general speech to the lords about having ‘restor’d the king to health’; it is fifteen lines before  she addresses a lord (‘Sir, wil you heare my suite’; TLN 972). Fredson Bowers thought the intervening material a later addition.[3]Stage directions occasionally designate characters in misleading ways. A Messenger is instructed to enter at TLN 2181 but speaks as a servant (his speech-prefix is Ser.); a gentle Astringer who enters at TLN 2601 continues to perplex editors.

Other textual problems abound. Short phrases are sometimes given to the wrong character. Bertram’s ‘hush, hush’ at the end of his line TLN 2226 should begin the speech of the French Lord which follows at TLN 2227. Similarly, ‘all’s one to him’ at the end of Paroles’ TLN 2245 should be the beginning of Bertram’s speech at TLN 2246.  Characters are sometimes given two consecutive speeches. The Countess purposes to interrogate the Clown:  ‘I will bee a foole in question, hoping to bee the wiser by your answer’ (TLN 861-3, with the speech-prefix Lady.). She immediately, without interruption, questions the clown but the text gives her a new indented speech and a duplicate speech-prefix (‘La.’) The Clown has two consecutive speeches at TLN 1243 and 1245. (For both these examples critics postulate an omitted speech in between.)

Gary Taylor identifies the MS underlying Folio All’s Well as ‘difficult’, ‘really foul autograph’.[4]Some of the oddities described above can be paralleled in other Folio plays believed to derive from authorial papers; but what is unusual in All’s Well is the number, variety, and inconsistency of problems. 


Added to All’s Well’s textual anomalies are a number of unusual stylistic features. The play has an unusually high percentage of rhyming couplets. Couplets form 19% of the play (Shakespeare’s Jacobean norm for rhyme, according to the Oxford Textual Companion, is 5%).[5]  All’s Well also has a notable quantity of hypermetrical blank verse lines with feminine endings.  (The play is only 55% verse, with 8% of the verse lines having feminine endings.) Blank verse lines often conclude with multisyllabic endings (tri-syllabic and even tetra-syllabic words or phrases are not unusual; some of these multisyllabic endings are hypermetrical, some are not).

The recent redating of All’s Well from 1602-03 to 1606-07 (or later)has gone some way to resolving some of the play’s stylistic anomalies. In 2001 MacDonald Jackson explored the unusual name of the character Spurio.[6] Spurio is named twice (TLN 643 and 2267) although he does not appear in the play. The name occurs only once elsewhere in Renaissance drama, in Middleton’s Revenger’s Tragedy (Q 1607) where it makes symbolically appropriate sense as the name of the Duke’s bastard (spurious) son. This led Jackson to conclude that the name originated with Middleton and was picked up by Shakespeare.  There may even be an intertextual joke: although Spurio is a character in Revenger’s Tragedy, he is never named in the dialogue (his name occurs only in stage directions); Shakespeare’s Spurio, on the other hand, is named in the dialogue but is not a character in the play.   This ludic inversion indicates Shakespeare’s familiarity with a written text of Revenger’s Tragedy, either in manuscript form (the play was written for the King’s Men at a time when Shakespeare and Middleton were collaborating on Timon of Athens, probably in early 1606) or in its first printed edition of 1607/08.


When Wells and Taylor edited the Complete Works of Shakespeare in 1986, they were unsure about the precise date of All’s Well; in the Textual Companion they offered a span of 1604-07. In an article in 2001, Taylor said he was now inclined to put it at the upper end of the spectrum.[7] The revised edition of the Wells-Taylor Oxford Complete Works  (2005) redates the play as 1606 or early 1607. Lois Potter’s new biography of Shakespeare (2012) includes All’s Well in a chapter on the years 1607-09;[8]Gordon McMullan claims it as one of the chronologically late plays in his article ‘What is a late play?’[9]Clearly Jackson’s post-1606 dating is winning widespread acceptance.

