CONTENT RATING: universal

Did Shakespeare really write, well, Shakespeare? Or is the Swan of Avon a five century old con? In the last episode of this slowburn mini-series, some big names in the field of Shakespeare attribution try their own hands at penning dramatic texts. Below you will find data, audio credits, further reading, and a transcript of the podcast.

Audio credits

Kai Engel – Difference
Kai Engel – Vintage Frame
Scott Holmes – Postcards
Aaron Dunn – Minuet – Notebook for Anna Magdalena (cropped)

Transcript

Case S02E08: Slowburn Shakespeare, part 6 of 6 – The stars are fire

By the time you’ve reached this point in the miniseries – the last episode about Shakespeare… probably – you might imagine that there really can’t be much to add to the debate around Shakespeare’s authorship. We’ve looked at the apocrypha – the plays that some want to include in the canon, and others want to exclude from it. We’ve looked at the canon itself, and the idea that someone other than the Stratford glover’s son, William Shakespeare, wrote them. We’ve gone through four individual candidates – Francis Bacon, Edward de Vere, Christopher Marlowe, and William Stanley. Whilst some of these have seemed faintly possible, or at least, a lot of fun, none of them has really reached a quantum of evidence sufficient to cast enough reasonable doubt to convince enough people. That evidence has taken the shape of biographical analyses – both comparing life histories against documents that survive from the times, and trying to divine an authorial profile from the content of the plays themselves. We’ve had cipher-based decryption – that is, attempts to hunt down secret messages hidden in Shakespeare’s writings. We’ve even had document analysis – aging and dating papers to see if they match the era they are supposedly, and we’ve had handwriting analysis, and even genetic testing.

But of course, the sort of evidence that listeners of this podcast are perhaps most interested in, and the kind that’s actually been fairly thin on the ground to press, is the language. Linguistic analyses. We’ve had a little, of course, but even some of that has sometimes been pretty ropey. In the weakest cases, various Antis have had nothing more concrete to point to than textual allusions, hints in poetry, clever plays on words with subtly alternative meanings, parallels between plays and lives. We have had some more slightly more compelling insights. In the strongest cases we have really marked similarities. Remember that example from the fourth part of the miniseries, two episodes ago? Listen again to these two lines from Christopher Marlowe:

But stay! What star shines yonder in the east?

The lodestar of my life, if Abigail!

And now compare them with these two lines from Shakespeare:

But soft! What light through yonder window breaks?

It is the East, and Juliet is the sun!

Extremely interesting, but it doesn’t quite a murder make, though.

We even looked into early computational linguistic analyses done by Claremont Colleges, and there, we stopped. It would be reasonable to imagine that by this stage, there cannot be anything left to say on the topic. No stone unturned. No corner left to peek round. But you’d be wrong. I’ve barely scratched the surface thus far and there’s a last huge mountain that we haven’t even started up.

Before we get there though, it’s also useful to notice that we could go back over a lot of the old ground and plough it again in an entirely new way. Perhaps the best thing about authorship analysis is that it’s a very broad church. If you have a reasonable chunk of digitised language, there are innumerable ways that you can come at it, and the modern era gives us advantages that the scholars of the 1800s and 1900s and even most of the 2000s simply didn’t have. Computing. Yes, you would be right to point out that Claremont Colleges also had computing, but I would argue that theirs was from an era that has now faded into relative antiquity, just as the computers of today will seem prehistoric in only a couple of decades. Our current machines have faster processors, and most importantly, they have friendlier software. The kinds of interfaces accessible to people whose expertise is in Shakespeare, rather than computing. That’s a big shift, and an important one. Computer-experts-who-happen-to-analyse-language can still do their thing but now the field is also open to linguists-who-happen-to-use-computers.

And it’s from the confluence of these ultra-modern phenomena – the digitisation of older texts, the radical advancement of computing, and the creation of widely accessible software – that we arrive at our last step in this story.

Welcome

Welcome to en clair, an archive of forensic linguistics, literary detection, and language mysteries. You can find case notes about this episode, including credits, acknowledgements, and, far more than usual, many extra links to further reading at the blog. The web address is given at the end of this podcast.

Joy’s soul

You thought we were heading into 2016. Wrong. As usual, before we go forward, Marty, first, we must go back. Thirty years back. Our story starts in in 1986, when Oxford University Press – OUP – published, The Oxford Shakespeare. One volume is William Shakespeare: A Textual Companion, one is William Shakespeare: An Old-Spelling Edition – does pretty much what it says on the tin. And the one we’re most interested in, the headlining volume, is The Complete Works. As you should have already realised by now, just the name itself is automatically, if implicitly, a statement about authorship, about the apocrypha, about the canon, about what a collection of people have decided to include or exclude based on sometimes fairly thin evidence. As seems to just be the case in this field, this new volume raised fierce debates since it tried to replicate plays and poems, not as they were printed, but as they were performed. And as I’ve discussed right back at the start of this miniseries, this is like taking a screenshot during a video, and deciding that this single freeze-frame is a fair, canonical representation of the whole video. This snapshot is Hamlet. This snapshot is Macbeth.

Lots of people were Not Happy with this performance-over-print decision because it meant, in some cases, substantial changes to comfortable, long-established ideas. Hamlet had whole speeches relegated to an appendix. Can you hear people falling onto their fainting couches? King Lear had to be put in twice because it came down to two very different versions, and it wasn’t clear enough to the editors which one had the greatest claim to being anointed the One True King. And perhaps most scandalous of all, some of the works had the names of collaborators included. Other co-authors. Thomas Middleton. George Wilkins. John Fletcher. Anonymous. For some absolute Shakespeare purists, this would have been enough to trigger a flurry of outraged letters to newspapers.

