Arden of Faversham

The Shakespeare Conference: SHK 30.098  Friday, 8 March 2019


From:        Gabriel Egan <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date:         March 8, 2019 at 5:12:07 AM EST

Subject:    Arden of Faversham




In late 2018, American Notes and Queries published Pervez Rizvi’s essay “Small Samples and the Perils of Authorship Attribution for Acts and Scenes” (doi:10.1080/0895769X.2018.1537841). Rizvi’s opening sentence set out his aim: “This note will challenge one of the tests used by MacDonald P. Jackson to attribute scenes 4-9 of the anonymous play ‘Arden of Faversham’ to Shakespeare”. Jackson’s test had been the subject of his chapter “A Supplementary Lexical Test for ‘Arden of Faversham’” in the Authorship Companion to the New Oxford Shakespeare.


Jackson has written a response to Rizvi’s critique, calling it “Reconsidering a Lexical Test of ‘Arden of Faversham’: A Response to a Critique” and it has been published in the journal ‘Shakespeare’ (doi:10.1080/17450918.2019.1573847). The abstract for Jackson’s response is:


This article is a response to a critique by Pervez Rizvi of an essay of mine entitled “A Supplementary Lexical Test for ‘Arden of Faversham’”, which employed a few high-quality lexical markers that are used much more often, or much less often, by Shakespeare than by a large control group of early modern dramatists, to support the view that Shakespeare was largely responsible for the central act of ‘Arden’. Rizvi drew on my essay to make certain sound points, notably that results for mere acts or groups of scenes cannot validly be mixed with results for whole plays, and that when one relies on only a small number of marker words the outcome can turn on the presence or absence of only one or two instances. Rizvi divided plays into 4657-word segments—the same length as ‘Arden of Faversham’, 4-9--and noted that, when my method was applied to them, some Shakespeare segments yielded very low scores and some non-Shakespeare segments yielded very high scores. My response aims to counter some of Rizvi’s strictures, clarify my assumptions and procedures, and show that the new data that he has generously supplied strengthen, rather than weaken, my original argument. 


As an editor of the journal ‘Shakespeare’ I’m glad to be able to offer a copy of Jackson’s new article to anyone who does not have access to that journal. Just ask and I’ll email you a PDF.


Being an editor of the journal that is publishing Jackson’s essay while being also an editor of the book (the Authorship Companion to the New Oxford Shakespeare) that published Jackson’s original essay obliges me to declare that interest. Naturally, I recused myself from making any judgements on Jackson’s new essay when it was submitted to the journal ‘Shakespeare’: my fellow co-editors handled the entire process. Of course, it may be suspected that my co-editors are naturally disposed to be favourable towards me and hence favourable towards the New Oxford Shakespeare that I’m a General Editor of. I can’t prove that they are not, but I can point to the journal ‘Shakespeare’ being even-handed in such matters.


In 2012 the journal ‘Shakespeare’ published Brian Vickers’s essay “Identifying Shakespeare’s Additions to ‘The Spanish Tragedy’ (1602): A New(er) Approach”.  In another publication, the Year’s Work in English Studies, I reviewed Vickers’s essay and pointed out what I think are its weaknesses. Vickers wrote a response to my review called “Shakespeare and the 1602 Additions to ‘The Spanish Tragedy’: A Method Vindicated” and sent it to the journal ‘Shakespeare’, which duly published it.


Amongst the phrases in which Vickers responded to my review are “Egan seems not to have realised this”, “Egan’s ignoring of the considerations of”, “Egan conveniently ignores”, “it is somewhat disingenuous of Egan to claim”, “Egan’s unjustified dismissal of”, and “Egan and Hoover fail to realise”. As with Jackson’s new essay, I recused myself from making any judgements on Vickers’s response when it was submitted to the journal ‘Shakespeare’: my co-editors handled the entire process. I hope that this stands as evidence that the journal ‘Shakespeare’ does not take sides in these debates. I’m happy to also offer anybody who wants it a copy of Vickers’s response to my review.