 The re-dating is important because it makes All’s Well not a misfit problem play but an early ‘late romance’. Literary critics have long recognised the connection betweenAll’s Well and the late romances. All’s Well begins where Pericles ends –   with a maiden healing the sick patriarch –  and ends where The Winter’s Tale begins – with the legible body of a pregnant wife. The resurrection of Helen at the end of All’s Well is analogous to that of Hermione at the end of Winter’s Tale, as is the language of grace and wonder when Helen restores life to the King. Joel Altman devotes an entire chapter of The Improbability of ‘Othello’ (2010) to the parallels between All’s Well and the late plays. The superficiality of characterization in All’s Well which has concerned many critics makes sense in the context of the late romances. The requirement for cornets in All’s Well’s stage directions, evidence for most critics of a Blackfriars revival in 1609, may even point to a date of 1609 for the play’s original composition. Certainly Jackson and Taylor’s new terminus a quo of 1606/1607 opens up All’s Well to further investigation. We need to reconsider the play in the context of what Shakespeare was writing at that time, and with whom.

This returns us to the textual problems with which we began. All attempts to explain the textual anomalies invoke doubleness of some kind: a double time scheme (Shakespeare was interrupted, returning to the play after an accidental break or he revised it, deliberately, later (these alternatives are both Fredson Bowers’ suggestions)[10]); or it was later adapted; or it was marked up for revival (incorporating act divisions for Blackfriars and directions for cornets, with E. and G. being shorthand designations for actors’ names). Duality has been used as a more general framework for understanding All’s Well; M.C. Bradbrook suggested in 1950 that the play ‘might well have as its sub-title “Two Plays in One”’ and the Janus term ‘tragi-comedy’ features prominently in recent critical commentary.[11] We suggest a different kind of doubleness as an explanation: collaboration. We believe that a second hand can be detected in this play – and suggest that this might fruitfully be investigated as the hand of Thomas Middleton – and that this suggestion can help us to understand both the play’s textual anomalies and its particular stylistic and dramaturgical qualities. Furthermore, like John Jowett, whose edition of Timon of Athens (2004)is a model for rethinking Shakespeare’s interactions with Middleton, we view collaborative authorship not as a way of palming off an unloved play on someone else, but rather as one way to re-engage with All’s Well on different terms and, in Jowett’s terms, ‘to turn the disconcertingly un-Shakespearian into the relishably Middletonian’.[12]


Thanks to work by MacDonald P.  Jackson, David J. Lake, and the general editors of the Oxford Middleton, Gary Taylor and John Lavagnino, we know more about Middleton’s stylistic fingerprint than any other early modern dramatist, more about his verbal and prosodic preferences as well as more about his thematic and ideational ones.[13] In short, these idiosyncratic ‘Middleton markers’ make his language easily recognizable. Much of this evidence can be brought helpfully to bear on the textual problems and anomalies that have long been identified in All’s Well.

Firstly, rhyme. We have already noted that the high proportion of rhyming lines in All’s Well is unusual for Jacobean Shakespeare; it is standard for Middleton. In fact, the 19% rhyme in All’s Well accords with Middleton’s norm of 20% in his contribution to Timon of Athens (where rhyme in Shakespeare’s part accounts for only 4%, which accords with his Jacobean average of 5%). Similarly, the feminine endings identified as one of All’s Well’s prominent stylistic features, and its fondness for tri- and tetra-syllabic endings, are associated more with the style of Middleton than that of Shakespeare. R.H. Barker gives a figure of 25% feminine endings in Middleton’s earlier plays such as The Phoenix (1603-4) and Michaelmas Term (1604) rising to around 45% for No Wit/No Help (1611)and Chaste Maid (1613).[14]Both these features have been established as important elements of Middletonian co-authorship in Timon of Athens.[15]  Although at 8% the overall figure for All’s Well is not high, there is striking variation within the play’s individual scenes, with (at the top end) 30.55%, 21.5%, 16.66% and 16% (of the play’s total of verse lines) for, respectively, 4.4, 1.3, 2.4 and 1.1 and (at the low end), 2%, 3.5% and 4.9% for 3.4, 2.3 and 3.5.