Anyway, this first edition of The Complete Works in 1986 was edited by John Jowett, William Montgomery, Gary Taylor – we’ll come back to him – and Stanley Wells. If you have an outstanding memory, you might recall Stanley Wells from this rather interesting comment, which was part of a longer article that vanished first from the Shakespeare Birthplace Trust site, and then from the Royal Shakespeare Company site:

The phenomenon of disbelief in Shakespeare’s authorship is a psychological aberration of considerable interest. Endorsement of it in favour of aristocratic candidates may be ascribed to snobbery – reluctance to believe that works of genius could emanate from a man of relatively humble origin – an attitude that would not permit Marlowe to have written his own works, let alone Shakespeare’s. Other causes include ignorance; poor sense of logic; refusal, wilful or otherwise, to accept evidence; folly; the desire for publicity; and even (as in the sad case of Delia Bacon, who hoped to open Shakespeare’s grave in 1856) certifiable madness.

To me it’s a little difficult to reconcile this statement with being the editor of a series of volumes that then goes on to explicitly reshape the authorship of Shakespeare by suggesting a range of collaborators. I imagine if Wells cared what I think, which I expect he doesn’t, I’m guessing he might say that his contention is with people who reject Shakespeare entirely and substitute their own reality. But yet, this edition was obviously open to the idea that authorship in works from this era is not as cut and dried as lots of people imagine. If it were crystal clear, there wouldn’t be plays marked as collaborations with multiple unidentified writers. This takes us right back to the cut-out theory – or one version of it, anyway – from the very start of this miniseries.

Remember the group theory had roughly three versions. Version one: Shakespeare doesn’t exist at all. The name is just a pseudonym that a bunch of people are using. He is as much a fiction as the three witches. Any similarities to people living or dead are purely coincidental, etc. etc.. Version two: Shakespeare exists, but he’s just the theatre manager, and he’s letting the writers use his name because reasons. Version three: he is either a part of, or the leader of a group of writers all publishing under his name. Bingo. Or as I mentioned earlier, collaboration. Some collaborate on some sets of plays. Some pen the lines for certain scenes or characters. Shakespeare reimagines something of theirs. They adapt something of his. Sometimes it’s just one author. Sometimes a few are free at the same time and all work together at once. But essentially, you end up with other, and sometimes very famous, co-writers working alongside Shakespeare without necessarily getting the credit for it later on down the line.

When the new edition of this series came out, the venture into this authorial territory – efforts to fill in blanks on the Shakespeare map, if you will – went even further. The 2005 Complete Works volume added in the plays, Sir Thomas More and Edward III, not because Shakespeare wrote them, but because both contained a few passages that might be by Shakespeare. In other words, he was, at best, second author. Maybe.

And now, at long last, we move forward to 2016, and OUP’s release of their top-of-the-line, ruinously expensive series, The New Oxford Shakespeare. After its predecessor, some may have been well-prepared for what was to come, but after the three-volume 1986 version, this new three (soon to be five) volume update included an entire 776 page book entitled, The New Oxford Shakespeare: The Authorship Companion, by Gary Taylor – again, we’ll come back to him – and Gabriel Egan. As the title suggests, in it, Taylor and Egan pick up this now ancient question: did Shakespeare write all the works attributed to him?

Actually, I should be fair. Their angle is perhaps better defined as, “Did Shakespeare write at least some of this play? Like, even just a bit of it. Couple of verses maybe? Was anyone else involved? Who did which bits? Why is this such a mess?” Okay maybe not that last question. That’s probably just people like me who mutter that when analysing data. Anyway, theirs isn’t automatically the “people hiding behind pseudonyms to avoid execution” subterfuge version of the question. Theirs is, flat out, someone saying, “yeah but this manuscript doesn’t even have a name on it, and someone’s patently added stuff in different handwriting afterwards so obviously other people were involved, soooo is this part of the collection or no?”

And, echoing the old 1986 series, the headlining volume of the new edition is, to give its full name: The New Oxford Shakespeare: The Complete Works – Modern Critical Edition. I guess it isn’t properly monumental unless it has a title, a subtitle, and a subsubtitle. Anyway, this particular volume lists the editors as Gary Taylor, John Jowett, Terri Bourus, and Gabriel Egan. Stanley Wells doesn’t appear but I assume that this was because he was in his eighties when this new edition was being put together, and he’s 91 as I record this. It did strike me as interesting, though, that in his absence, as implied by the existence of the Authorship Companion, this edition goes even further still with the authorship question, by including even more possible co-authors, adapters, and revisers.

The 2016 Complete Works includes fifty-three titles. Remember that the First Folio contained only thirty-six, and the canon generally stands at thirty-eight, so that’s already fifteen more. These works are mostly plays, some poems and so forth, and thirty-seven of them have no author listed at all. Implicitly, the editors do not challenge Shakespeare’s authorship, either because the evidence to do so is insufficient or there simply isn’t any. But in sixteen cases now, both Shakespeare and other names are listed, either as people who have collaborated on that work, or as people who have subsequently adapted that work. And the ordering of the names is indicative of who is deemed to have contributed most.

So how does this list look? Well, let’s take some heavyweight classics: A Midsummer Night’s Dream, The Comedy of Errors, The Tempest, Romeo & Juliet… all the ones I really hated. According to the New Oxford Shakespeare, these are all Shake, no fake. And as I’ve said, there are like thirty-three more titles in that list.

But what about the ones with other authors. Well, there’s The Tragedy of M. Arden of Faversham. This was supposedly mainly authored by… Anonymous. Yeah, I know. But also by Shakespeare. So that’s new. What about Titus Andronicus? That one’s attributed to Shakespeare first, Peele second, and it might even have an added scene possibly by Thomas Middleton, question mark. What else. Apparently Marlowe and Anonymous collaborated on all the Henry VI titles. Nash comes in on the third. And remember our Shakespeare apocrypha? One of those titles was The Passionate Pilgrim? Well that’s also included, but it’s a messy one. Authors are listed as Shakespeare, Barnfield, Griffin, Deloney, Marlowe, Raleigh, and Anonymous, which sounds like really high-interest-rate credit card. Let’s do some big hitters. Measure for Measure, All’s Well that Ends Well, Macbeth. These are all listed as adapted by Middleton. Perhaps he had a talent for jazzing up plays. Who knows. Last few. You may have forgotten this by now, but remember I said how, just after his death, Shakespeare was overshadowed for a while by that pair, Beaumont & Fletcher, who are another fun little authorship case in their own right? Well, apparently, Fletcher is first author, and Shakespeare second, on The History of Cardenio, Henry VIII, and The Two Noble Kinsmen.