Gabriel Egan




The Completion of Arden 3

The Shakespeare Conference: SHK 30.097  Friday, 8 March 2019


From:        Charles Weinstein <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date:         March 7, 2019 at 3:13:10 PM EST

Subject:    The Completion of Arden 3


According to, Measure for Measure, the final volume in the Arden 3 series, will be published in the US on November 14th. Although A.R. Braunmuller was originally the sole editor, the cover of the impending volume lists Braunmuller and Robert N. Watson as co-editors.


— Charles Weinstein 




The Shakespeare Canon and NOS

The Shakespeare Conference: SHK 30.096  Thursday, 7 March 2019


[1] From:        Gabriel Egan <This email address is being protected from spambots. You need JavaScript enabled to view it.>

     Date:         March 6, 2019 at 4:34:58 PM EST

     Subj:         Re: SHAKSPER: NOS


[2] From:        Gerald E. Downs <This email address is being protected from spambots. You need JavaScript enabled to view it.>

     Date:         March 7, 2019 at 2:35:32 AM EST

     Subj:         Re: NOS




From:        Gabriel Egan <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date:         March 6, 2019 at 4:34:58 PM EST

Subject:    Re: SHAKSPER: NOS




I’m grateful to Pervez Rizvi for explaining the nature of Gerald Downs’s query. Rizvi asks us to “imagine a text that uses just three function words: ‘and’, ‘it’, ‘that’” in which the only sequences of these words are:










In Rizvi’s hypothetical, the text uses these phrases an equal number of times but he doesn’t say how often.  Let’s assume two times; it shouldn’t make any difference.  Because by the rules of Rizvi’s hypothetical only the above four sequences are the sequences in which these three words occur, each occurrence of one of these pairs must be separated from the next by a run of words that are not IT, AND, or THAT. I’ll represent these intervening words with “XXX” to produce this possible source text that meets Rizvi’s criteria:






Rizvi doesn’t state how big is the window in which words have to fall be the considered adjacent (in the WAN method it’s five words) so I’ll assume that it’s just two words since he has asked us to consider word pairs.


If we construct the raw Markov chain for the adjacencies of IT, AND, and THAT in the above text we get:



 IT  0   2    0

AND  2   0    2

THAT  0   2    0


In this matrix, the column “IT, AND, THAT” represents the nodes (words) you start on and the row “IT, AND, THAT” represents the nodes (words) you get to. Observe three things:


1) If you start at IT there's only one place to go: AND.


2) If you start at AND there are two places to go: IT and THAT.


3) If you start at THAT there's only one place to go: AND


The ‘limit probabilities’ calculation does indeed give AND a heavier weighing that IT or THAT.


Rizvi wrongly summarizes this as “the phrases ‘and it’ and ‘and that’ are given twice the weight of the phrases ‘it and’ and ‘that and’”. That’s not true: the method isn’t at this point giving weight to phrases (which are represented by the edges between nodes). Rather it is giving weight to words (which are represented by nodes).


Rizvi asks a rhetorical question: “Is it obvious to you why this [weighting one phrase over another] is the right thing to do for authorship attribution purposes? Me neither.” Rizvi’s implied question, asking why we do this, is unanswerable since our method doesn’t do what he thinks it does.


Rizvi could have asked why we should weight the word AND more heavily than the words IT or THAT? One answer would be that we do this because in the text AND can be followed by IT or THAT (2 choices) whereas IT can be followed only by AND (1 choice) and THAT can be followed only by AND (1 choice). That is, the weighting reflects an authorial preference about the word AND that is demonstrated in the sample text.


In fact, the calculation of limit probability is subtler than that, because it apportions weight according to the inbound as well as the outbound edges of each node, meaning that it reflects the authorial willingness to use AND before or after another word. The limit probability is a way of expressing the author’s preference for AND in combination with other function words without simply reducing this preference to the frequency of use of AND.



Gabriel Egan





I’ll try to condense both Gabriel Egan’s replies to my questions and my further responses.


I thought I’d at least laid the PageRank analogy to rest . . . with my explanation that the ‘limit probability’ calculation of node-weight in the Markov chain that is PageRank’s model of a series of webpages is the same . . . calculation . . . in the Markov chain that is a [WAN]. But that’s all I really can say. The analogy . . . is that the same mathematical notion . . . solves the . . . problem of apparent infinite regression . . . in both cases.


