Current Postings

The Shakespeare Canon and the NOS

The Shakespeare Conference: SHK 29.04408  Sunday, 16 December 2018

 

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

     Date:         December 15, 2018 at 2:39:33 PM EST

     Subj:         Re: SHAKSPER: NOS 

 

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

     Date:         December 16, 2018 at 10:42:36 AM EST

     Subj:         Re: SHAKSPER: NOS 

 

 

[1]-----------------------------------------------------------------

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

Date:         December 15, 2018 at 2:39:33 PM EST

Subject:    Re: SHAKSPER: NOS

 

Pervez Rizvi makes this point:

 

“As with the business of the missing NOS data files, Egan makes it sound like he’s just a bystander. Taylor’s article was published in 2014 but he tells us in a footnote that Egan had read and commented on it before then. The two scholars then edited the Authorship Companion together. They allowed about half a dozen chapters in the book to use the microattribution method, to support their attributions. Was it not irresponsible of them to do that, when Egan knew that the method had not been properly validated?”

 

That is certainly arguable; but who is to say what transpired in the editorial discussions between Gabriel and Gary Taylor. As the senior editor, Taylor was probably in a position to make the final decision, which the entire editorial team would be obliged to support as the “edition’s” product. For this reason, I believe that scholarship is best served by identifying a single person as the “editor,” as Riverside did with its collection attributed to G. Blakemore Evans as “general editor,” even though several other scholars edited the particular plays and poems or contributed magisterial articles. That practice allows the other participants the freedom to disagree with editorial choices in which they themselves did not participate. Of course, editors could always change their minds, but it would be best to avoid controversy by making that explicit.

 

A related point: I feel strongly that recurrent textual and orthographic issues should be resolved the same way in all cases throughout the Canon, and this choice should be referenced in the “Editorial Practices” section of the prefatory material. So, for example, if the general editor prefers not to modernize “murther” to “murder,” the reader knows that when she sees “murder” that is the way the word is spelt in the copytext. Consistent practice was the expressed rule in, for example, Riverside; I am under the impression that NOS leaves many such matters to the discretion of the individual editors. I could be wrong about that; in which case any variants in such matters should probably be attributed to failure to precisely proofread the texts. 

 

[2]-----------------------------------------------------------------

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

Date:         December 16, 2018 at 10:42:36 AM EST

Subject:    Re: SHAKSPER: NOS

 

Letter to Shaksper 16.12.18

Gentle Shaksperians,

 

In the Humanities, as in the Natural and Life sciences, it is always pleasing when two or more independent research projects produce the same result. This forum has received my report on Pervez Rizvi’s demolition of the microattribution method espoused by Gary Taylor and colleagues, and his convincing rebuttal of Gabriel Egan’s apologia for it. I can now report on a parallel study by Darren Freebury-Jones and Marcus Dahl, “The Limitations of Microattribution”, 

 

https://www.utexaspressjournals.org/doi/pdf/10.7560/TSLL60404

 

which discusses the essay by Taylor and Nance, “Imitation or Collaboration? Marlowe and the Early Shakespeare Canon”, Shakespeare Survey 68 (2015): 32-47. Freebury-Jones and Dahl make three main points, starting with the paucity of data presented. The NOS authors allowed themselves more latitude than Taylor’s 63 words in Macbeth, choosing 173-word excerpts, but still believed that a mere half-dozen verbal parallels suffice to make a reliable authorship attribution. Taylor and Nance deny the evidence that Thomas Kyd wrote the major part of 1 Henry VI, awarding that honour to Marlowe. They analyse the first 173 words of Joan la Pucelle’s speech in 5.2, finding 9 matches with Marlowe and not one with Kyd. Freebury-Jones and Dahl cite five matches with Kyd’s plays, two each with Soliman and Perseda and Cornelia, and one with The Spanish Tragedy. As with other NOS attributions, the search for matches is biased in favour of the chosen author, either Shakespeare or Marlowe, and always against Kyd.