And for me at least, this is a significant shift in a huge tectonic plate that has, historically, resisted almost any attempts at being moved. It contradicts those quarters who have previously insisted that having the audacity to doubt Shakespeare’s authorship, especially within the canon – plays like Macbeth – was embarrassing, misconceived, and even a “psychological aberration”, to quote some of the harsher critics. In the Complete Work’s brief discussion on authorship, the editors write that,

Most of the works in this Modern Critical Edition were written entirely by Shakespeare. But a significant number of his plays were written collaboratively, usually with one other playwright as the subordinate partner. […] Some early modern plays were first written by one author, but later adapted by someone else for a theatrical revival. Shakespeare himself sometimes adapted other men’s plays: this edition identifies him as the adapter of three plays originally written by other playwrights. […] But whenever this edition is the first to identify a collaborator or adapter, we treat the attribution more tentatively. […] In this way, we […] distinguish between the most secure and well-established attributions and those that are more recent—and therefore, intrinsically, not yet as widely known, widely tested, or widely endorsed. […] In cases of collaboration, adaptation, or disputed authorship, the introductory box always provides a cross-reference to the relevant chapters of the Companion.

I have enough respect for OUP’s general credentials – their peer reviewers, editorial processes, quality controls, and so forth – to believe that they would not publish absolute snakeoil. Not intentionally, anyway. For me, if OUP thinks this series is worth being published, then in and of itself, its very existence and its undeniable focus on authorship provides evidence that there is indeed room for reasonable doubt, not just around the apocrypha and the lesser works, but even around the canon. I should add, however, that I don’t think it should require a publication like this to exist before anyone is authorised to ask the question for themselves. I stated this position at the start anyway. If someone wants to question gravity, I feel that’s their fundamental right. Heck, take on the findings of Stephen Hawking. Challenge the idea of time. Get stuck in. Just do it well.

So let’s get back to the 2016 New Oxford Shakespeare with its handful of authorship reattributions. The New York Times hailed its arrival with a relatively muted piece, opening with the tentative lines:

Shakespeare may have had a little more help than previously suspected.

Not exactly throwing this out as a breaking world exclusive, are they. The BBC was faintly more enthused:

Scholars working on New Oxford Shakespeare […] said his collaboration with other playwrights was more extensive than has previously been known. The research, by 23 international scholars, has identified 17 of 44 Shakespeare plays as being co-written with other authors. The new research involved both traditional textual analysis and the use of computerised tools to examine the scripts. Gary Taylor, one of the New Oxford Shakespeare’s general editors, told The Guardian: “We have been able to verify Marlowe’s presence in those three plays strongly and clearly enough. We can now be confident that they didn’t just influence each other, but they worked with each other. Rivals sometimes collaborate.”

Even so, perhaps because of write-ups like this, and despite coming out with Oxford University Press, a fairly prestigious publisher, New Oxford immediately met with some vigorous pushback. In the selfsame BBC article, we get a response from Carol Rutter, a professor of Shakespeare and performance studies at the University of Warwick:

It will still be open for people to make up their own minds. I don’t think [Oxford University Press] putting their brand mark on an attribution settles the issue for most people. […] I believe Shakespeare collaborated with all kinds of people… but I would be very surprised if Marlowe was one of them. The reason for that is that while these were being written, Marlowe was the poster boy of theatre writing. Why would he agree to collaborate with a non-entity of an actor? I would suggest we should look not to another playwright but to the actors. […] Yes, Shakespeare collaborated. But it’s much more likely that he started his career working for a company where he was already an actor, and collaborated not with another playwright but with the actors – who will have had Marlowe very much in their heads, on the stage, in their voices. They were the ones putting Marlowe’s influence into the plays.

In case you were wondering, at this stage in the miniseries, I am Team Gary Tayor, and I am Team Carol Rutter. What, you might ask, does that mean? How can you be both? Well, just as I absolutely champion the right of someone like Gary Taylor to ask questions, and come up with their own answers based, hopefully, on solid scientific principles, I also absolutely champion the right of anyone, like Carol Rutter, to put any set of methods or results or analyses under the microscope. And I say that no matter how esteemed the person who made the claim, and no matter how junior the one putting it to the test. I’ve seen good undergraduates debunk the work of distinguished professors. Good science is good science. It doesn’t care about your paygrade, the letters after your name, your seniority, provenance, country club friends, ego. Accordingly, in the rest of this miniseries, I’ve put all the other approaches through their paces, from Earls and Lords right through to spies and statesmen, pointing out the various issues with each, and I see absolutely no good reason why that same standard should not apply here. The New Oxford doesn’t get a pass just because it came out with OUP under the auspices of some long-standing figures in the field.

But so far, all I’ve given you are the New Oxford conclusions, and if the past five episodes have taught us absolutely anything, it’s that the methods, analyses, and interpretations of the results are everything. Woolly cipher-based ghost-chasing? You may find yourself digging up perfectly innocent riverbeds. Over-confident assertions based on incomplete evidence? You may end up emailing out some awkward concessions. Or worse. So what did the New Oxford Shakespeare say, and do, about authorship? Should we view their results as any more legitimate or meaningful than all the others we’ve seen so far?

Let’s take a look at their claims.