This analogy reminds me of one reported by the justly eminent American historian Richard Armour: Columbus convinced Queen Isabella that the world was round by showing her an egg. It’s not enough to claim the “same mathematical notion” applies to the earth as to an egg. We must presume that Isabella learned something more of the analogy than that. Pervez Rizvi observes that Gabriel Egan is obligated to say why a WAN is a Markov chain. Otherwise, a “calculation” is not an analogy. I accept that Gabriel can say no more about that.


Take his sentence “He [Egan] may have been training his co-authors and readers to accept the Shakespeare Folio as good data for WANs”. Is that sarcasm? Is Downs saying that Shakespeare’s Folio texts are for some reason unsuitable for computational analysis of his style?


No sarcasm. I’m of the opinion that Gabriel and the NOS promote a view of Shakespeare’s texts more agreeable with their past ‘theories’ than with the truth—and that they must know better. F ought to be considered generally unfit for ‘too too’ stylometry until textual questions are answered. I assume co-authors from other disciplines are not knowledgeable enough to judge as they might otherwise be; most Shakespearians aren’t concerned with textual problems. That’s no knock on those interested in different aspects of a wide subject, though many should care more than they do when text matters to their work.


I once expected rejection of a Lear submission (as puzzle-pieces were missing). I’d wanted to write after Knowles’s Variorum (yet to appear) and tried unsuccessfully to learn what Blayney had withheld in 1980 (and finally revealed in 2017). What surprised me were statements from the editor and a reader: 


“It has been extremely difficult to find qualified readers to review the essay.” True, though “willing, qualified readers” may have been meant. One of those found made a point I hadn’t fully considered: 


There is doubtless a handful of scholars who would read this essay with some interest, but it's a tiny handful. And the struggle through the essay has a small payoff. It raises good questions (which is why I say it probably deserves publication elsewhere), but only in the mind of a reader prepared to understand and follow it.


So much for good questions. But isn’t this a little off-putting for a major journal—one that now believes its readership is up to WANs, Markov, entropy, and averaged playwrights?


I also struggle with Downs’s phrasing. He writes that “In practice, Gabriel describes the exclusion of evidence negative to a professed belief” and my guess is that Downs here uses “negative to” to mean “that is inconsistent with”. I’ve never heard anybody use “negative to” that way . . .


Gabriel’s guess was right. Noticing the odd usage myself, and pondering “a first time for everything,” I opted to test a question I’ve never heard asked: “Is there a second time for everything?”


Actually, I’d altered the sentence to justify the margin manually—as is my habit—when an unconsciously placed function word was suddenly overweight. But to did fit; knowing it ‘had little or no meaning,’ I let it pass until the next margin reset, which I forgot. I do believe it’s unfortunate that Shakespeareans have left so many ‘critical’ questions unanswered—or unaskt—for me to raise. I wonder how interesting the canon would be if it were not contaminated. But interestingly corrupt it is.


Gerald E. Downs





Q1 Hamlet

The Shakespeare Conference: SHK 30.095  Thursday, 7 March 2019


From:        Brian Vickers <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date:         March 7, 2019 at 11:33:06 AM EST

Subject:    Q1 Hamlet


Is Q1 Hamlet Reported? 


The aim of this research project has been to test the hypothesis that Q1 Hamlet (1603) was put together by actors and reporters and represents an unauthorised text. This hypothesis was supported by two major scholarly studies published during the Second World War: G. I. Duthie, The ‘BAD’ QUARTO of HAMLET (Cambridge, 1941), and Alfred Hart, Stolne and Surreptitious Copies. A Comparative Study of Shakespeare’s Bad Quartos (Melbourne, 1942; reprinted 2014). This explanation was endorsed by a generation of Hamlet editors, including Harold Jenkins (1982), Philip Edwards (1985), and George Hibbard (1987), and those who edited the complete works between 1974 and 1986. It remained the consensus scholarly judgment until a backlash in the late 1980s against the New Bibliography, and especially the work of W. W. Greg, succeeded in discrediting the theory of memorial reconstruction and the whole concept of “Bad” Quartos. Setting aside Pollard’s unfortunate name for these texts, and the outdated notion of “pirates” violating the rights of stationers, the theory of memorial reconstruction was validated by Kathleen Irace in her book Reforming the “Bad” Quartos (1994) and in her edition, The  First  Quarto of Hamlet (1998). 