 

This may seem only a local issue, but a bigger principle is at stake. Freebury-Jones and Dahl dismiss these samples as “Lilliputian”, quoting a scientist’s caveat that “the smaller the studies conducted in a scientific field, the less likely the research findings are to be true”, since “small sample size means smaller power”. The NOS group can’t have it both ways: if they're going to base their attributions on purely quantitative measures they’ll have to respect the laws of statistics. And in fact, as David Auerbach has shown, the Authorship Companion contains several elementary errors in statistics by Egan, Craig and Burrows, Taylor, and Loughnane.

 

In the second stage of their critique Freebury-Jones and Dahl take up the challenge of replicating the NOS method, searching plays performed between 1576 and 1594. They selected the first 173 words from 1.2 of Kyd’s The Spanish Tragedy for analysis, and discovered five unique matches with Marlowe, two from 2 Tamburlaine and one each from 1 Tamburlaine, The Jew of Malta, and The Massacre at Paris. Kyd himself only provided one unique word-string matching this sample, the phrase “Except some fewe that”, which occurs in Soliman and Perseda and Cornelia. As they point out, by the criteria of Taylor and Nance, this speech would be denied its true author. Freebury-Jones and Dahl then examined a 173-word passage in 4.1 of Kyd’s Soliman and Perseda, with the same result: eight matches with Marlowe outscored three with Kyd. Even after additional tests, as they put it, “No matter how we interpret these overall findings … Marlowe would be considered the more likely author on the basis of Taylor and Nance’s criteria”. To discover that their preferred method cannot reliably indicate the true author of a play should be rather chastening for the NOS project, in which, as Taylor told the Guardian in October 2016, where its 1986 predecessor had only claimed 8 plays as co-authored, it had increased that total to 17. Shrinking Shakespeare while aggrandizing their importance as self-proclaimed authorities is a perversion of textual editing.

 

The final stage in Freebury-Jones and Dahl’s critique of quantitative methods draws on the remarkable work of Martin Mueller, a classicist, a Shakespeare scholar, and a pioneer in what he calls “digitally assisted text analysis”. Mueller was co-creator of the wonderful Chicago Homer interactive website, marked up to facilitate the study of repetitive or formulaic language in the Iliad and Odyssey. The site identified 14,000 phrases that vary in length from two words to twenty lines, marking the shift from oral to written composition, with a total of 34,000 occurrences. Drawing on this experience, about a decade ago Mueller began to create an interoperable digital corpus of non-Shakespearian plays called “Shakespeare his contemporaries” (see  http://www.digitalhumanities.org/dhq/vol/8/3/000183/000183.html)

 

Being an old-fashioned scholar, more concerned with advancing knowledge than boosting his own reputation, Mueller’s data was available only on his website. It included a list of all the n-grams in nearly 500 plays, arranged pairwise: that is, for each play Mueller listed all the other plays which showed a significant number of matches. Freebury-Jones and Dahl have extracted from this database the plays in Marlowe’s canon, listing 21 plays having strong affinities with Edward II [1592]. The five leading plays are The Massacre at Paris [1593], followed by Sir John Oldcastle [1599], Alphonsus [1587], Thomas Wyatt [1602], and The True Tragedy of Richard III [1589]. The striking feature of this data, as Freebury-Jones and Dahl point out, is that 

 

there are a large number of matches with authors such as Anthony Munday, John Fletcher, Peele, and Kyd, for plays which they are not supposed to have written. The relatively high number of matches between these dramatists is very close and were the question of authorship to be answered purely quantitatively, would not reveal very likeable results for the traditional canon e.g. Peele co-wrote the Tamburlaine plays; Lyly co-wrote Doctor Faustus; Fletcher The Jew of Malta; and Munday Edward II. We could also argue on this basis that Marlowe had a hand in Shakespeare’s Richard II and Richard III.

 

In other words, attributions based solely on quantitative measures would be grossly inaccurate. It seems obvious to me that the corrections of such errors can only be supplied by qualitative methods, which treat plays as individual artefacts, produced in a historical context. In the five plays listed above that share a significant number of n-grams with Edward II, I have added the dates allotted by Martin Wiggins in British Drama 1533-1642: A Catalogue, which reveal that Greene’s Alphonsus and the anonymous True Tragedy antedate Edward II. This fact shows the importance of chronology, regularly ignored in quantitative studies, for on the face of it Marlowe was indebted to those plays. Qualitative analysis would involve examining each of these matching n-grams to see if it constituted a reliable authorial marker—a criterion not often met by Taylor and Nance, as Freebury-Jones and Dahl have observed—to establish Marlowe’s indebtedness, if any.