Modest doubt

Bear in mind in this next section that I’m going to summarise a book of almost 800 pages into a few short minutes, and then do the same with some rebuttals, and even squeak in some counter-rebuttals. The inevitable is going to happen. I am going to give an incomplete picture that simplifies some details and misses other things out entirely. Sadly that’s just the way data compression works. There is a solution of course. You can read the Authorship Companion yourself, in full… if you can afford it. As I write this, I checked and it’s currently retailing from OUP’s site at a cool £145. That’s almost exactly $200. And that’s just for the Authorship Companion. If you want all three volumes, you’re looking at £340. That’s a little under $500. Mercifully my library has the ebook, otherwise I wouldn’t even get to look at a copy, never mind butcher it in a podcast.

OUP’s astronomical overpricing of knowledge aside, Taylor and Egan don’t just look at language, though we’ll spend much more time on that shortly, for obvious reasons. They also discuss palaeographic and biographical evidence, and theatrical provenance. When it comes to language itself, they consider quotations, echoes, prose, rhetorical evidence, chronology, vocabulary, oaths, interjections, cue burden, stage directions, verbal parallels, image clusters, and a whole range of metrical tests including speech lengths, sentence/verse correlation, alexandrines, feminine endings and more.

We’ll start with the three Henry VI plays. The first play is credited as being written by Marlowe, Nash, and Anonymous, and then subsequently adapted by Shakespeare. So Shakespeare is not given as the author but as a later influence on the work. Presumably, though, Shakespeare then got a taste for this particular storyline because he’s credited as first author on the second and third plays in the Henry VI trilogy. Both of these are also supposedly co-authored by Marlowe, and anonymous, and then also later revised by Shakespeare. Convoluted. But let’s keep going. The justifications for ascribing authorship in this way are outlined in Egan’s chapter about the types of statistical tests that were run on the data. However, several scholars have questioned this chapter on various grounds. For instance, in his 2020 review of the Authorship Companion, Auerbach (2020: 236-241) writes the following, which I have very heavily edited for length:

There is too much that is good in the Companion to permit disregarding it, yet one concludes the volume with a strong sense of disappointment. A truly useful companion would have summarized the state of the field past and present, breaking down Shakespeare’s attributed and questioned works and the history of attributions made to them. It would have set out attribution techniques and shown their applications and failings. And it would have presented detailed and decisive cases for the New Oxford Shakespeare’s bolder attributions, such as assigning part of Arden of Faversham to Shakespeare and parts of the Henry VI plays to Marlowe. Such a work would be of tremendous use. Instead, what is presented is a partial (in several respects) grab bag, and worse, an unreliable one, where valid and invalid methods rub shoulders without the reader being easily able to distinguish between them, and a sense of priorities and proportion is lost in a sea of data. […] The volume, overall, has a shambolic feel. […] Egan sacrifices too much space to unhelpful analogies with forensic ear-print evidence and redundant computer storage. […] Also problematic are questionable priorities – the dubious and marginal cusum analysis of A Funeral Elegy is treated at greater length than the entire attribution history of Pericles – and statistical infelicities which I have discussed elsewhere (Auerbach, “Statistical Infelicities”). [I’ll come back to these.] We reach steadier ground with MacDonald P. Jackson’s short but thoughtful piece critiquing recent trigram studies. […] Jackson’s second chapter in the volume […] offers much to the expert, but little beyond confusion to the “readers new to the field” at whom the volume is aimed. […] Taylor and Duhaime’s study of the fly scene in Titus Andronicus […] demonstrates the pitfalls of quantitative analysis, however, as a flood of poorly summarized data frequently drowns out the subjective measures being employed. […] Moreover, in a subsequent article, […] Taylor leaves the reader uncertain of what standards are being applied and whether they are being applied consistently. […] Foot-shifting between quantitative number-crunching and qualitative hermeneutic interpretation, drawing attention to whichever supports the underlying contention of the moment while disparaging the other, is unfortunately far too frequent an occurrence in the Companion (see Auerbach, “Statistical Infelicities”). […] Anna Pruitt provides a sober corrective, follow[ing] a procedure that, ironically, a number of other contributors did not. Her article is clear and accessible. I have critiqued two of the case studies, those by Elliott and Greatley-Hirsch and by Pascucci, in detail elsewhere (Auerbach, “A Cannon’s Burst Discharged Against A Ruinated Wall”; Auerbach, “A Critique of Giuliano Pascucci’s”). The methodological problems in both are so great as to cast a pall over the remainder of the volume, and the endorsement of these chapters by the editors, who cite them as foundational support for their summary judgments, is quite troubling. Elliott and Greatley-Hirsch’s attribution of Arden of Faversham to Shakespeare is ultimately based on a handful of primitive factors far simpler than their elaborate and opaque machine learning-influenced procedures suggest. […] Pascucci makes a number of very specific and very audacious claims. […] Both chapters fall significantly below the standards for authorship studies, including those standards set forth in this very volume by Pruitt and Taylor. […] The volume fails to enforce a consistent standard of scientific rigor, and so the crucial foundation of such an anthology – the authority of its editorial judgment – is fatally jeopardized. There are consequences for the New Oxford Shakespeare more generally (Taylor et al.). […] More generally, a very real opportunity to provide a solid, comprehensive, and authoritative guide to the often arcane field of authorship studies was missed by this volume, and such opportunities do not occur frequently. To a neophyte, the volume is less accessible than purportedly more specialized studies by Jackson and Vickers, as well as less rigorous (Jackson; Vickers). Too often this Companion is partial, confusing, and worst of all unreliable. (Auerbach 2020: 236-241)

Okay so I mentioned Auerbach’s 2020 paper, Statistical Infelicities in The New Oxford Shakespeare Authorship Companion. Given that the Authorship Companion is heavily statistical in nature, it’s worth picking out just a couple of specific instances that can be relatively easily conveyed in audio form. At the start of this paper, Auerbach argues that the entire volume exhibits…

a lack of statistical rigour that vexes and ultimately dooms this effort, undermining any certainty that could be assigned to its resolutions. (Auerbach 2020: 28)