The starting point for all these scholars was the fact that Q1, allegedly “acted by his Highnesse seruants in the Cittie of London”, was soon replaced by Q2 (1604-5), firmly stated to be “By William Shakespeare. Newly imprinted and enlarged to almost as much againe as it was, according to the true and perfect Coppie”. It seemed obvious to the scholarly tradition to start by comparing the two editions, to see what it was about Q1 that caused Shakespeare’s company to replace it with the authorised text. Yet Laurie Maguire, in her schematic overview, Shakespearean suspect texts (1996), although acknowledging that Hamlet Q1 and Q2 are “parallel texts” (342 n. 38), announced (155) that 


I approach the Shakespearean suspect texts as if no parallel text existed i.e. I ignore any help or bias offered by a Q2 or F version. This puts the Shakespearean suspect texts on an equal footing with the non-Shakespearean, and facilitates contextual understanding. 


That remains a puzzling decision. Where does the implied “bias” reside? What is “contextual understanding”? Most serious, how can we put Q1 Hamlet on “an equal footing” with Q2, of which it is an attempted clone? Whatever its merits as a shortened acting version, preserving about half the text, it mangles Shakespeare’s language to the degree that a reader can often only understand it by referring to the authorised edition. Maguire’s example encouraged Terri Bourus, in Young Shakespeare’s Young Hamlet (2014), which not only avoided comparing Q1 with Q2 but barely quoted from Q1. Bourus also avoided engaging with the scholarly tradition, briefly dismissing G. I. Duthie, and never mentioning Alfred Hart (Maguire at least mentioned Hart a few times, albeit never engaging with his case). Bourus revived the long-discredited thesis that Q1 represents Shakespeare’s first draft (c. 1589), a theory treated with great respect by Gary Taylor and Rory Loughnane in the New Oxford Shakespeare Authorship Companion (542-8).


It seems a matter of some urgency to re-assess the possibility that Q1 Hamlet is a reported text. Previous scholars who have endorsed this hypothesis have studied internal evidence, ranging from sequences of high accuracy (derived from one actor still having his original “part”) to corrupt and unintelligible passages and the frequent misplacing of the text by anticipation and repetition. As I reported here (SHK 30.024, 19 January 2019) MacDonald Jackson recently devised new tests of the play’s prosody and vocabulary that decisively date Q1 to 1602-3. I endorse his dating and agree with his assessment of the play’s un-Shakespearian language, its confusedly elliptical utterances, narrative incoherence, “pedestrian, clumsily repetitive, and sometimes barely meaningful verse”. My own approach derives from having noticed that Q1 contains many reminiscences of plays in the repertoires of the London theatres between 1587 and 1603.  I am not the first person to compile such a list: Alfred Hart did so, briefly listing ‘Inter-play Borrowings’ (Stolne and Surreptitious Copies, 391-402).  


Hart relied on his memory and wide reading, I have had the benefit of Pervez Rizvi’s database of 527 plays,  marked up to automatically identify every instance of verbal repetition, whether n-grams (contiguous sequences of two or more words) or collocations (dispersed sequences within a ten-word window, the approximate length of a line of verse). Rizvi’s database contains 13,364 n-grams shared by Q1 Hamlet and the 526 other plays performed between 1552 and 1657. I have concentrated on the first 5,000, and given priority to unique n-grams, verbal matches occurring only twice in the corpus – that is, in this text and only one other -- which are the most reliable form of evidence. Some of the matches occur 3 or 4 times elsewhere, but I have only included phrases within the relevant period (1587-1603) that are echoed for the first time in Q1. I have worked with a copy of that text in modern spelling, reproducing the lineation of the Arden3 edition by Ann Thompson and Neil Taylor (2006), which I have marked up to identify material not included in Q2. I have chosen the 1604-5 text as my reference point to bring out the many differences that Shakespeare and his contemporaries would have noticed had they compared it with the unauthorised edition. I am aware that some material from Q1 occurs in the Folio but not in Q2; however, that does not affect the present enquiry. The advantage of marking Q1-only material is that it enables us to see that most echoes from earlier plays occur in the reporters’ own text. Reminiscences of alien material are much rarer in the passages of authentic Shakespeare, suggesting that the reporters’ memories were more intensely exercised when they were forced to create their own material. In several instances we find echoes clustering together as they eked out their invention. The full list of matching phrases can be found on my website Hart identified about 50 borrowings, I have added over 100 more. But I have also omitted about 50 instances of a unique parallel – occurring only in Q1 and in one other play – expressed in everyday language that any character in such a situation might have been expected to use. These “ordinary” utterances are often dismissed by readers unfamiliar with attribution studies, where the significance of verbal matches does not reside in their rarity. I am not claiming that every match is significant: readers will vary in their acceptance of each. But the total evidence, I submit, allows us to affirm that Q1 Hamlet is indeed a reported text, put together in 1602.