 

This is the second study refuting the NOS microattribution method that I have reported on here, and another is being peer-reviewed. In a previous posting Egan advised SHAKSPER readers that if they hear “that Rizvi has now devastated the scholarship on which the New Oxford Shakespeare is based” they should “read Rizvi’s articles and the forthcoming rebuttals rather than taking that summary on trust.” In the three reports that I have so far published I have provided links to the essays concerned, so that readers can see for themselves whether my summaries are reliable. I have presented the counter-arguments in a non-technical manner, bringing out their wider significance and giving my own honest verdict. I will of course report on “the forthcoming rebuttals” with the same candour, as and when they appear. 

 

They conclude that quantitative attribution studies are unreliable unless they use large samples, a finding that other recent studies would endorse.

 

Warm regards,

Brian

 

 

 

The Shakespeare Canon and the NOS

The Shakespeare Conference: SHK 29.04407  Saturday, 15 December 2018

 

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

Date:         December 15, 2018 at 6:05:52 AM EST

Subject:    Re: The Shakespeare Canon and the NOS

 

All except one of the points Gabriel Egan makes in his latest post were ones he also made last time, and I’d respectfully ask people to read my response to them from last time. The exception is his following comment:

 

I agree with one aspect of Rizvi’s critique of Taylor’s actual method...I made exactly that point—that I would want more validation runs before drawing such a conclusion about method—when I reviewed Taylor’s article in The Year’s Work in English Studies 95 (2016): 402-405. 

 

As with the business of the missing NOS data files, Egan makes it sound like he’s just a bystander. Taylor’s article was published in 2014 but he tells us in a footnote that Egan had read and commented on it before then. The two scholars then edited the Authorship Companion together. They allowed about half a dozen chapters in the book to use the microattribution method, to support their attributions. Was it not irresponsible of them to do that, when Egan knew that the method had not been properly validated?

 

 

 

The Shakespeare Canon and the NOS

The Shakespeare Conference: SHK 29.04406  Friday, 14 December 2018

 

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

Date:         December 13, 2018 at 1:39:16 PM EST

Subject:    Re: SHAKSPER: NOS

 

Dear SHAKSPERians

 

Pervez Rizvi reckons that “the [microattribution] method probably is what Egan tells us it is, but Taylor described it wrongly in his SQ article”.  I’m glad to hear this, since it means that whatever other merits Rizvi’s article possesses it cannot stand as a demonstration (by replication) that the microattribution method doesn’t work. By definition, refutation-by-replication requires one to replicate the method. Rizvi has shown that the method he used, which led to 759 matches, doesn’t work; but refuting that method is not a comment on the microattribution method that led to 18 matches. In his article, Rizvi could have claimed that Taylor’s phrasing should have been clearer about just what he did, and I’d agree to that. But then I’m always saying to people in this field that they need to be clearer about exactly what they did.

 

I agree with one aspect of Rizvi’s critique of Taylor’s actual method, when he complains that Taylor switched from counting tokens to counting types upon finding that a single validation run showed that counting tokens gives the wrong result. I made exactly that point—that I would want more validation runs before drawing such a conclusion about method—when I reviewed Taylor’s article in The Year’s Work in English Studies 95 (2016): 402-405. (This review is available in Open Access form on my website.) I also made other criticisms of Taylor’s article, which I won’t mention here because I’d like those who are interested to go to the review and see for themselves that I did. I point this out because it has been claimed on this list that I never criticize what my fellow New Oxford Shakespeare editors have published. I do make such criticisms and I welcome the same in return. That’s how we progress.

