Afterthoughts: Week 4

As we turn, next week, to the question of character, we pause temporarily in assembling formal resources (prosodic, rhetorical, statistical) to apply them to a new problem. In the next few days I’m going to try to set down a few thoughts about the territory we’ve covered, but for the moment a few—somewhat fewer, I hope!—remarks about where we were on Wednesday.

Matthew Harrison, a former graduate student in the department, was enormously helpful in gathering the resources we used, and he made a useful distinction to me: between distant reading (the Franco Moretti high-altitude survey of features that emerge across corpora of many texts) and something like computer-assisted close reading. The latter is what Hope and Witmore do with “the” in Macbeth, and what Mary was suggesting with the unusual prominence of the word “through” in MND. The digital finding shows us something we didn’t notice (at least not consciously? does some share of the authority of digital methods depend on their pointing out phenomena that might affect us, even if we are not aware of being affected?), but we then go back and read the way we always read.

Our characteristic use of the technologies, however, especially Wordhoard, was something else again—perhaps closer to the use of digital technologies to make attributions and detect forgeries. For we are in the business, after all, week to week, of making forgeries. Everyone used one of the other of the tools to test the frequency or typicality of a given construction, whether and where it appeared elsewhere in Shakespeare’s works generally or in our play in particular. Mary checked the n-gram for “through” vs. “thorough” and found the former spelling increasingly prominent—one way in which we can estimate whether a given word in Shakespeare is coming in (and he sounds up-to-date, fashionable, etc.) or going out (and he sounds old-fashioned, archaic, etc.). Yan’s valorous by-hand comparison of local repetitions across three plays suggested that MND might be an outlier in that respect…

…which led into a very interesting discussion of the fact that the aristocrats speak twice as often as the mechanicals in the play—something that surprised us all, given how memorable the mechanicals’ lines are. That’s pretty basic DH, but still, it got us thinking about memory and meter and also, per Yan’s result, about repetition. We were perhaps being a little more sophisticated (whatever sophistication is worth—the play certainly has questions about that) when we observed that many of the words that were notably frequent in the play, relative to the rest of the corpus, were the mechanicals’ and the faeries’ words. If they speak less, they must therefore repeat more. Interesting: repetition and class? Repetition and song?

Incidentally, I had a very interesting discussion with Yan in office hours about the methodologies of some of our tools, their assumptions and procedures…Yan, could I tempt you to set down a few thoughts here? I had hoped to open all that up in class, but we went elsewhere.

Let’s see: also an interesting exchange with, if I remember rightly, Maddy and Will and others about punning: metonymy or metaphor? The adjacency (or identity?) of the sounds seemed metonymic, but Maddy pointed out that the tenors (soul and sole, for example) can be quite far apart. But perhaps a pun does not become a metaphor unless its discrepant meanings are made to illuminate one another, as opposed to just making us laugh by their dissonance? Anyhow this one is worth filing away for when we return to wordplay in earnest with Hamlet (the first of our two weeks with the play, oriented generally toward the historicity of language, will treat words, words, words, as Hamlet himself puts it).

I’ll mention briefly the nice work we did with Eli’s passage, the back-and-forth between Puck and Oberon about fairy lifeways. Eli helped us see the stylistic variety, and Scott pointed out how the somewhat hectic enjambment and relatively free placement of caesura in Puck’s agitated (or just puckish) account of the dangers of sunlight contrasted with how evenly, how magisterially Oberon laid his syntax across the lines. (Per Derek Attridge, Scott described this as the ascendence of the phrase as a rhythmic unit over the line; that’s a thought we might return to.)

Things we wish we had: tagging not just for foreign words but words by etymological origin; and by verse type (Maddy wondered where else, besides Macbeth, we could find catalectic tetrameter). What else?

I ended up by giving us a taste of the Shakespeare produced by the recurrent neural network (sent to me by Yan); a few choice lines (if “choice” is the word):


  As I have all the very line, that gave me your highness,
  Where you will hear the single spirit of my business,
  Plant down flives on your son, and even
  And open with their own conusteries; and thinking
  your grave ship should ne’er break Humphrey’s eyes,
  I am poor dear party to make his chamber
  And hospish shameless frozen pride. Here name,
  And light in plot legely in whom I said,
  Glimmed by an argument of it sweet fears your other mouth,
  Such a great estimation would be run as this,
  ‘Tis fit for them, ’tis talk before yourselves.

I confess, when I read this, I can get almost weepy from laughter. We talked just a bit about why Shakespeare’s own language is funny, and Bergson’s idea that comedy is the encrustation of the mechanical upon the organic; that seems to have something to say about certain compulsive, repetitive speakers in the play. Does it account for the comedy (if you find it comic!) of machine Shakespeare? What about the lines below, from the Shakespeare pastiche in “Beyond the Fringe” (the pre-Python British sketch television show)?

  Get thee to Gloucester, Essex. Do thee to Wessex, Exeter.
  Fair Albany to Somerset must eke his route.
  And Scroop, do you to Westmoreland, where shall bold York
  Enrouted now for Lancaster, with forces of our Uncle Rutland,
  Enjoin his standard with sweet Norfolk’s host.
  Fair Sussex, get thee to Warwicksbourne,
  And there, with frowning purpose, tell our plan
  To Bedford’s tilted ear, that he shall press
  With most insensate speed
  And join his warlike effort to bold Dorset’s side.
  I most royally shall now to bed,
  To sleep off all the nonsense I’ve just said….

It’s maybe a cheap, melodramatic question: is this Shakespeare? But I’d love to try to rescue it from the temptations of donnish posturing. Does Shakespeare consist in a set of texts? Or a way of writing? Or (that word again: we’ll really go after it I hope in Week 8) a sound? Maybe there is some version here of what in AI is called the Turing Test. The test: if, exchanging messages at a terminal, you cannot tell the difference between a machine and a human interlocutor, then the machine is intelligent. (I.e., who cares how the machine works, what’s behind the curtain, if it can do what we do?) You could apply a similar test to style. If it sounds like Shakespeare, if you really can’t tell the difference, is it Shakespeare—regardless of when it was written, by whom or by what? Who thinks so, who thinks not, why?


PS a little more “Beyond the Fringe,” when they (Jonathan Miller, Allen Bennett, Peter Cook, and Dudley Moore) reenter as “Rustics”:

Miller: Is it all botched up, then, Master Puke?
Bennett: Aye, and marry is, good Master Snot.
Moore: ‘Tis said our Master, the Duke, hath contrived some naughtiness against his son, the King.
Cook: Aye, and it doth confound our merrymaking.
Miller: What say you, Master Puke? I am for Lancaster, and that’s to say for good shoe leather.
Cook: Come speak, good Master Puke, or hath the leather blocked up thy tongue?
Moore: Why then go trippingly upon thy laces, good Grit.
Cook: Art leather laces thy undoing?
Moore: They shall undo many a fair boot this day.
All: Come, let’s to our rural revel and with our song enchant our King.

Surely that is Shakespeare.

Leave a Reply