Middleton tends to favour ‘All’as a speech prefix (A Mad World My Masters Q 1608, E4;  Michaelmas Term Q 1607, A3) whereas Shakespeare’s preference is for ‘Omnes’ (used as a speech prefix 6 times in F Antony and Cleopatra, once in Coriolanus). Although ‘All’ occasionally appears as a speech prefix in Shakespeare quartos, it occurs only twice in the Folio, both times in All’s Well: TLN 960 (designating the bachelor Lords at the French Court) and TLN 1970 (grouping together the soldiers who ambush Paroles and pretend to be the enemy). Further, Jonathan Hope’s quantitative analysis of auxiliary ‘do’ forms in Shakespeare and other dramatists points out that Shakespeare’s use of the unregulated form (‘I did go home’, as opposed to the regulated ‘I went home’) is distinctive, never exceeding 84%, whereas no other author shows a regulation rate below 85%. Middleton’s average is 90%, a strong preference for the regulated form. The unregulated figure for All’s Well is at the top of Shakespeare’s range, 84%, although a scene-by-scene breakdown shows a range of values from 54% (4.1) to 100% (3.1 and 3.3). According to this analysis, 13 of the play’s 22 scenes fall outside Shakespeare’s normal range. Interestingly, Hope’s analysis of auxiliary ‘do’ forms (in The Authorship of Shakespeare’s Plays) leads him to be sceptical about claims that Middleton was a co-author of Timon of Athens, since, at 84%, that play’s regulation rate is ‘at the upper limit of, but not outside, the Shakespeare comparison sample’: it is only the same high value, 84%, for All’s Well that brings Timon within the ‘Shakespearean’ limit.[16]

Several items of vocabulary make their first (and sometimes only) appearance in the OED in All’s Well. Some of them have an interesting pre- or after-life in Middleton’s plays.  When Paroles writes to Diana, warning her against Bertram, he accuses Bertram of being ‘ruttish’ (TLN 2318). The adjective is clearly a coinage from animal terminology; in Pericles a visitor to the brothel in Mytilene is converted by Marina and exits with the declaration ‘I am out of the road of rutting for ever’ (4.4). But the only other occurrence of ‘ruttish’ as an adjective in Elizabethan / Jacobean drama comes in  Middleton’s The Phoenix (Q 1607) when the captain describes his oversexed father. A related form occurs in Middleton’s The Witch, wherea character lodges at ‘Rutney’s’ (3.2.38), an apparently fictitious establishment whose name advertises  the activities carried on under its roof.

The King’s fistula has raised eyebrows, starting with the onstage reaction of Lafew to Bertram’s introduction of the topic (‘I would it were not notorious’ sounds like a rebuke to the young count for his indelicacy; TLN 38); but literary critics and medical historians have also expressed surprise at this ailment.

In  the source, William Painter’s Palace of Pleasure (a shared property of Shakespeare, who used it for Romeo and Juliet, and Middleton, who used it for The Revenger’s Tragedy, and for the writing duo together, for Timon of Athens) the King suffers from ‘a swelling upon his breast’.  Fistulas may occur in sundry body parts but are usually anal (hence the invention of the fistula chair -- a chair with a hole in its seat to provide sedentary comfort for the sufferer, illustrated in early modern surgical manuals). The probable change to an anal fistula connects All’s Well with two Middleton plays The Widow (1615-16) and A Game at Chess (1624). (The only other occurrence in Elizabethan and Jacobean drama is in Marston’s The Fawn 1606).

Once the possibility of a Middleton collaboration is mooted, other textual, stylistic and thematic peculiarities of All’s Well make more sense. Cornets were already a part of the boys’ company soundscape from the beginning of the century (in, for example, Marston’s Antonio plays and Middleton’s Your Five Gallants) and thus might register a Middleton contribution, a simpler hypothesis than the complicated revival scenario, particularly since there are no records of performance. Editors have long noted the typographical oddity of the long dash in the Countess’ speech in 1.3.153:

(Gods mercie maiden) dos it curd thy blood
To say I am thy mother? what’s the matter
That this distempered messenger of wet?
The manie colour’d Iris rounds thine eye?
------------------------ Why, that you are my daughter? (TLN 473-77)

Although Folio dashes usually indicate an interruption, Susan Snyder notes that here the dash ‘seems to indicate a pause while the Countess waits for an answer’.[17] There is another example, in Paroles’ speech at TLN 1145. Long dashes to indicate pauses for stage business are not unknown in printed playtexts (as in George Wilkins’s The Miseries of Enforced Marriage, a prodigal-son play in the King’s Men’s repertoire at the same time as All’s Well and published in 1607).  But the dramatist in whom they are unignorably frequent is Middleton. Middleton’s manuscripts and quartos abound with long dashes mid-line, indicating pauses, prompts or changes of address: The Phoenix (66 examples),The Revenger’s Tragedy (60 examples),and The Lady’s Tragedy (81 examples).