What is Auerbach talking about, specifically? Well, one example is thus. In the Authorship Companion, Egan states that,

Once some plays are entirely removed from the accepted canon of Shakespeare, the ranges within which various phenomena must fall in order to be typical of Shakespeare are likely to become narrower simply because we are generating them from a smaller sample. (Egan 2016: 30)

In response, Auerbach writes,

Here Egan seems to have it backward. There is no reason that an arbitrarily smaller sample of Shakespeare should generate more homogeneous results than a larger sample. Rather, the law of large numbers dictates that a smaller sample should show less regularity and more aberration, and simple tools of statistical analysis such as standard deviations operate on this fundamental principle. Extreme values are more likely to be tolerated as acceptable outliers in a smaller sample precisely because certainty is lower. More evidence increases certainty, rather than decreasing it. The only case in which adding a play to the canon would broaden rather than narrow a model’s tolerance of outliers would be if the play were severely and uniquely disjoint in its characteristics compared to all other plays in the canon. The most likely cause of such a disjoint nature, however, would be if a different author had written the play, and the play had been added to the canon in error. (Auerbach 2020: 28)

Or to put it more succinctly, more data, more certainty. Less data, less certainty. And in corpus linguistics there is a general preference for larger corpora over smaller ones for precisely this reason. To quote, well, Stalin, quantity has a quality all of its own. Sheer brute size smooths out weird anomalies that can easily skew smaller datasets. Of course with historical corpora the amount of available data is finite. Barring exceptional circumstances such as the surprising discovery of ancient manuscripts, there simply isn’t going to be any more of it, so we have to work with what exists, and choosing to downsample from what little there is to even less is a precarious step. To do so as a means of homogenising the data only works if you know that the ones you’re removing are, for some reason, unusual. But then you end up facing a whole host of other questions. What are the aberrations and why do they otherwise undermine your research question? What evidence led you to determine that your now-excluded pieces were aberrations at all? With more data – who knows, with plays and poems lost to time – perhaps this was actually a norm and now you’ve edited it out. If you’ve artificially homogenised your dataset then what do your results even mean any more since you’ve now essentially massaged them into showing you something want to see?

Anyway, what else? Well, in the Authorship Companion, Egan argues, fairly uncontroversially, that various statistical tests differ in their reliability, but that if we apply multiple independent tests to the same data that all point to the same author, then the likelihood that the answer is incorrect drops considerably. Essentially, statistical triangulation is the gold standard, and such a notion is generally widely accepted. Unsurprisingly, then, Auerbach (2019) agrees that this is correct in principle, but he argues that little effort is made across the entire book to actually establish whether the various statistical tests are genuinely independent. In fact, he argues that they are very likely to not be independent because most of the tests are based on one thing: word frequency. Thus, if the tests are not independent, then the fact that their results agree tells us relatively little about who the author is.

For the non-statistically minded, I’ll try to give a quick analogy. Imagine I hold up a picture of a Labrador and ask people to rate how cute it is. I start out by asking five people from the Labrador Appreciation Society. Remarkably, they all adore it. I then declare that five independent judges have approved of my picture, thus it is, objectively, an adorable dog. Well, not really. The five people are not independent judges. They all belong to the same society and are impelled by the same bias. By contrast, imagine the pandemic is over, I stand in a busy street somewhere, and I ask passing strangers on a busy street to rate the same photo. There’s a good chance that at least some of these strangers will be independent from each other. I might accidentally ask a pair of friends or both members of a couple or a parent and child, sure, but even if that happens, should all five random people now agree that my photo is adorable, then yes, I probably do have a very nice picture of a dog. We can be a bit more certain that this result is a better reflection of a wider opinion because there’s a far better chance that at least some of these people will be uninfluenced by each other or by an identical background factor.

So back to Egan and Auerbach. Auerbach is essentially suggesting that the Authorship Companion might be an extended work of showing the Labrador picture to multiple members of the Labrador Appreciation Society – that is a bunch of people who are not independent judges, Moreover, Auerbach is suggesting that we have good reason to believe that they really are all be connected in some way because they’re all wearing Labrador Appreciation Society badges. Or, in the case of the data, the tests are mostly based on word frequency. (Statisticians, please don’t @ me. I’m trying my best here.)

Aaaanyway, in his Statistical Infelicities paper, Auerbach (2020) also takes issue with what appears to be cherry-picking – the decisions to include statistical tests when they agree and discount them when they don’t; the acceptance of some parts of prior studies that confirm the Authorship Companion’s results and rejection of other parts of those selfsame studies where they contradict, and so on. In summary,

Just as the results of these studies must be taken in total or not at all, rigor cannot be applied selectively. The presence of fundamental statistical errors in significant passages in the Companion has the effect not of casting doubt merely on those particular passages and points, but of bringing into question the robustness of all such conclusions reached. Here is a scientific sloppiness that pervades the entire Companion and so disadvantages its overall tenability. (Auerbach 2020: 30)

In summary, then, Auerbach is not really impressed with this volume. But does anyone else share these concerns? Is this just a one-person axe-grinding campaign or is there consensus? Well, another critic, Erne (2018) argues that:

What is striking about these case studies is the confidence with which the conclusions are presented, even when the sample is small (the Fly Scene is 84-lines long) or when the extant play (Double Falsehood) is argued to be an eighteenth-century rewrite of a late seventeenth-century rewrite of the play to which Shakespeare contributed (Cardenio). Even if we grant that significant advances have been made in authorship attribution studies (and this reviewer certainly does), such confidence may nonetheless seem surprising. At the end of his chapter on the authorship of two poems – ‘When God was pleas’d’ and ‘Shall I die’ – that were attributed to Shakespeare in manuscript, Gary Taylor writes: ‘Until we have better tools, readers must simply make up their own minds. Or perhaps, preferably, refuse to make up their minds’ (230). Yet elsewhere in the volume, there is little such commendable reluctance to jump to firm conclusions, including by Taylor himself. (Erne 2018: 3)