 

I’ll end with a point of substance rather than commentary on how we argue. Rizvi writes that he “used exactly the attributions I state in the files he [Egan] downloaded from my website, i.e. the attributions I got from S[hakespeare H[is] C[ontemporaries]” (that is, Martin Mueller’s moribund project of that name). Since these attributions are, by common agreement, wrong in claiming Shakespeare’s sole-authorship of 1, 2, 3 Henry VI, Henry VIII, Pericles, The Two Noble Kinsmen, Timon of Athens, and Titus Andronicus, how did Rizvi avoid giving (in his tables of whose canons are being hit by his 759 matches) Shakespeare the credit for what other men wrote? I’m not asking about his calculations of overall canon size—I accept of course that even when we factor in co-authorship Shakespeare’s is still the largest canon—but rather I’m asking about the counts that underpin his tables. Surely the counts of whose canons are being matched will be wrong if they are based on attributing to Shakespeare substantial chunks of writing that we all agree aren’t his.

 

Regards

Gabriel Egan

 

 

 

MM Ending Query

The Shakespeare Conference: SHK 29.0445  Friday, 14 December 2018

 

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

     Date:         December 13, 2018 at 12:55:08 PM EST

     Subj:         Re: SHAKSPER: MM Ending 

 

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

     Date:         December 14, 2018 at 10:30:56 AM EST

     Subj:         MM Ending 

 

 

[1]-----------------------------------------------------------------

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

Date:         December 13, 2018 at 12:55:08 PM EST

Subject:    Re: SHAKSPER: MM Ending

 

I notice the lengths that many commentators go to try to preserve the view that Shakespeare always attempts deep psychological and spiritual efforts to draw characters in his plays.

 

Can’t we see that Shakespeare has in Measure for Measure given a tour de force in order to be able to spin a narrative in which the judgments that characters make are later visited on them, giving them a taste of what it’s like to be in the other person’s predicament. In this, the character and the play’s audience learn humility in their intensely seeing and feeling the other person’s challenges as their own. This is a viewpoint actually drawn in Shakespeare’s, 2henryIV, in which the Chief Justice urges Henry V in his making a judgement to consider the following: “Question your royal thoughts, make the case yours.”

 

This I think is a worthy goal of a dramatist, which I think succeeds in MM in that it is not so obvious that Duke Vencentio is actually tampered with to become the dramatist’s pawn to create this dramatic world in which all these reversals can happen.

 

[2]-----------------------------------------------------------------

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

Date:         December 14, 2018 at 10:30:56 AM EST

Subject:    MM Ending

 

Thanks to all who responded to my questions about MM. Though I have seen many productions (the most recent being last weekend’s production of MM at FDU), I cannot comment usefully on the various performance choices listed by Alan Dessen. Nor can I illuminate contemporary references in the plays, such as Thomas Krause proposes. As to Brian Bixley’s quotes from MM and the Odyssey on the discomforts of death, I can only say that I am old enough to feel their force.

 

I especially appreciate Robert Projansky’s thoughtful reading, particularly as he traces the difficulty of Isabella’s final act of mercy after the violence of her initial reaction: 

 

When she is first told in IV iii that Angelo has betrayed the bargain and has had Claudio executed, her natural reaction is to exclaim

 Isa.

Oh, I wil to him, and plucke out his eies.

 

By contrast with this reaction, reminiscent of Beatrice’s violent desire to eat Claudio’s heart out in the marketplace in MAdo, Isabella’s final act of mercy clearly signals her transformation. But Robert Projansky does not credit the Duke with a similar change or any change at all:

 

the Duke doesn’t show much in the way of growth and change over the play

 

This assumption lies behind readings of the Duke as a mere plot device or an omniscient manipulator. I would suggest that it also flies in the face of the rest of the canon. In no other mature play that I can recall does a major character remain static from beginning to end. In order to maintain that view we must discount the Duke’s “Be absolute for death” speech, as Projansky suggests. But there is plenty of other evidence that the Duke is a human, flawed character, (“an advocate for death”), before encountering Isabella. 

 

Among the ironies of the Duke’s monk’s disguise is his ignorance of Catholic doctrine, specifically in confusing a “sin” with the “sinner.” Even the novice Isabella knows better and makes the distinction the basis of her argument with Angelo. In their first encounter she pleads:

 

I have a brother is condemn’d to die;

I do beseech you, let it be his fault,

And not my brother.

