Electron probability density cloud of hydrogen atom
We have thought a lot about what a “text” is in literary studies over the last few decades, spurred on by editorial theory, deconstruction, new media studies and book history. A nominalist by inclination, I tend to think of a text (real or digitized) as a provisional state of something, this other something being a hypothetical ideal or a fiction of analysis. So when I encounter a print version of a Shakespeare play, I am encountering an entity (for example, 1 Henry VI) in a state that is more or less suitable to the medium of print. But the printed play is not the performance. Nor is it whatever idea Shakespeare had when he began working with his company on the play.
An additional complication: versions of any given Shakespeare play in print — those found in the 1623 First Folio — may contain variation at the level of the individual word or character, variation that (in the case of the Folio) is corrected during the print run. Whatever is “behind” the First Folio, then, that original is a reconstruction of something that can only be said to exist in an ideal sense. We can think of that meta-object as having a probabalistic character: different letters in particular positions have a likelihood of being x or y, for example. But in the end, the actual identity of even an individual character must be understood as a likelihood.
None of these ideas except the last is particularly novel in Shakespeare studies. Peter Stallybrass and Margareta de Grazia, among many others, have already made the point that the sources behind Shakespeare plays are an editorial ideal — approximated in practice but unreachable in an ideal sense. Less has said, however, about the probabalistic nature of the text itself: its existence as a set of likelihoods that realized provisionally in different cases. A text as a cloud of probabilities. That’s interesting.
One Comment
Fred Jelinek is supposed to have said “Every time a linguist leaves the room our results improve.” Jelinek was an important figure in the debate about whether computational modeling of language should follow rules or probabilities. The latter approach won, by and large. For the humanist scholar in a “readerly” mode this is a bit of a challenge. You will get more out of the digital text if you let the machine work in ways that you would not work yourself. Treat the text “as” a network of probabilities, and it will help you with your “reading.” Perhaps “reading” can be demystified and turns out to be nothing but its own cloud of probabilities. To use another anecdote, there is the little old lady who believe the earth sat on a big turtle. When somebody asked her what the turtle sat on, she said “you naughty man, it’s turtle all the way down.” But it may be probabilities all the way down.