Erne seems to suggest that Taylor is preaching, but not practicing, and what I infer in the softest whisper from this is that perhaps the lure of the Shakespearian holy grail – attributing authorship in a brand new way that will crown Taylor’s career with immortal success – has gotten in the way of rigor and objectivity. Anyone else have issues with this work? Well, Freebury-Jones has written a few things on it, but to take just a little snippet:

The New Oxford Shakespeare authorship team argues that Christopher Marlowe had a hand in all three Henry VI plays, and that Shakespeare wrote parts of Arden of Faversham. However, the quantitative methodologies employed to reach these conclusions are dubious. (Freebury-Jones 2018: abstract because paywall…)

Anything else? Well, last of all in this gigantic saga, there’s the Gary Taylor versus Brian Vickers… thing.

Just think Montagues and Capulets all over again… <sigh>

Honestly, I’m… yeah okay well how about I just quote some of Vickers and then some of Taylor and you can form your own conclusions. In the Times Literary Supplement in 2020, in a 3,500-word article about The New Oxford Shakespeare, Brian Vickers writes the following:

The prime mover in all three revisions to Shakespeare’s canon was Gary Taylor, a young American scholar who had just been promoted from editorial assistant to co-editor alongside Stanley Wells, a respected Shakespeare scholar. [Interesting implicature there. Anyway, moving on.] Since no external evidence exists for these ascriptions, Taylor had to rely on other evidence or his own aesthetic judgement. The small rocky island on the new map was a banal lyric, “Shall I die?” which Taylor inserted in the canon because in one manuscript the scribe signed it “William Shakespeare”, increasing its value to other collectors. In fact, it looks like a poem written for music in around 1610. None of Taylor’s co-editors, and indeed no other critic, ever endorsed the ascription, but if you’re editor of the Oxford Shakespeare you don’t need anyone else’s approval. […] In the press releases and interviews that Taylor gave when the new edition was launched, he described his efforts in triumphalist terms. […] The New Oxford Shakespeare contained no justification of the attributions; this appeared a year later in an Authorship Companion, edited by Gary Taylor and Gabriel Egan. This bulky volume […] seems to be a professional scholarly production, [but evaluating it properly] needs considerable expertise in mathematics, statistics and computing as well as Elizabethan drama and textual scholarship. […] This Authorship Companion is unfortunate proof that scholars, journal editors and publishers in the Humanities are prone to being abused by pseudo-scientific methods. […] Every attribution is false. Oxford University Press has a proud record as the world’s leading publisher of scholarly editions of English literature. The trust that senior editors placed in Gary Taylor has been repaid with an opportunistic bundle of untested methods set loose on the greatest author in our language. Shakespeare is not just a national, but an international treasure [heh, see what I mean?] and it is tragic to contemplate the damage done to culture in general by these editions being used to teach students, and being sold in bookshops to unsuspecting laymen. The Press has just commissioned the New Oxford Marlowe. Among its editors are members of Taylor’s editorial team, and rumour suggests that it will include the Henry VI plays. Many people will fervently hope that on reflection the editors will think it enough to have ruined one major author’s canon. (Vickers 2020)

You might be forgiven for thinking that some of this actually seems kind of bitter and personal, and you will not be surprised to learn that it really does seem to be yet another Shakespeare-driven grudge match between two big names in the field. Two years earlier, in 2018, Vickers advertised in the very same Times Literary Supplement for a publisher for his latest book, and when asked why he needed to go to such a length after the success of publishing forty previous titles over the span of fifty years, Vickers replied that his reputation as a scholar had been damaged by a string of hostile reviews by people associated with the New Oxford Shakespeare. According to Vickers, he had been turned down by Cambridge University Press, Manchester University Press, and Bloomsbury because their referees deemed him controversial (Lea 2018). Taylor’s response to this?

It is ironic that Brian, who has been writing savage ad hominem reviews in the TLS for 30 years, is blaming his failure to find a publisher on other people’s reviewing. […] I doubt that his conspiracy theories will convince anyone who isn’t already paranoid. […] Brian’s approach to Shakespeare is that there is only one proper way to interpret him – Brian’s. The New Oxford Shakespeare, by contrast, is a collaborative edition, and its critical introductions give readers many possible approaches: 1950s approaches such as Brian’s – who is quoted on a number of occasions – but also theatrical, historical, political, formalist, feminist, cinematic, psychological, and eco-critical interpretations. For Brian, Shakespeare is fuel for angry, narcissistic monologues. For us, Shakespeare inspires thousands of fascinating conversations.

Interestingly, Vickers also seems to imply a sort of narcissism about Taylor when he claims that the authorship question is a matter of scholarly judgment, yes,

but also of headline-seeking, power and prestige. My position is that, as a great artist, Shakespeare should not be parcelled off to other writers or saddled with plays that he did not write, such as Arden of Faversham – especially not by such shoddy and bogus scholarly methods.

Hmmm. Overall, then, whilst The New Oxford Shakespeare’s Authorship Companion certainly adds to the conversation on the Shakespearian authorship question, it seems to throw up far more questions than it answers. At the very least, as I noted at the start of this episode, it has successfully generated plenty of controversy and engagement, and in its own right that new interest can be useful, if channelled properly into analytical and methodological advancements.

Unfortunately, however, though the Elizabethan dramatist died five hundred years ago, he has still managed to create an awful lot of drama over the ensuing centuries, right through to the present day, and the passions and fervour that his work inspires have arguably held the authorship question back at least as much as they have propelled it forward. For many people in the field, both Stratfordians and Antis, the Shakespeare authorship question in whatever guise seems to have been a mechanism not so much for advancing science and society and civilisation, but rather, for advancing CVs, and status, and success.