The scene in which this speech of the Countess occurs is one in which the variant speech-prefixes prompted Bowers to conclude that the scene was revised, with the dialogue between Helen and the Countess about Helen’s love for Bertram being a later insertion.  He concluded this because the Countess begins the scene with the consistent speech-prefix Coun.  When Helen enters, the Countess’s  speech-prefixes change (also consistently) to Ol. Cou. (TLN 450-513).  At TLN 514 her speech prefixes revert  (consistently) to Cou. We suggest that Shakespeare wrote the opening of scene 1.3, the Countess’ dialogue with the clown and the steward, in which the steward reveals that he overheard Helen declare her love for Bertram. The dialogue concludes with the Countess telling the steward to keep this secret (TLN 447): ‘stall this in your bosome’. (‘Although ‘stall’ is frequent in Elizabethan/Jacobean texts as a noun, only Shakespeare uses it as a verb.) We suggest that Middleton enters with Helen into this scene, when the Countess’s speech prefixes change. (Rhyming couplets and feminine endings also proliferate.) Although the speech prefixes subsequently revert to the pre-Middleton form, the dialogue retains some Middleton preferences – ‘in’t’ (twice), ‘I’de’, and a number of feminine endings.

Middleton’s favoured contractions, colloquialisms, exclamations, collocations and orthographical preferences have been well studied by Lake, Holdsworth, Jackson, Taylor, and Jowett.  Of their lists of Middleton’s preferences, All’s Well features ‘t after prepositions (as in for’t, on’t) 29 times; the unusual do’s for does ten times; and ha’s for has ten times. His preference for the modern forms does (over doth) and has (over hath) occurs ten and sixteen times respectively. His liking for contractions with I (such as I’d, I’m, I’ve) is noticeable in six appearances of I’d. The numbers may not seem especially high. But what is notable is the way in which these Middleton preferences cluster in certain scenes.  Act 1, scene 1 has for’t, in’t (twice), by’t, with’t (twice). Act 4, scene 3, the gulling of Paroles, has ha’s (six times), in’t (twice), on’t (twice), for’t (twice), do’s (twice), does (three times) and I’de. This scene also contains a distinctive piece of Middleton vocabulary (‘ruttish’; distinctive, as explained above, because the adjective is used by Middleton but occurs nowhere else in early modern drama).  The novelistically explanatory stage direction Enter Paroles with his interpreter - which reminds us of the plot rather than simply cueing an entry – is also recognizable from Middleton’s practice in, for example, The Revenger’s Tragedy: Enter Lussurioso, with Hippolito, Vindicies brother (D2); Enter the discontented Lord Antonio, whose wife the Duchesses yongest Sonne ravisht (Cv). So too is the feature seen in 4.3 of introducing a character with a generic or status marker in the stage direction and giving them a specific name in speech-prefixes. All’s Well has Enter Count Rossillion followed by speech-prefixes Ber., akin to Middleton’s practice. In Michaelmas Term the stage direction provides an entrance for Quomodoes wife (C3v) but her subsequent speech-prefix is Toma. (for her name, Thomasine). At the end of Your Five Gallants the heiress, Katherine, is instructed to enter as follows:  Enter the Virgin betweene two antient gentlemen. (I2v); her speech-prefix is Kathe.   Middleton is also fond of appositive explanations in stage directions: Enter Lussurioso, and Infesto two Lords (Phoenix I2r); Enter Vindici and Hippolito, Vindici in disguise to attend L. Lussurioso the Dukes sonne (Revenger’s Tragedy B2v); Enter Misters [sic] Katherine with Fitzgraue a Gentleman (Your Five Gallants B2r). It seems likely that the unusual stage direction Enter a gentle Astringer. at TLN 2601 is a misreading of a gentleman, a stranger as the subsequent speech-prefixes are Gent.; this prefix is more likely to be an abbreviation of a noun (Gentleman) than of an adjective (gentle). If so, the original stage direction, with its noun in apposition, is typical of Middleton. Collectively these features invite attention.


