The Shakespeare Conference: SHK 22.0075  Monday, 16 May 2011

From:         Gerald E. Downs <This email address is being protected from spambots. You need JavaScript enabled to view it.>
Date:         Wednesday, 23 Feb 2011 18:49:27 -0500 (EST)       
Subject:     Groats-worth of Wit: Greene or Chettle?

Tom Reedy called attention to an article that I had not seen:
>Why is the attribution to Henry Chettle of Greene's Groatsworth of Witt
>so universally accepted among Shakespeare scholars? John Jowett,
>Katherine Duncan-Jones, Brian Vickers, and many others wholeheartedly
>endorse Warren B. Austin's 1969 computer-based study, despite the
>many flaws in its methodology pointed out by Richard Westley in his
>"Computing Error: Reassessing Austin's Study of Groatsworth of Wit,"
>_Literary and Linguistic Computing_, 21:3 (2006), pp. 363-78.

>As far as I have been able to determine, Westley's paper has never been
>rebutted or even reviewed, and I find his arguments persuasive.

Having now read the article I venture a few(?) comments. I have long accepted the likelihood that Chettle was behind Groatsworth, in part because of Austin's study. Although a number of prominent scholars agree, there is no reason in principle for the inquiry to end. Yet Westley's attempt to reopen discussion is faulty.

First, his criticism of Austin is needlessly ungentlemanly and accusatorial. Those who are impressed with Austin's collegiality and argumentation will not welcome personal attack as if it were argument. Here are a few examples of tone, if not substance:

"Austin deftly avoids . . ."; "One compelling reason to exclude _Messenger_ from study is that it does not support the hypothesis Austin wants to advance" (365); "Austin's bias emerges . . ."; "pseudo-mathematical. . ."; "Austin later admits . . ." (366); "Austin applies a perverse and unnecessary standard . . ." (368); "a kind of scholarly gerrymandering . . ."; "And he fudges . . ."; "And fantastically . . ."; "Austin's glibness . . ."(369); "an irreparable research gaffe"; ", , , for no good reason"; "The charge of derelict practice . . ."; "a pretext . . ." (371); "manipulations of the data . . ." (372); "Consequently, it is disingenuous . . ." (376). And so on.

It surprises me that name-calling was allowed by the journal, though of course an author's "style" is his own, and though some may be influenced by what others take to reflect a weak argument (else why resort to it?). Eminent scholars will simply ignore a whole production rather than condone such goings-off, or become the next targets of abuse. One of the problems facing Internet discussion is its ability to draw the acknowledged authorities out of solitary; criticism is essential, but there are right and wrong ways. Also, a faulty argument accompanying name-calling increases reluctance to reply. Many who approve of the case for Chettle's "main finger" (though they usually claim only a probability and call for further study) will spot Westley's missteps.

Perhaps all agree that if Chettle wrote Groatsworth (to shorten the proposition), we would like to know. Yet, far from universal acceptance of the hypothesis, many scholars don't even mention it when they should, or seem unaware. Personally, and more as time passes, I find Henry Chettle to be a central figure in the history of the literary age. I think well especially of the work and insight John Jowett brings to Chettle-based topics.

I take Westley's observations in no particular order. (An argument I lately find interesting is discussed farther along). To begin, he seems largely to miss the purpose of Austin's study. For convenience, Austin limited a control text, Greene's _Farewell to Folly_, to its first 25,000 words. Westley finds something sinister in that, "thereby creating interest in what the omitted part might reveal" (369). It reveals a "Folly" clause, "to whom hee had so many times beene beholding", paralleling two clauses in the Groatsworth letter to three scholars. Westley observes that "such stylistic congruence is a staple of Austin's study except that he only highlights those cases that connect [GW] to Chettle's corpus" (370).

But GW is full of parallels to Greene, as everyone is aware; to suppose them necessary to the study is to misunderstand the hypothesis. In the first place, what is surprising about GW that leads to the hypothesis? Nothing, actually; it seems and claims to be Greene's work. The surprise is in the preface to Chettle's Kind-Heart's Dream (KHD), where we learn from Chettle himself that he is suspected of forgery. Nashe calls GW "a scald trivial lying pamphlet," but Chettle supplies his own indictment. This might be accounted for by the obvious possibility that Chettle forged GW. That is the hypothesis, backed by a number of circumstances. It is not tested by parallels, the putative forger's tools, but by the habits of the forger, as differentiated from the usage of Greene that does not include, as far as possible, his distinctive verbal characteristics.