Finally, then, for the very last time in this miniseries, we return to the question that we started with: Who wrote the works of Shakespeare? And what is our answer? Well, we have to acknowledge that as it stands, there is still, as yet, no one answer that satisfies everyone, and to even try to propose one here would be to come between the dragon and his wrath.

End of part 6 of 6.

If you’re interested in more Shakespeare content, from linguists, at Lancaster, then search the internet for Future Learn, Shakespeare’s Language. This free online course is all about both revealing meanings and exploring myths, and as a bonus, you get introduced to corpus-based methods for analysing Shakespeare’s language. What’s not to love!

Outro

The episode was researched and fact-checked by my research assistant, Rebecca Jagodzinski, and my intern, Debbi Tomkinson, and it was narrated and produced by me, Dr Claire Hardaker. I am also extremely grateful for all the input I’ve had from the renowned Shakespeare authority, Jonathan Culpeper – creator of that online course I mentioned – who has patiently entertained this whole miniseries idea from inception to gruesome, bloody execution.

However, this work wouldn’t exist in its current form without the prior efforts of many others. You can find acknowledgements and references for those people at the blog. Also there you can find data, links, articles, pictures, older cases, and more besides.

The address for the blog is wp.lancs.ac.uk/enclair. And you can follow the podcast on Twitter at _enclair. Or if you like, you can follow me on Twitter at DrClaireH.

References

This is an incredibly incomplete list of potential introductory sources (many deliberately introductory) that you could turn to if you were interested in reading more about this subject, but as is the nature of link-rot, some of these sources will have disappeared or changed in the intervening time. Suffice to say, there are way more texts out there on this subject than you could probably ever want to read, but this should get you started…

Baconian theory of Shakespeare authorship. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Baconian_theory_of_Shakespeare_authorship [Accessed 11th June 2020]

Brewer, C. (2013). Shakespeare, word-coining, and the OED. Shakespeare Survey, 65, 345-357.

Christopher Marlowe. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Christopher_Marlowe [Accessed 13th June 2020]

Craig, H. (2011). Shakespeare’s Vocabulary: Myth and Reality. Shakespeare Quarterly, 62(1), 53-74.

Derbyite theory of Shakespeare authorship. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Derbyite_theory_of_Shakespeare_authorship [Accessed 11th June 2020]

Edward de Vere, 17th Earl of Oxford. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Edward_de_Vere,_17th_Earl_of_Oxford [Accessed 13th June 2020]

Elliot, W., & Valenza, R. (1991). Was the Earl of Oxford the True Shakespeare? A Computer-Aided Analysis. Notes and Queries, 38(4), 501-506.

Emilia Lanier. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Emilia_Lanier [Accessed 15th June 2020]

Francis Bacon. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Francis_Bacon [Accessed 13th June 2020]

Friberg, J. (2016). The Seven Steps to Mercy: Cracking the Shakespeare Code [Documentary Series]. Syndicado.

Friedlander, A. (2015). Five myths about William Shakespeare. Retrieved from: https://www.washingtonpost.com/opinions/five-myths-about-william-shakespeare/2015/09/04/bc986ef6-524d-11e5-9812-92d5948a40f8_story.html [Accessed 9th June 2020]

Henry Neville (died 1615). (n.d.). Retrieved from https://en.wikipedia.org/wiki/Henry_Neville_(died_1615) [Accessed 15th June 2020]

Holloway, C. (n.d.). Shakespeare’s Stratford Monument. Retrieved from: https://www.hollowaypages.com/Shakespearemonument.htm [Accessed 18th June 2020]

Kathman, D. (n.d.). Part 5 of “Critically Examining Oxfordian Claims”. Retrieved from: https://shakespeareauthorship.com/ox5.html [Accessed 18th June 2020]

Maranzani, B. (2020 [originally 2019]). Was Shakespeare the Real Author of His Plays? Retrieved from https://www.biography.com/news/shakespeare-real-author-theories [Accessed 9th June 2020]

Marlovian theory of Shakespeare authorship. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Marlovian_theory_of_Shakespeare_authorship [Accessed 11th June 2020]

Mary Sidney. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Mary_Sidney [Accessed 15th June 2020]

Meares, H. (2020 [Original 2019]). Why Is William Shakespeare’s Life Considered a Mystery? Retrieved from https://www.biography.com/news/william-shakespeare-life-mystery?li_source=LI&li_medium=bio-mid-article&li_pl=208&li_tr=bio-mid-article [Accessed 9th June 2020]

McCrum, R. (2010). Who really wrote Shakespeare? Retrieved from https://www.theguardian.com/culture/2010/mar/14/who-wrote-shakespeare-james-shapiro [Accessed 9th June 2020]

Nevillean theory of Shakespeare authorship. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Nevillean_theory_of_Shakespeare_authorship [Accessed 11th June 2020]

Oxfordian theory of Shakespeare authorship. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Oxfordian_theory_of_Shakespeare_authorship [Accessed 11th June 2020]

Peer, M., & Jacobi, D. (2004). The Shakespeare Conspiracy [Film]. TMW Media.

Rubinstein, W. (2001, 08). Who was Shakespeare? History Today, 51, 28-35.

Schmitz, E., & Utzt, S. (2011). The Shakespeare Enigma [Film]. Atlantis Film.