A detailed discussion of this evidence is impossible here, but some general comments can be made on the repertoire echoed by the reporters. I emphasise the plural, since previous discussions have focussed too much on Greg’s theory of a single reporter, such as the Host in The Merry Wives of Windsor. My reading of the evidence suggests that all the actors who had played major roles – Hamlet, the King, Gertred, Corambis, Leartes, Ofelia, the Player and the Gravedigger – were involved in the recreation of the text, a task that would have been impossible for one person, let alone a shorthand taker. I accept the traditional theory, confirmed by Irace in her computer-assisted study, that the most accurate reporter (probably because he had access to his own “part”) was the actor who played Marcellus, doubling that role with   Voltemand and two personages in the Mousetrap, Lucianus and the Prologue.


As for the texts they echoed, these included 40 non-Shakespeare plays, amounting to 70 borrowings. The most often cited was The Spanish Tragedy, probably revived in 1599, which they recalled 14 times, referring less often to other plays attributed to Kyd (Soliman and Perseda, King Leir, and Fair Em). They drew on five plays by Marlowe (1 and 2 Tamburlaine, Dido, Faustus, and Edward II), two by Lyly (Campaspe, Endymion), two by Peele (Edward 1, The Troublesome Reign), two by Greene (Friar Bacon and Friar Bungay, Orlando Furioso), three by Heywood (1 and 2 Edward IV, The Four Prentices of London), four by Marston (Antonio and Mellida, Antonio’s Revenge, Histriomastix, Jack Drum’s Entertainment), two by Jonson (EMIH, Poetaster), and one each by Yarrington, Dekker, Chapman, and Chettle, together with three anonymous plays.


The reporters of Q1 Hamlet drew proportionally more often on Shakespeare, amassing 98 borrowings from 25 plays, confirming earlier scholars’ theories that they must have belonged to his company. Tabulating their echoes, I was surprised how many came from the 8 history plays, 39 in all: 8 from 2 Henry VI (counting 2 from The Contention), 8 from 3 Henry VI, 7 from Richard III, 3 from King John, 4 from 1 Henry IV, 3 from 2 Henry IV, 5 from Henry V, 1 from Richard II. Either Shakespeare’s Histories were a staple part of their repertory, or the actors remembered them well. The 12 comedies they drew on produced 32 borrowings, more from the later 90s: 5 from The Merchant of Venice and from Twelfth Night, 3 each from The Merry Wives of Windsor, As You Like It, and Measure for Measure. From the tragedies they recalled Titus Andronicus and Julius Caesar three times each, Romeo and Juliet seven times, and Othello thirteen times, a frequency that raises again the problem of dating. Q1 was entered in the Stationers’ Register on 26 July 1602, while Othello is conventionally dated to 1603. Having noted five of these echoes, Alfred Hart suggested that Othello must have been produced “not later than the early months of the year 1602” (p. 401). Hart found it improbable “that the Shakespeare of Hamlet and Othello borrowed lines from the pirated editions of his own plays”, and his verdict has been supported in modern times by E. A. J. Honigmann, in arguments most conveniently available in his Arden edition of Othello (1997). 


I conclude that the unauthorised first quarto of Hamlet was not an early draft by Shakespeare, but a collective reconstruction made, for whatever purpose, by members of the King’s Men in the spring and summer of 1602.