That may seem unfair, but the investigator has to guard against parallels and unusual usage because of what I call "fallacy of the odd occurrence," or the "forger's choice." The perp is free to choose coincidental material that may later seem meaningful to attribution studies. If Greene "yarkt up" a book in a day-and-a-half, Chettle might do the same in a couple of weeks with a supply of Greene publications, including 'Folly.' Study should be designed to bypass likely tools, which Austin does; honestly, it seems. Westley continually recommends the mistake of testing GW for Greene (while he refers to the suspect Chettle as an "eyewitness." Unless we clep burglars eyewitness, which they are in a sense, the term presumes innocence a bit much).

Westley cites the deft derelict Austin himself to the effect that "Elizabethan compositors were 'faithful to the author's copy'" (375), to prepare an argument that spelling differences in GW and KHD corroborate Chettle's claim that he had only copied GW from Greene's holograph (The explanation, apparently, for both a missing Greene ms. and printer's copy in Chettle's hand). Forgery needs no such explication, though a forger might. Westley finds these differences: leisure / leysure; aire / ayre; Maistership / Mastershippe & Maship; mistris / mistrisse; musike / musicke; otherwise / otherwayes; sorie / sory; and verie / verye. In all, he lists 37 (or so) pairs or multiples of different spellings. They bespeak Westley's unfamiliarity with the reality of compositorial practices, which allow little opportunity for inference in these works. Rather than argue this point I'll similarly list words from Greene's _Planetomachia_ (1585), alongside the same words in _Penelope's Web_ (1587), from a passage of a little over 250 words transferred from the early work into the later. Greene copied almost word-for-word; what do we learn about his spelling? (Taken from Goree, Phil. Q 3, 1924, u/v modernized).

1) marveile / marvaile 2) Egipte / Aegipt 3) woont / wont 4)  councell / counsell 5) Parleamente / Parliament 6) hee/he 7) leaste / least 8) counsayle / counsaile 9) fall / falles 10) cuppe / cup 11) lippe/lip 12) hindrance/ hinderance 13) sodayne / sodaine 14) onely / only 15) hear / heare 16) denial / denyall 17) yeeld / yeelde 18) dispight / despight 19) menne / men 20) wee / we 21) shee shall / she shal 22) al / all 23) marry / marrie 24) persuade mee / perswade me.

Goree quotes other similarly large Greene "borrowings" from himself, e.g. 300 words from _Planetomachia_ to _Perimedes_ (1588). The first of the lines run:

. . . hardly admitting any into familiaritie unlesse he might sell his courtesie for profite, and they buy his favour with repentaunce.

. . .  he admitted none into familiarite, unless he might sell his courtesie for profit, and they buie his favor with repentance.

One-third of the spellings vary; the same words from the same author. It could go without saying that authors alter their spelling and that even the same compositors spell variantly, sometimes for reasons (i, e, y, ie, and 'Mastershippe & Maship' come and go to justify line-to-margin); and sometimes for no reason. Only evidence out of the ordinary might point to Chettle or Greene. Westley's short list for GW & KHD (and a cursory look of my own) indicate unusually invariant spelling, which one might expect from two manuscripts in the hand of Chettle; who as a professional closely associated with the printers probably directed the composition. None of the spelling exonerates Chettle from forgery. And it may be that Chettle's and GW's 'O' and Greene's 'Oh' preferences are not meaningful, as Westley suggests (he makes much of orthography otherwayes). Yet the usage ratios are perhaps large enough to observe Greene's borrowing again. From R W Dent, "_Gwydonius_: A Study . . . in Plagiarism":

Greene (_Gwydonius_ 1584, one of Westley's alternatives not included in Austin's study):

. . . O lawlesse Love, O witless will, O fancie, fraught full of phrensie and furie.

George Pettie's _Petite Pallace_:

. . . O love without law, O rage without reason. O will without wit, O fansy fraught full of fury and frensy.