Act 4, scenes 1 and 3, in which the blindfolded Paroles  fears he shall lose his life for ‘want of language’ (4.1, TLN 1984), are comic tours de force in which Paroles’ comrades pretend to be the enemy and speak a nonsense-language to deceive him. In 4.1:

All.      Cargo, cargo, cargo, villianda par corbo, cargo. (TLN 1979)

In 4.3:

Int.      Bosko Chimurcho.  [TLN 223]

Middleton is a playwright who is drawn to mock-languages. The Phoenix gets standard comic mileage out of Tangle’s love of legal Latin phrases but then takes the linguistic comedy to a new level when Tangle combines legal and fencing terms to indicate the thrusts, parries and counter-thrusts of the law. In The Puritan Widow the comedy is fuelled by the mock conjuring language of Captain Idle. When the Captain seeks ‘boisterous words and horrible names’, Pieboard tells him:  ‘any fustian invocation, captain, will serve as well as the best, so you rant them out well. Or you may go to a ’pothecary’s shop and take all the words from the boxes’ (3.5.98-102).  This could be the advice followed by the Dumaine brothers in All’s Well.

All’s Well is the only Shakespeare play to begin with a female character (the Countess) speaking. Shakespeare’s preference elsewhere in the canon is to open a play with a male (Macbeth is an ambiguous counter-example). Nor is he alone in this preference: as Leslie Thomson points out, ‘none of the plays by Greene, Marlowe, Marston, Jonson, Massinger, Webster, or Ford begin with a woman’.[18] Three Middleton plays do:The Puritan Widow, A Chaste Maid in Cheapside and Women Beware Women.

That the Countess might have a Middletonian edge enables us to see past the play’s recent sentimentalized stage history, which has conditioned us to see her as a fairy godmother figure. One thinks of Peggy Ashcroft (RSC, directed by Trevor Nunn in 1981), Judi Dench (RSC, directed by Greg Doran in 2003) and most recently (and obviously) Clare Higgins in Marianne Elliott’s modern fairy tale at the National Theatre in 2009. In Elliott’s production, the ending of the play offered an exquisitely tender relationship, a couple reunited after a marathon journey – the couple in question being not the husband and wife but the daughter-in-law and mother-in-law. Such attractive productions conceal the essentially bawd-like activities of both the Countess and, later, of the Widow, mother to Diana (who are at times reminiscent more of the pragmatic mothers of Castiza in The Revenger’s Tragedy or Frank Gullman in A Mad World My Masters than of Cinderella’s fairy godmother).

This more Middletonian Countess fits the play’s city comedy realism which several critics have noted with degrees of surprise or discomfort. The French lords go to war not for ideological reasons but for cardiovascular exercise (‘It well may serve / A nursserie to our Gentrie, who are sicke / For breathing, and exploit’; TLN 258-60). Paroles’ betrayal of his comrades’ military secrets includes the prosaic calumny that one of the Dumaine brothers wets the bed (‘he will be swine-drunke, and in his sleepe he does little harme, save to his bed-cloathes about him: but they know his conditions, and lay him in straw’; TLN 2358-60). Helen is not (just) a magician or a healer but a shrewd mercantile negotiator who bargains with the King (‘But if I helpe, what do you promise me’; TLN 801) and buys up the services of the Florentine women in Act 4:

Take this purse of Gold,
And let me buy your friendly helpe thus farre,
Which I will over-pay, and pay againe
When I have found it. (TLN 1872-75)

Critics are agreed that this is not what we think of as Shakespeare’s kind of play.

The heroine and the plot she engineers bore the brunt of nineteenth-and twentieth-century prudery (disguised as critical distaste). Helen’s frank discussion with Paroles about how women may lose their virginity to their own liking, her business-like attitude in capturing her man, and her perpetration  of the bedtrick led Milton Shulman in the Evening Standard to call Helen a ‘heroine with the brash moral standards of an ambitious strumpet’.[19]But Bertram also attracts condemnation. He cowardly compels Helen to tell the King that her hasty post-wedding departure is her own initiative (‘make this hast as your owne good proceeding, / Strengthned with what Apologie you thinke / May make it probable neede’; TLN 1260-62). He is underage, shallow and immature. His immaturity is stressed in ways that suit the genre of prodigal-son comedy, to which this play belongs, but, unusually among both prodigal and Shakespeare heroes, he commits further crimes in the final scene, the very scene that is designed to acquit him of his prior transgressions.  When the King recognizes Helen’s ring (a gift from the King) on Bertram’s finger (where it was placed during the bedtrick), Bertram quickly improvises a lie: ‘In Florence was it from a casement throwne mee, / Wrap’d in a paper’; TLN 2804-05).

