RE: Data-mining Hamlet

["Nat Gustafson-Sundell" <n-gustafson-sundell@northwestern.edu>]

This coverage is surprising since Moretti's approaches, at least 
as reported here, have been around for awhile.  At the most 
recent Digital Humanities and Computer Science colloquium, there 
were several variations on his approaches (and lots of different 
ways to make network analyses look different from one another), 
though one of the underlying currents was to question if there is 
anything really to be gained from such crunching, whether really, 
it is anything but a toy or will ever amount to something really 
substantial.  Obviously, there are a significant number of 
practitioners who seem to think there is something there, but the 
savviest among them seemed to be careful about how they expressed 
their expectations of the new tools.

Obviously, for example, a Shakespeare scholar already knows that 
Rosenkrantz and Guildenstern don't speak to each other.  One 
prominent DH and Shakespearean scholar, Professor Martin Mueller, 
speculated (at the colloquium) that DH computational analysis has 
not really offered any new insight, although it has been fun 
playing with the graphical representations of the data (...lots 
of sound and fury signifying nothing?  How does that line go?) 
which has otherwise been knowledge internalized and thunk upon by 
scholars, although he seemed committed, perhaps from a sense of 
duty or perhaps from a sense of play, to keep experimenting.

Personally, I too find the graphics appealing.  I also think 
there may be something to some of the approaches.  Mere text 
analysis seems to me unlikely to yield much more than toys, 
although the more sophisticated approaches enabled by some of the 
xml mark-up projects underway may allow some meaningful depth. 
Story analysis could be interesting.  The network analysis 
described in this article is one fairly simple approach (see for 
comparison the 2 network analyses presented in 
http://chicagocolloquium.org/dhcs-program-2010/).

I recently set up a DB for a scholar to do a kind of story 
analysis which could end up being very interesting, but I won't 
describe it here since she is just getting started -- but I think 
the best of these kinds of projects will require a great deal of 
scholarly interaction and interpretation in order to "uncover" 
the data. (as an aside, this project reminded me of Propp's 
formal analysis of Russian fairytales, so I looked to see what 
use was being made of his work in a DH context, of course to find 
a couple of sites allowing one to build one's own fairytales)

The issue of value, I suppose, must be personal.  As a kid, I 
loved books of lists and probably still would if I had more time, 
and some of these analyses feed right into that interest.  When 
did certain words appear or become prominent in parlance in 
newspapers or fiction?  Do we learn anything from such an 
analysis?  I think the answer depends on the person who sees the 
analysis.  A scholar probably already knew it in some way, but 
the use of the analysis could make for a point, probably among 
many, in an article or chapter.

As for demonstrating the merit of mass digitization projects, I 
think that is already happening in lots of different ways -- 
among them the fact that students are enabled to close-read texts 
(and compare editions) that were otherwise less available (see 
some of the reading lists at the Hathitrust). I recently saw an 
article (sorry no link since I can't remember where I saw it), 
about how there's a movement among foodies to go back to 17th, 
18th and 19th century recipes and food preparation techniques. 
The source docs enabling this "movement"? Digitized texts now 
available for free download.

-Nat

-----Original Message-----
From: owner-liblicense-l@lists.yale.edu
[mailto:owner-liblicense-l@lists.yale.edu] On Behalf Of Joseph Esposito
Sent: Monday, June 06, 2011 9:01 PM
To: Liblicense-L@Lists. Yale. Edu
Subject: Data-mining Hamlet

Fascinating piece on computational analysis of Hamlet at 
boston.com:

http://bo.st/lEa35M

And the long link:

http://www.boston.com/bostonglobe/ideas/articles/2011/05/29/hamlet_and_the_region_of_death/?page=full

I think this is the kind of work that will eventually demonstrate 
the merit of mass digitization projects.  I would love to see 
someone working in this field put together a demonstration site 
for a popular audience--data-mining of the complete works of Jane 
Austen, for example.

Joe Esposito