Spring Attractions from the Shakespeare Guild

The Shakespeare Conference: SHK 30.094  Thursday, 7 March 2019


From:        John F. Andrews <This email address is being protected from spambots. You need JavaScript enabled to view it.>

Date:         March 6, 2019 at 6:53:50 PM EST

Subject:    Spring Attractions from the Shakespeare Guild




A Special Evening with the Extraordinary Jim Dale


Wednesday, March 13, at 7 p.m.

The Players

16 Gramercy Park South, Manhattan

Free, but Reservations Required


In the 1950s and '60s, Jim Dale was known primarily as a singer and songwriter, composing such hitas the theme song for Georgy Girl,” which was nominated for an Academy Award. Meanwhile he was garnering plaudits as a film and television comic, with eleven Carry On features that made him a celebrity in Britain. Next came stage roles like Autolycus and Bottom with Olivier’s National Theatre Company, and Fagin in Cameron Mackintosh’s Oliver. In 1980 Mr. Dale collected a Tony Award for his outstanding performance in the title role in Barnum. Since then he has been nominated for TonyDrama Desk, and other honors for his work in plays such as CandideJoe EggMe and My Girl, and ScapinoAnd he's earned two Grammy Awards as the “voice” of Harry Potter. We look forward to a memorable gathering with one of the most versatile artists in entertainment history. 



Speaking of Shakespeare with Director Ethan McSweeny


Thursday, March 14, at 8 p.m.

The National Arts Club

15 Gramercy Park South, Manhattan

Free, but Reservations Requested


A director whose productions have been lauded by leading critics as “daring” and “impeccably stylish,” Ethan McSweeny has recently accepted a new position as Artistic Director of the American Shakespeare Center in Staunton, Virginia, where he’ll be reviving Elizabethan and Jacobean classics in a beautiful replica of London’s storied Blackfriars playhouse. Mr. McSweeny has earned eight Helen Hayes Awards for his work at the Shakespeare Theatre Company and other venues in Washington. He has also directed in such settings as the Guthrie Theater in Minneapolis and the Stratford Festival in Ontario. In 2000 his Broadway revival of Gore Vidal’s The Best Man won Drama Desk and Outer Circle accolades and was nominated for a Tony AwardPlease join us for a lively conversation with the artist American Theatre has lauded as a  “wunderkind with a Midas touch.



The Ides of March at Edwin Booth's Final Home


Friday, March 15, at 2 p.m.

The Players

16 Gramercy Park South, Manhattan

Free, but Reservations Required 


As part of a new Afternoon Salon series for NAC members and their friends, the Guild has arranged a special tour of the historic club that adjoins it on Gramercy Park South. Our guide will be Raymond Wemmlinger, who oversees the Hampden-Booth Theatre Library and presides over the Players Foundation. Ray is an authority on America’s foremost acting family, and he’ll show us playbills, costumes, portraits, and other artifacts from the legendary Edwin Booth, who lived in this elegant building from 1888 until his death in 1893. In Edwin’s bedroom we’ll see a photograph of John Wilkes Booth, who restaged a “lofty scene” from Julius Caesar five months after he, Edwin, and Junius Brutus Booth the younger (named after their famous father) had presented that Roman tragedy as a benefit for the Shakespeare Statue that now adorns Central Park.   



You may reserve for these events with an email to This email address is being protected from spambots. You need JavaScript enabled to view it.And you’ll find details about them and additional programs at www.shakesguild.orgComing attractions will include a salon with filmmaker Melinda Hall (2 p.m., Tuesday, April 23, at the NAC), a staging of Jonathan Richards and Lois Rudnick’s Ever the Twain: William Shakespeare in Mark Twain’s America (7 p.m., Tuesday, April 23, at The Players), a salon with Artistic Director Bonnie J. Monte of the Shakespeare Theatre of New Jersey (2 p.m., Wednesday, April 24, at the NAC), a conversation with actor John Douglas Thompson, who’ll be playing Kent in Broadway’s King Lear (2 p.m., Tuesday, May 14, at the NAC), and a special evening with actress Michael Learned, who’ll discuss her roles in television’s The Waltons and in such theater classics as Edward Albee’s Three Tall Women (8 p.m., Tuesday, May 14, at The Players).     






Subscribe to Our Feeds


Make a Gift to SHAKSPER

Consider making a gift to support SHAKSPER.