This is a republishing of an article I wrote for the Strange Bedfellows project. The original url is here.
Early modern drama, and Shakespeare’s works in particular, have a very troubled past. So troubled that, especially in Shakespeare’s case, editors have been fighting over what exactly were Shakespeare’s “true words”. Early modern methods of printing, copying, and ideas of copyright were varied to say the least. Thanks to nearly infinite variation that comes from this, any play that has Shakespeare’s name should also have the names of its editors. So much is involved in editing Shakespeare and early modern plays, the length of time involved in editing the Cambridge Edition of Ben Jonson’s works comes to mind, that any editor could easily be considered a coauthor. Editors resolve variant issues anywhere from single words, like “hath” or “have”, “Ay” or “I”, etc., to complete passages, like in Act 3.1 of King Lear. Editors have fought over these issues for so long, contrasting schools have formed, and different editions have be published in response to others, that I feel like a particularly modern question must be asked of this early modern “drama”: “Do individual words matter, specifically in literature?”
On one hand, this seems like a question of summation. If individual words are changed throughout the narrative, eventually the effect will reach a critical mass and be substantially different. But on the other hand, modern stylistic, i.e. statistics driven, modes of literary analysis suggest otherwise. Sample sizes ranging anywhere from 2,500+ words can be statistically unaffected by changes in random words.  With works of literature like Shakespeare’s ranging upwards of 30,000 words (Hamlet), it would not seem that literary variation would matter much. However, when analyses depend on counting specific words, like Arthur Kinney’s method which counts words like “gentle, beseech, answer, spoke, and purpose“, are we dealing so much of a matter of authorial stylistics or correctly preparing the data source?
Messy data and text files are fairly common; see the Moby Shakespeare and Project Gutenberg’s text of Shakespeare’s plays. “Definitive” editions like the Internet Shakespeare Editions try to prepare the text much like a print edition. However, given the outstanding problems and inherent complexities of the original works, computational stylistics, at least on corpora of early modern drama, might seem like a moot point. John Burrow presents an interesting retort to Vicker’s damnation of word frequency tests (a sub-division of computational stylistics) in noting that humanists are like statisticians in the comparative nature of their scholarship. Further mathematical complexities are the result of trying to make these comparisons as “equitable” as possible. However isn’t there a human element in literary analysis that is being forgotten; actually reading?
No matter how much I am interested or involved in digital humanities and computational approaches, I still feel like the literary element of texts is what separates us from statisticians. But then again, at what point during the process of reading do I actually pay attention to individual words? Usually I am reading for plot, motifs, characterization, or other literary elements. If I am trying to make a particular point about interpretation, I might hone in on individual words but that is always in the secondary phase of understanding a text, once the initial reading has been completed. When reading a narrative, whether or not for the first time, I honestly believe that I do not notice individual words as much as I would have thought. In this case, is the computer functioning like my brain, by normalizing variation within a certain frequency? At this point, the question posed above might be better rephrased as “when do words matter in literature?”
Obviously words matter in the very largest sense; literature would actually not be what it is without them. But I am going to notice the textual variation between two similar versions of Hamlet? That I am not so sure about. This is a conundrum that I have been thinking about for several years and particularly its implications. Could one word change a story completely? Possibly and possibly not.
 Eder, Maciej. “Does Size Matter? Authorship Attribution, Small Samples, Big Problem.”
<http://dh2010.cch.kcl.ac.uk/academic-programme/abstracts/papers/pdf/ab-744.pdf>. Accessed 17 February 2013.
 Vickers, Brian. “Shakespeare and Authorship Studies in the Twenty-First Century”. Shakespeare Quarterly (62) 2011: 106-142. p. 127.
 Burrows, John. “A Second Opinion on ‘Shakespeare and Authorship Studies in the Twenty-First Century”. Shakespeare Quarterly (63) 2012: 355-392. p. 359.