The precarious happy ending occurs not because Bertram is converted but because he is cornered. The couplet in which he accepts Helen (addressed, we note, not to his wife but to the king) simply repeats his reaction to marriage in Act 2 by again imposing a condition:  

If she my Liege can make me know this clearly,
Ile love her dearely, ever, ever dearly. (TLN 3053-4)

Theatrically the couplet is, as J. L. Styan notes,  ‘too short to show a psychological change of heart, and too long to suggest a miraculous bolt from heaven’.[20] Shakespeare could easily have written a speech in which Bertram explained that he saw the error of his ways and now loved Helen; clearly Shakespeare did not wish to write such a speech. This may be in part because endings which overlook, rather than reform, youthful failings are typical of Middletonian comedy (see for example, A Mad World My Masters). 

Some of the play’s grittiness of environment and character has been explained by its being a ‘problem’ play, one of that group of late Elizabethan/early Jacobean plays that F. S. Boas first identified in 1896 in Shakespeare and his Predecessors. Boas’s group of four plays – Hamlet, Troilus and Cressida, All’s WellMeasure for Measure (plays he grouped together as being on the cusp of comedy and tragedy, plays that address a social problem) – has shrunk to three: All’s Well, Troilus and Measure. However, even among this group of three,All’s Well is anomalous. Troilus and Measure satirise politics and ideals of honour;All’s Well satirises comic convention. Bertram is trapped not by Helen but, as A. D. Nuttall writes,  ‘by comedy plotting’; and Shakespeare’s finale in which Bertram ‘must be reclaimed for the happy ending’ explores and explodes a genre that is traditionally defined by its ending.[21] But including Middleton in the mix allows for different comic possibilities. We know that Shakespeare and Middleton were collaborating in 1606-07. Middleton had just finished his comedy A Mad World My Masters (not quite a city comedy: it has city comedy’s values but not its location). A Mad World is a comedy in which, as George E. Rowe first argued in Thomas Middleton and the New Comedy Tradition, Middleton interrogates the genre of New Comedy: it is a comedy about comic genre. All’s Well takes this one stage further.[22] David Scott Kastan analyses the ways in which the play tries so (too?) hard to fit into different comic models: Helen plots a New Comedy; the Countess attempts a Christian prodigal-son comedy, urging the King to forgive her son’s transgressions (‘make it /Naturall rebellion, done i’th blade of youth’; TLN 2702-03, our emphasis); with the ring and the letter it becomes a Terentian comedy of intrigue.[23]


Recent scholarship has done much to re-establish collaboration as part of Shakespeare’s professional practice at different stages of his career. It is now broadly acknowledged that he collaborated with Peele on Titus Andronicus and with Nashe on 1 Henry VI in the early 1590s, with someone as yet unidentified on Edward III, and that his work with Fletcher at the end of his career takes in Two Noble Kinsmen, All is True and, apparently, the lost Cardenio. Add to this the collaboration with Middleton on Timon and with Wilkins on Pericles, the suggestion that Shakespeare may have written the additions to The Spanish Tragedy, ongoing investigations into his part, if any, in Arden of Faversham, and Jowett’s recent allocation of his play-patching on the problematic Thomas More manuscript to around 1604, and there emerges a more complicated picture of collaborative practice throughout his career into which our suggestion about All’s Well can be factored.[24]

There are methodological implications too. Much of the scholarship on attribution studies has been shaped by sophisticated statistical analyses of linguistic and prosodic style, giving its conclusions the veneer of laboratory verifiability but disenfranchising most of the literary community from understanding, let alone challenging, its method. Our analysis has proceeded differently, by eye rather than by computer, in reading not data-crunching, and via literary, qualitative judgements rather than mathematical quantitative ones. We freely acknowledge that further work on ‘Middleton markers’ in All’s Well would benefit from additional analytical and stylometric tools, but in suggesting that the play might be collaborative we have not undertaken such work. Rather, where we have noted something unusual or revisited long-standing editorial questions about the play, we have made use of LION and other online resources to follow it up.  We have made use of the numerical and statistical tables supplied by Wells and Taylor, Jackson, Hope, Lake, Vickers and Barker but we have not sought to repeat or mimic the large-scale data sweeps undertaken by them and other colleagues in this field. Our more narrative approach here will doubtless draw opprobrium from adherents of purely quantitative method, but we strongly believe that literary critics should not leave these issues entirely to statisticians and statistical approaches. The discussion of attribution is not separate from, nor prior to, the most basic procedure of literary criticism: that of interpretation.

