For some time now I have suspected that Anthony Munday and Edward de Vere might have written the play. The maddening aspect of presenting evidence in Munday's behalf, however, is that his writing style is free of tells. It is more modern than other playwrights of the time, true: only moderately Euphuistic; competent with prosody but never inspired; primarily a prose writer — a freelance writer in a strangely modern sense.
Anyway, taking a break, recently, from yet another session tallying endless minutiae, it came to me to ask the Microsoft Copilot AI Mode search feature what was the most common adjective in the plays of Kyd. I had already compiled the clear answer. It dawned on me, however, that it would possibly teach me more than a little about AI to ask it this question.
The first thing I seem to have learned was that AI Mode was too lazy to go to the works themselves. It reviewed online linguistic studies.
There isn't a definitive linguistic study pinpointing the single most common adjective in Thomas's Kyd's plays....
Next it offered a half-dozen adjectives that
mirror the emotional intensity and moral complexity of his characters.
Finally, it ended with more distinctly “human” touches.
If you're diving into a textual analysis or building a corpus study, I can help you set up a method to track adjective frequency across his works. Want to go that route?
Most human of all, rather than admit it did not know the answer it changed the subject.
I could not help but reflect upon the fact that I did not ask what “definitive linguistic studies” said was the most common adjective. Nor did I ask to be invited to create a “corpus”.
Some time later I did ask the AI for a definition of adjective. It answered with a clear, concise and correct definition. It knew that there were such things as linguistic studies and knew to look to one for an answer. From those studies, it had an idea that there existed such a thing as “the plays of Thomas Kyd”. But what it gave no sign of knowing was what “adjective in the plays of Thomas Kyd” meant. Actual adjectives in the wild, as it were, it seemed to know nothing about.
We all may struggle to see objects out of their expected context. AI, it seems, can cannot see them at all. Microsoft Copilot AI Mode1 can only detect that words are written about them in studies. There are “adjectives” and there are “adjectives in the plays of Thomas Kyd” and the two exist in separate worlds.
The answer was the word “sweet”. The most common adjective in Arden of Feversham is also sweet. (The matches between Kyd and the play go further.) My question did not indicate in any way a desire to build a “corpus” of the vocabulary of Thomas Kyd. In fact, the final suggestion called upon me join an effort to instruct it how to discover the answer I sought.
The only thing that could change the answer from one researcher to another would be to select different plays as being by Thomas Kyd and/or what part of the plays in question had he written personally. It is this that I had been driving myself crazy with. This that I had taken a break from.
Munday began writing around 1577. His earlier plays and poetry are written predominantly in rhymed fourteener couplets as was common circa 1580. He doesn't pronounce the suffix -ed for the past tense as a separate syllable (unless to regularize the scansion of a line), does end the third person present tense of verbs “-eth” (unless the modern “-s” will regularize the scansion of a line). He loved trendy words and sayings (a la Euphuism), used the word nick to mean all of “steal,” “in the best of shape,” and “just in time,” and was the rare playwright to use the word “delicate” in the few of his plays that survive.
Munday often referred to his time as a spy in the English Seminary in Rome and the English county of Kent. By these, and his constant flattery of Edward de Vere, the 17th Earl of Oxford, we know that he co-wrote the fictional Chronicle of some of the principal events in the Life, Adventures, and Times of Edward Webbe, chief master gunner, his trauailes (1590) and other prose works. In the late 1570s, he briefly served as apprentice to the printer John Alde who published a ballad entitled “A ballad of the Deliuery of 266 prasoners from the Turkes,” in 1579, an event that relates to Edward Webbe. I've yet to find a copy.
 |
| Edward de Vere's Ulysses & Agamemnon (1584)! |
According to Collier's Extracts From The Registers Of The Stationers Company, page 96, a prose version was licensed in the same year to Thomas Dawson and Stephen Peele. The prose version was reprinted in 1608 and is the apparent source for some details in Edward Webbe.
It is by such extrinsic means as these that one may identify Munday as the author of an anonymous work. Patterns of vocabulary, grammar and/or prosody are not sufficiently pronounced.
As it turns out, the most common adjective in Munday's plays also seems to have been “sweet” (but not so as to be obsessive). The county of Kent is the setting of Arden of Feversham. And Munday had published a highly popular prose account of famous murders and violent deaths in England entitled A View of sundry Examples. Reporting many straunge murthers (1580). Of much more interest, the saying
When two bones are at strife for a dog, it is commonly séen:
That the third comes and takes it, and wipes their mouthes cléen.
appears both in Arden of Feversham and Munday's Fidele and Fortunio (1584). A third playwright, however, was a still better match than either Kyd or Munday for at least half of the play. But that will have to wait another day.
As for Microsoft Copilot AI Mode (and Google AI), while they are not the most developed applications of AI, their failures do inform us. Textual analysis and author attribution are tedious and delicate tasks. The first question the AI needed to answer was what plays it would identify as being by Kyd. Then it needed to identify what portions of the plays were actually written by Kyd. These tasks are highly demanding.
Having made its choices in those categories, determining the most common adjective is a simple counting exercise. There is no indication that the AI even attempted to do either. Admittedly, after many years of practice this all will still take me many hours and I will likely never be totally satisfied with the precision of my data. But I am not a glorious AI Large Language Model computer program, yet I have at least arrived, in the process, at any number of simple word counts across dozens of plays all of which seem to be beyond the capability of AI at any speed.
Much bigger players than I also report that AI is only being trusted to generate output to be carefully reviewed by high level experts before being accepted.
1Google AI Mode did no better. “a definitive answer requires detailed textual analysis”.
Also at Virtual Grub Street:
Invention in a Noted Weed: the Poetry of William Shakespeare. September 21, 2024. “The coward conquest of a wretches knife,...”
The Sonnets of Shakespeare: Sonnet 108. Edward de Vere to his son, Henry. “That may expresse my love, or thy deare merit?”
- Sonnet 130: Shakespeare's Reply to a 1580 Poem by Thomas Watson. September 7, 2024. “Interesting to see our Derek Hunter debating with Dennis McCarthy, at the North group,...”.
- Rocco Bonetti's Blackfriars Fencing School and Lord Hunsdon's Water Pipe. August 12, 2023. “... the tenement late in the tenure of John Lyllie gentleman & nowe in the tenure of the said Rocho Bonetti...”
Check out the Shakespeare Authorship Article Index for many more articles and reviews about this fascinating time and about the Shakespeare Authorship Question.
- Check out the V.G.S. Oxfordian Shakespeare Poetry Page for many poems by Shakespeare together with historical context.
No comments:
Post a Comment