A complete count of O's and Oh's must of course reduce Greene's O numbers by these three, atop a big can of Greene worms. In 1906 H C Hart (N&Q) noted (before any need for 'fudge charges') "one interesting result in this process of appropriation: Greene's text and glossary become possessed of many terms to which he has no claim whatever." He has no better claim to the spelling. Hart reports as many as ten plagiary pages at a clap, plenty of which Greene uses again in later yarkings-up. For example, Thomas Bowes's translation of La Primaudaye (1586), 138-148 "is bodily lifted into Greene's . . . _Second Part of Tritameron_ (1587)."

Closer to home, lots of Primaudaye is copied into "Folly," most of which Austin used as a control text; about which Westley faults Austin for not using it all. In another instance, Greene uses Primaudaye in "Perimedes" and he "copies this whole passage into 'The Royal Exchange' . . . ." Before any early work of Greene's could be properly used in a comparative study, his "carloads" of plagiarism, his direct translations, his repetitions aplenty from his own books (stolne or not), would have to be excised. They're either not his words nor style, or they double or triple the data. In this respect Austin is probably right to have limited his study to later text, though "Folly" shows itself a poor control. The extent of plagiarism is unknown, but massive.

Keeping Chettle's works in the forefront of the study helps matters. The accurate way to handle the stylistic mess is to ask whether GW agrees more with Chettle than it does with "Greene and others." After all, Henry is the suspect and the data for Greene will never be accurate. Expanding it to the Greene canon will not help and since any such result would be skewed, improvement on Austin's choice is unlikely. That ties in with the Westley criticisms, which seldom actually dispute Austin's results. An imperfect study that points to Chettle still points to Chettle, but Austin's work easily withstands the confused attack.

Westley seems to think that all possible evidence had to be evaluated by Austin, rather than to allow the researcher to formulate a limited number of tests determined by his own rationale. For example, Austin was led "to exclude much of the data needed to determine the authorship . . ." by the criteria of a specific test, in this case "Greene- or Chettle-plus words" which [as I understand the quotation] occurred in one or the other corpus at least ten times, where one author's usage was at least one-and-one-half times as frequent, but that it's variation within the works of the high-use author had to remain less than between the two authors. Now obviously these restrictions are meant to pare down the list of tested words. Westley says the words "are anything but randomly chosen" (368), but the criteria have nothing to do with the choice of candidate words, from which the list derives.

The last criterion is that a test word's "range of use in the works of one author had to be clearly distinguished from, and not overlap, its range in the works of the other." Westley notes that it "is not clear what the fourth criterion is saying nor why Greene and Chettle had to differ in the grammatical forms of certain words they used, assuming that is what Austin means." If one does not understand a sentence, why assume a meaning only to criticize the assumption? The rule is not about grammar (other than its own, of course). What Austin means is that the average of the numbers of each author's use of the word must reflect the use in each work. Anomalous use in a work or two (of either author) removes a word from those measured against Groatsworth. This appears to be a fairly designed test. Westley doesn't report results, doesn't disagree with them, doesn't say how the test is rigged against Greene, or why the results should be distrusted. That is his way.

Westley's alternative insistence is that rare words should be the evidence. As I noted, that is the mistaken way to judge possible literary forgery, which can draw words out of a hatful of Greene and others. Westley remarks that "because such words are unique or rare, they are statistically decisive when they occur in discernable patterns." But how do unique words form patterns?

"Austin follows this up with still more heavy-handed exclusions: all auxiliary verbs, inflected verb forms, and even personal pronouns are omitted from his study, the latter because they are 'dependent on the relative prominence of male and female characters, the author's choice of point of view, and the relative amount of dialogue' in a given work . . . . This is tantamount to carving out the particular words Austin does not want studied. [And? Seems intelligent carving to me.] . . . He studies only those exceptions that favor Chettle as the author of _Groatsworth_" (369). Westley is accusing Austin of rigging results by designing tests around evidence that had already been found to favor Chettle. Isn't it possible that Austin (an eyewitness, amongst others, no doubt) devised the tests before they were run, and that they favor Chettle's authorship because he was the author? Again, Westley does not dispute the results.

"Of course, scholars often must use their materials selectively in order to detect and study the right lexical patterns. Still, Austin's persistence in overlooking vital data constitutes an irreparable research gaffe" (370). So why is Austin criminally selective? Why is non-use of data irreparable? The data hasn't gone anywhere. Where is the argument against Austin's tests and results? Alternative tests are only meaningful in the doing.