So far as we are aware, in suggesting that the play is collaborative we have been anticipated by only one previous critic, John Dover Wilson. Identifying many of the textual problems discussed here, and taking a rather harsher view of the play’s tone, Dover Wilson identified another hand in All’s Well. Writing in 1929, he posited ‘a process of expansion by some inferior dramatist’ who ‘had a passion for sententious couplets and a mind running on sexual disease’.[25] He could not name the playwright, but suggested the same collaborator might have had something to do with the seamier aspects of Measure for Measure. Thanks to the Oxford Middleton it is clearthe Folio text of that play represents a Middleton adaptation of Shakespeare, dating from the early 1620s (as the editors argue in Thomas Middleton and Early Modern Textual Culture)[26] – and the editorial work on Middleton has also given us a much richer and more appreciative view of his writing than Dover Wilson’s thumbnail sketch.


Our preliminary survey of Middleton markers in All’s Well suggests that Middleton’s putative presence is most prominent in the Paroles scenes and gives responsibility for the bedtrick to Shakespeare. We find particular concentrations of Middleton forms in 1.1, 1.3, 2.1, 2.3, 4.1, 4.3 and 5.3. Further work is needed to corroborate Middleton’s presence in All’s Well, and to investigate whether this might be collaboration, like Timon, or a revision, like Measure, or a different form of collaboration altogether such as one author functioning as plotter. Contextually, All’s Well needs to be reinvestigated as one of a clutch of prodigal-son plays in the King’s Men’s repertoire c.1607: A Yorkshire Tragedy, The London Prodigal, The Miseries of Enforced Marriage. (The first two of these plays are associated with both Middleton and Shakespeare.) But stylistically it is striking how many of the widely acknowledged textual and tonal problems of All’s Well can be understood differently when we postulate dual authorship. A broad-brush summary might look like this. One author knew that the two French lords had names, the other did not. One preferred personal names over types, and drew for many of them on his earlier plays. One used different speech prefixes from his collaborator. One wrote narrative stage directions as explanation to his partner at point of handover. One was more inclined to rhyming couplets and to hypermetric verse. One wrote like William Shakespeare and one wrote like Thomas Middleton.


[1]Stanley Wells and Gary Taylor, with John Jowett and William Montgomery, William Shakespeare: A Textual Companion (Oxford: Clarendon Press, 1987), p.492.

[2]Fredson Bowers, ‘Foul Papers, Compositor B, and the Speech-Prefixes of All’s Well that Ends Well’, Studies in Bibliography 32 (1979): 60-81, p.76.

[3]Fredson Bowers, ‘Shakespeare At Work: The Foul Papers of All’s Well That Ends Well in John Carey (ed.), English Renaissance Studies Presented to Dame Helen Gardner in Honour of Her Seventieth Birthday (Oxford: Clarendon Press, 1980): 56-73, p.61.

[4]William Shakespeare:A Textual Companion, p. 492, 493.

[5]William Shakespeare: A Textual Companion, p. 96.

[6]MacDonald P. Jackson, ‘Spurio and the Date of All’s Well that Ends Well’, Notes & Queries 48 (2001): 298-9.

[7]Gary Taylor, “Divine [ ]sences," Shakespeare Survey 54 (2001): 13-30, p.24n.

[8]Lois Potter, The Life of William Shakespeare: A Critical Biography (Oxford: Wiley-Blackwell, 2012).

[9]Gordon McMullan, ‘What is a late play?’ in Catherine Alexander (ed.), The Cambridge Companion to Shakespeare’s Last Plays (Cambridge: Cambridge University Press, 2009): 5-27.

[10]Bowers, ‘Shakespeare at Work’ and ‘Foul Papers’.

[11]M.C. Bradbrook, ‘Virtue Is the True Nobility: A Study of the Structure of All’s Well that Ends Well’, Review of English Studies 1:4 (1950), 289-301, p.289

[12]John Jowett (ed.), Timon of Athens (Oxford: Oxford University Press, 2004), p.145.

[13]MacDonald P. Jackson, Studies in Attribution: Middleton and Shakespeare (Salzburg: University of Salzburg, 1979); David J. Lake, The Canon of Thomas Middleton’s Plays: Internal Evidence for the Major Problems of Authorship (London: Cambridge University Press, 1975); Gary Taylor and John Lavagnino (eds), Thomas Middleton: The Collected Works and Thomas Middleton and Early Modern Textual Culture (Oxford: Oxford University Press, 2007).