Shakespeare authorship question. (n.d.). Retrieved from https://en.wikipedia.org/wiki/Shakespeare_authorship_question [Accessed 8th June 2020]

Syme, H. (2011). Shakespearean Mythbusting I: The Fantasy of the Unsurpassed Vocabulary. Retrieved from http://www.dispositio.net/archives/501#:~:text=If%20one%20only%20counts%20the,was%20indeed%20an%20impressive%20number [Accessed 9th June 2020]

Wiggins, M. (n.d.). Who Wrote Shakespeare? Retrieved from https://www.bbc.co.uk/programmes/articles/211LBPTmBYp2rbh4bSQlSTS/who-wrote-shakespeare [Accessed 9th June 2020]

William Shakespeare. (n.d.). Retrieved from https://en.wikipedia.org/wiki/William_Shakespeare [Accessed 8th June 2020]

William Shakespeare. (n.d.). Retrieved from https://www.oed.com/view/source/a644?rskey=MpP0Ls&result=2 [Accessed 19th June 2020]

William Stanley, 6th Early of Derby. (n.d.). Retrieved from https://en.wikipedia.org/wiki/William_Stanley,_6th_Earl_of_Derby [Accessed 15th June 2020]

Other relevant texts

The Annotations contained within Oxford’s Bible can be found here: https://shakespeareauthorship.com/oxbib.html

William Webbe’s Discourse of English Poetrie is accessible here: http://eebo.chadwyck.com/search/full_rec?SOURCE=pgimages.cfg&ACTION=ByID&ID=99846918&FILE=&SEARCHSCREEN=param(SEARCHSCREEN)&VID=11915&PAGENO=2&ZOOM=FIT&VIEWPORT=&SEARCHCONFIG=param(SEARCHCONFIG)&DISPLAY=param(DISPLAY)&HIGHLIGHT_KEYWORD=param(HIGHLIGHT_KEYWORD)

The Art of English Poesie can be found here: http://www.gutenberg.org/cache/epub/16420/pg16420-images.html

Parallels between Promus and Shakespeare’s works are accessible at these two links: http://www.sirbacon.org/mp.html and http://www.sirbacon.org/mp2.html

Palladis Tamia can be found here: http://eebo.chadwyck.com.ezproxy.lancs.ac.uk/search/full_rec?SOURCE=config.cfg&ACTION=ByID&ID=99845635 and https://shakespearedocumented.folger.edu/exhibition/document/palladis-tamia-one-earliest-printed-assessments-shakespeares-works-and-first

A Review of James and Rubinstein’s work can be found here: https://academic.oup.com/sq/article/58/2/245/5064487

William Barksted’s poem is found here: https://shakespeareauthorship.com/barksted.html

The Dedication poem at the beginning of the sonnets can be found here: http://www.shakespeares-sonnets.com/dedication

The Compleat Gentleman can be found here: https://archive.org/details/peachamscomplea00peacgoog

The First Folio can be accessed here: https://www.folger.edu/the-shakespeare-first-folio-folger-copy-no-68#page/To+the+Reader/mode/2up

Shakespeare’s Epitaph can be read here: https://www.williamshakespeare.net/william-shakespeare-epitaph.jsp

Rev. Dr. John Ward’s diary is accessible here: https://archive.org/details/diaryrevjohnwar00sevegoog/page/n16/mode/2up

One of the 1819/1820 articles on Marlowe can be found here: https://books.google.co.uk/books?id=RQcwAAAAYAAJ&pg=PA61&redir_esc=y#v=onepage&q&f=false

Information on the Northumberland Manuscript can be found here: http://www.sirbacon.org/NMANUSCR.HTM

Ignatius L. Donnelly’s proposal can be accessed here: https://archive.org/details/greatcryptogramf00donnrich/page/n7/mode/2up

Wilber G. Zeigler’s proposal can be accessed here: http://www.gutenberg.org/ebooks/57810

J Thomas Looney’s proposal can be accessed here: https://archive.org/details/bub_gb_B004AAAAIAAJ

One of Alfred Hart’s Studies is accessible here: https://academic-oup-com.ezproxy.lancs.ac.uk/res/article/os-X/39/274/1621907

Stritmatter and Anderson’s analysis of the Bible is accessible here: https://shakespeareoxfordfellowship.org/shakespeares-bible/

Dudley, M.Q. (2013). “‘My Library Was Dukedom Large Enough’: Academic Libraries Mediating the Shakespeare Authorship Debate”, The Canadian Journal of Library and Information Practice and Research, 8(2), 1-9.

Franssen, P. (2013). “Fictional treatments of Shakespeare’s authorship”. In: P. Edmonson and S. Wells (Eds), Shakespeare Beyond Doubt: Evidence, Argument, Controversy (pp. 189-200). Cambridge: Cambridge University Press.

Holderness, G. (2013). “The unreadable Delia Bacon”. In: P. Edmonson and S. Wells (Eds), Shakespeare Beyond Doubt: Evidence, Argument, Controversy (pp. 5-15). Cambridge: Cambridge University Press.

Martin, Patrick & Finnis, John. The Secret Sharers: “Anthony Rivers” and the Appellant Controversy, 1601–2. Huntington Library Quarterly, Vol. 69, No. 2, pp. 195-238

McCrea, S. (2005). The Case for Shakespeare: The End of the Authorship Question. Westport, CT: Praeger Publishers.

Schiff, J. (2015). “A genius, but mad”, Yale Alumni Magazine, Nov/Dec. Last accessed 18^th September 2019. Available at: https://yalealumnimagazine.com/articles/4189-delia-bacon

Stewart, A. (2013). “The Case for Bacon”. In: P. Edmonson and S. Wells (Eds), Shakespeare Beyond Doubt: Evidence, Argument, Controversy (pp. 16-28). Cambridge: Cambridge University Press.

Taylor, G. and Bourus, T. (2016a). “Why Read Shakespeare’s Complete Works?” In: G. Taylor, J. Jowett, T. Bourus, G. Egan (Eds), The New Oxford Shakespeare: William Shakespeare, The Complete Works (pp. 1-44). Oxford: Oxford University Press.

Taylor, G. and Bourus, T. (2016b). “Why Read This Complete Works?” In: G. Taylor, J. Jowett, T. Bourus, G. Egan (Eds), The New Oxford Shakespeare: William Shakespeare, The Complete Works (pp. 45-58). Oxford: Oxford University Press.