Examples of Westley's loaded terminology are bad enough in themselves; where do they ultimately lead? "And fantastically, Austin reverses himself to make a single exception to his exclusionary rules: 'Thus *you* and *ye* were eliminated with the class of personal pronouns; but each writer's preference for *ye* and *you*, the alternatives being open to him, was noted and proved significant.' . . . Consider Austin's glibness. [These words] are excluded . . . but then are reinstated because they represent each author's preferences . . . ."

Yet there is nothing wrong with testing the perceived difference. Even Westley has "no objection to including *you* and *ye* as data helpful to determining the authorship of [GW] . . . (369)." Then why fantastic and glib? Though Austin's counts seem meaningful, showing Greene seldom uses *ye*, whereas Chettle and GW have a greater *ye* percentage (38% and 19%, respectively). Westley returns to the issue with results of his own, derived from the six LION Greene prose texts. He reports a 20% *ye* use, in agreement with GW; his figures are *ye* 218 times, *you* 874 times -- far from what would be predicted by Austin (372, 377n.).

Observant readers might notice that Austin utilizes the text of _A Quip for an Upstart Courtier_; and so does Westley. "Quip" uses more than 200 *you* and only 4 *ye* (my counts are subject to correction, but close). That means the *ye* percentage must be more like 25% in the rest of the LION texts. But "Pandosto" uses 'thee' and 'thou' predominately, where you/ye is only 25/1. The later "Menaphon" is 120/8, so we are down to about 450 *you* and 210 *ye* to account for in three texts. "Morando" shows 90/3. What is going on?

I was unable to find satisfactory texts of Mamillia and Morando on the Net; the other texts are available in modernized form, which preserves the you/ye usage well enough. But I noticed in C J Vincent's "Pettie and Greene" (MLN, '39) Mamillia spellings reproduced as "ye fairest" and "ye wolfe"; it occurred to me that "old spelling" texts left to dumb-bundly computers would render the "thorn and superscript e" definite article as "ye" when an uncontrolled Control-F would up ye pronoun count. As it happens, the LION texts do just that; Mamillia has 180 *you* and 94 *ye*, none pronominal. Morando has 91 you, 3 ye pronoun, and 11 "ye the." Double-checking Gwydonius, where I had counted zero "ye's," LION shows 162/88 you/ye; no pronoun ye.

Westley's count of 218 *ye* in the six LION prose texts can only be accounted for by his failure to separate ye from ye. One may generously infer that he did not read a single "ye" clause in all these works and that he has no knowledge of "ye" as an abbreviation of "the." The corrected results confirm Austin's test as favoring Chettle's usage in GW by a wide and consistent margin. I am not especially concerned to conclude that the data pool is over Westley's head; more importantly, personal attack shouldn't be rewarded by easy approval of his hugely mistaken counter-argument.

Westley's introductory remarks reflect a poor understanding of the issues surrounding GW: "And so the justly famous attack on the playwrights that identified Shakespeare as a _Johannes fac totum_ (jack of all trades) in the midst of a roiling London theatre scene in 1592 may be a fiction, or else does not refer to Shakespeare at all (Erne, 1998; see Carroll, 1994a)." But Erne accepts that Shakespeare was the person alluded to in GW, as does Carroll. The "attack" was primarily on the player Shakespeare and Erne reargues the case that Chettle's apology in KHD was not to Shakespeare, the player attacked, but to George Peele, one of the three playwrights specifically addressed in the GW letter. The apology is a different issue. Carroll, in an article (1994a) separate from his edition of GW, discusses the identity of the player in "Roberto's Tale," not that of the player attacked in the letter to the playwrights. There's no way these immensely human topics should be neglected to pursue an ill-informed inhuman bean count.

Much more can be said of Westley's essay, only to continue on an unproductive tangent. But the topic itself is not played out. From the work of Chauncey Sanders, Austin, Carroll, and Jowett, a substantial hypothesis developed, that Chettle is most responsible for GW. The alternative to Greene's authorship should get more attention.

Gerald E. Downs
SHAKSPER: The Global Shakespeare Discussion List
Hardy M. Cook, This email address is being protected from spambots. You need JavaScript enabled to view it.
The SHAKSPER Web Site <http://www.shaksper.net>

DISCLAIMER: Although SHAKSPER is a moderated discussion list, the opinions expressed on it are the sole property of the poster, and the editor assumes no responsibility for them.

Subscribe to Our Feeds


Make a Donation

Consider making a donation to support SHAKSPER.