[14]R. H. Barker, ‘The Authorship of The Second Maiden's Tragedy and The Revenger's Tragedy’ Shakespeare Association Bulletin, 20 (1945): 51–62, 121–33

[15]Jowett, p.136.

[16]Jonathan Hope, The Authorship of Shakespeare’s Plays: A Socio-Linguistic Study (Cambridge: Cambridge University Press, 1994), p.35, p.100. 

[17]Susan Snyder (ed.), All’s Well that Ends Well (Oxford: Oxford University Press, 1993), 1.3.153n.

[18]Leslie Thomson, ‘Making a Woman of the Boy’ in A.L. Magnusson and C.E. McGee (eds), The Elizabethan Theatre XIV: Papers Given at the International Conference on Elizabethan Theatre held at the University of Ontario in July 1989 (Toronto: P.D. Meany, 1996), p. 158

[19]Milton Shulman in The Evening Standard, quoted by J.L. Styan, All’s Well that Ends Well (Manchester: Manchester University Press, 1984), p. 14.

[20]Styan, p.115.

[21]A.D. Nuttall, Shakespeare the Thinker (New Haven and London: Yale University Press, 2007), p.247.

[22]George E. Rowe, Jr, Thomas Middleton and the New Comedy Tradition (Lincoln: University of Nebraska Press, 1979.

[23]David Scott Kastan, ‘All’s Well that Ends Well and the Limits of Comedy’, English Literary History 52 (1985): 575-589.

[24]Brian Vickers, Shakespeare, Co-Author: A Historical Study of Five Collaborative Plays (Oxford: Oxford University Press, 2002), and ‘Shakespeare and Authorship Studies in the Twenty-First Century’, Shakespeare Quarterly 62 (2011): 106-42; Richard Proudfoot, ‘The Reign of King Edward the Third (1596) and Shakespeare’, Proceedings of the British Academy, 71 (1985): 169–185; MacDonald P. Jackson,  Defining Shakespeare (Oxford: Oxford University Press, 2003), and ‘Shakespeare and the Quarrel Scene in Arden of Faversham’, Shakespeare Quarterly 57 (2006): 249-93; John Jowett (ed.),  Sir Thomas More (The Arden Shakespeare, London: Methuen, 2011).

[25]John Dover Wilson (ed.), All’s Well that Ends Well (Cambridge: Cambridge University Press, 1929), pp.100-11.

[26]Thomas Middleton and Early Modern Textual Culturep.681. 

Appendix: Tabulation of forms favoured by Middleton

As John Jowett explains in Appendix C to his Oxford World’s Classics edition of Timon of Athens (2004), ‘in the works of Shakespeare and Middleton … the following forms are markedly favoured by Middleton, though few are exclusive to him’. Here we give TLN references from the Norton facsimile of the Folio, and give Folio spellings, although we have divided the list into acts and scenes according to modern editions.


in’t       87

for’t     138

in’t       140

by’t      151

with’t  151

ne’re   157

with’t  159



[   ]



doe’s   336

on’t     459

in’t       518

in’t       577

I’de      582



too’t    628

for’t     668

do’s     706

does    746



[   ]



I’de      952

All.       960

ha’s     1008

ha’s     1012

I’me     1050

do’s     1088

I’de      1159

I’de      1169

in’t       1192

ha’s     1208




do’s     1215

do’s     1227

do’s     1252



does    1332

does    1359



[   ]



[   ]



[   ]



does    1590



does    1690



for’t     1767

on’t     1778

too’t    1796



on’t     1903



All.       1979

Oh       1991

Oh       1992



ha’s     2051



in’t       2210

ha’s     2207

ha’s     2209

does    2212

ha’s     2220

on’t     2244

for’t     2258

do’s     2312

in’t       2321

does    2234

does    2359

ha’s     2361

ha’s     2368

do’s     2384

ha’s     2390

I’de      2419

on’t     2422  



[   ]



do’s     2557



[   ]



ha’s     2661

does    2682



on’t     2735

on’t     2784

do’s     2907

do’s     2912

does    3028

Oh       3042

[Shakespeare’s preference for ‘O’ occurs at TLN 21 (Act 1.1); TLN 865, 868, 871, 873, 875, 882 (Act 2.2); TLN 1999 (Act 4.1); and TLN 3057 (Act 5.3)]