 

Angelo responds with the necessary secular view:

 

Condemn the fault, and not the actor of it?

Why, every fault’s condemn’d ere it be done:

Mine were the very cipher of a function 

To fine the faults …

And let go by the actor. (2.2.34-41)

 

At this refusal, Isabella accedes to the “law” and turns to go. Her only prepared petition was to advance the Church’s spiritual view that sinners should be punished for their sins but must not be identified with them.  Like Angelo—even going him one better—the Duke not only condemns the sinner but also condemns the innocent fruit of fornication. “Repent you, fair one,” he asks Juliet, “of the sin you carry?” The metaphor is telling. (See too his curious mentions of children and youth in 1.3.) Might not this confusion explain the Duke’s reluctance to impose order on Vienna, for to eradicate the “sins” would require “gelding and splaying all the youths of the city” and ultimately to impose the precise sentence levied on Claudio?

 

And the Duke’s badgering interview with Juliet also exposes his callousness. After she reiterates her repentance, he parts from her with this speech: 

 

                                    There rest.

Your partner, as I hear, must die tomorrow,

And I am going with instruction to him.

Grace go with you. Benedicite!  Exit.

Juliet

Must die tomorrow! O injurious love,

That respites me a life, whose very comfort

Is still a dying horror!

 

Is this the consolation that a friar brings, the “charity” that the Duke told the Provost motivated his visit? To drop the news of Claudio’s impending death and then walk out?

 

Finally, there is the parallel between the Duke and Angelo to be considered. For both, their “substitute selves” protect them from slander, as the Duke explains in 1.3 (40-43) and Angelo echoes in 4.4: 

 

For my authority bears so credent bulk

That no particular scandal once can touch

But it confounds the breather (24-26).

 

In addition to their reclusive natures, the Duke and Angelo have in common, for different reasons and to different degrees, the need for a disguise. I find the Duke’s reiterated fear of slander to be an important aspect of his character. Given the importance of forgiveness in other Shakespeare plays, from the deft hints in KJ through the mutual forgiveness of Hamlet and Laertes to Prospero’s sweeping pardons, the forgiveness the Duke extends to Lucio for his slanders signals a significant change. For one obvious thing, it confirms that the Duke has disentangled his personal feelings from his governing responsibilities; Lucio’s slanders are forgiven, but his treatment of Kate Keepdown is not. Publicly and in his own person, the Duke brings to each of the couples a just, if often painful, resolution in the generically comic ending.

 

There is another less immediate reason that I locate the Duke’s change in 3.1. Folio MM has seventeen scenes; the Duke introduces the bed trick in the scene immediately following the central scene, the position in which the narrative turn and psychological changes occur in most other mature plays: Falconbridge’s “thousand businesses,” Othello’s “seduction,” the initiation of the bed trick in AWW, Imogen’s male disguise, etc. I realize that mere assertion does not make this claim convincing and that even to isolate the “central scene” of most plays is to invite dispute. But I have defended the claim at length elsewhere, and I mention it here only to point out a seldom remarked but pervasive feature of Shakespeare’s construction.

 

These observations do not, of course, exhaust the Duke’s character. But I hope they encourage a closer look at its interiority, perhaps along the lines remarked by Robert Polansky and by Harry Keyeshian’s acute observation on the similarity of MM and WT.

 

 

 

Boycott International Journal of English and Cultural Studies

The Shakespeare Conference: SHK 29.0444  Friday, 14 December 2018

 

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

Date:         December 13, 2018 at 1:08:43 PM EST

Subject:    Re: SHAKSPER: Boycott

 

Good of John Cox to bring this up. I get multiple solicitations a month from the likes of (making this up) The Journal of Interdisciplinary Humanities Research. This pay-to-play faux academic journal thing is a big industry. 

 

I’m thinking older scholars, who haven’t grown up steeped in and savvy to the internet culture and techniques of fraud and deception, might be particularly susceptible. But youngsters should watch out too. Publishing in these journals won’t do much, or anything, for your tenure track, will they?

 

 

 

Subscribe to Our Feeds

Make a Gift to SHAKSPER

Consider making a gift to support SHAKSPER.