Make a Donation

Consider making a donation to support SHAKSPER.

Subscribe to Our Feeds

Current Postings RSS

Announcements RSS

Home :: Archive :: 2004 :: March ::
Stylometrics
The Shakespeare Conference: SHK 15.0775  Monday, 29 March 2004

[1]     From:   Ward Elliott <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
        Date:   Friday, 26 Mar 2004 15:29:14 -0800
        Subj:   RE: SHK 15.0767 Stylometrics

[2]     From:   David Evett <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
        Date:   Friday, 26 Mar 2004 22:03:07 -0500
        Subj:   Re: SHK 15.0750 Stylometrics


[1]-----------------------------------------------------------------
From:           Ward Elliott <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
Date:           Friday, 26 Mar 2004 15:29:14 -0800
Subject: 15.0767 Stylometrics
Comment:        RE: SHK 15.0767 Stylometrics

I. William Davis wonders how stylometrics are received by the full body
of Shakespeare scholarship and whether Shakespeare's writing evolved.
Neither Valenza nor I, nor our Shakespeare Clinic students, are
Shakespeare regulars who could in any way properly offer to speak for
the full body of Shakespeare scholarship. But we have published over a
dozen articles in the top refereed journals, including the Shakespeare
Quarterly, so I would guess that our work is at least within the
regulars' range of doctrinal tolerance.

Many SHAKSPERians are understandably jumpy about drawing conclusions
from outsiders' novel, stylometric evidence which still has not yet been
digested by the profession.  So were we in 1990-96 when our findings
really were novel and tentative.  But we are much less jumpy now, for
three reasons:  1) our internal-evidence conclusions have no gross
conflicts with external evidence.  Our tests don't say that Sejanus was
by Shakespeare or Hamlet by someone else.  2)  For plays we have found
so much redundancy of evidence that half our tests could be thrown out
for whatever reason and the remaining tests would still put almost all
of Shakespeare's and others' plays on different planets.  All those
gazillions of relative odds mean we've got a large margin of error.  And
3) our actual errors turned up since 1996 have turned out to be
vanishingly small, and certainly not for lack of vigorous adversary
scrutiny.  Besides passing muster with our publishers' referees, our
evidence has survived highly adversary, shock-and-awe challenges from
Don Foster and others and emerged all but unscathed.  I have no doubt
that further errors will turn up under further examination, but I would
be surprised if they were big enough to change our outcomes.  And I'm
confident enough to back it up with a $1,000 bet and to hope that it
will help make evidence like ours more digestible to Shakespeare regulars.

II. Bill Godschalk wonders what our tests say about Arden of Feversham
and Edward III.  The short answer for the whole plays is that they don't
look much like Shakespeare's solo work.  Arden has 13 rejections in 48
tests, Edw3 has 10.  No core Shakespeare play has more than two.
Comparative odds against Shakespeare authorship are very long, in the
mid-gazillions.
Whether Shakespeare could have written the "Shakespeare Scenes" of Edw3
is a much harder question, for which we have done much of the
preliminary counting, but not yet the full analysis.  For these shorter,
more variable passages we have only 12-14 validated tests, not 48.  Of
the suspected "Shakespeare scenes," 1.02 has one rejection and could be
Shakespeare (again, note that we don't say *must be* or *is*
Shakespeare's).  2.01, 2.02, and 4.04 have two rejections each, and fall
outside our Shakespeare profiles, but only by one order of magnitude,
not gazillions.  The odds that three out of the four "Shakespeare
scenes" would fall outside the Shakespeare envelope and still be
Shakespeare are lower.

Most, but not all, of the "non-Shakespeare scenes" also fall an order or
two of magnitude outside the Shakespeare envelope.  1.01, 3.01-3.05, and
4.01-4.03 have two or three rejections each - but 5.01 has only one
rejection and 4.05-4.09 none at all, which means that our tests by
themselves can't rule them out as "possibly Shakespeare."  Bill
Godschalk and other SHAKSPERians are more into Edw3 than we are, and
perhaps can cast further light on these outcomes.  I know that some
people think the whole play is Shakespeare's (we don't), but I'm not
aware of anyone who has singled out 5.01 or 4.04-4.09 as particularly
Shakespearean.

We expect to do more work on Edw3 this summer, looking to see how useful
tests we consider well-validated for long, single-authored plays can be
for sorting out shorter, possible multiple-authored passages.  Among
other things, having looked at our preliminary results and compared
notes with present-day screenwriters, we are becoming more and more
skeptical, not that Shakespeare did his share of co-authoring, but that
he and his co-author divided the work up neatly scene by scene to make
it easier for stylometrists to determine neatly and precisely who wrote
what.

For now, on balance, I am doubtful that Shakespeare wrote either the
"Shakespeare Scenes" or the "non-Shakespeare scenes" of Edw3.  The
comparative odds against most of the individual scenes are already an
order or two of magnitude worse than the worst core Shakespeare outlier;
the comparative aggregate odds against either *set of scenes,* taken as
wholes, coming from Shakespeare's hand would be lower yet.  If this is
so for our "couldn't-be" scenes, we would guess that the individual
"could-be" scenes are probably false positives, of which some are to be
expected in non-Shakespeare samples shorter than 3,000 words.  The tests
are not perfect, and they get less and less perfect as the text blocks
at issue get shorter.  Hence, we see it as a close call, too close for
us to say that reasonable people could not conclude otherwise.  If they
do, I would think they would want a look at our evidence before making
up their minds.

Whether Shakespeare wrote the whole play solo is not such a close call.
  Our tests say the comparative odds against it are many orders of
magnitude greater than those against the individual scenes, and the task
of justifying a Shakespeare ascription, correspondingly many times more
daunting.

Ward Elliott

[2]-------------------------------------------------------------
From:           David Evett <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
Date:           Friday, 26 Mar 2004 22:03:07 -0500
Subject: 15.0750 Stylometrics
Comment:        Re: SHK 15.0750 Stylometrics

 >In any
 >case, I should expect Milton to have the greater vocabulary, with his
 >polyglot education.

I haven't seen Ward's full argument - I'd be curious to know whether he
included in his count Milton's Latin as well as English works - but I'm
not persuaded that level of education is the only relevant issue.
Milton's elitist genres promoted use of learned words with Greek and
Latin roots.  Shakespeare's work for the popular theater invited
recourse to the demotic, mostly Germanic roots though with much French,
including the jargons of farmers and sailors and soldiers and tailors;
I've always loved his exhilarating plunges and leaps of register.  More
narrowly, the kind of education and the uses to which it is put are also
factors.

David Evett

_______________________________________________________________
S H A K S P E R: The Global Shakespeare Discussion List
Hardy M. Cook, 
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 
The S H A K S P E R Web Site <http://www.shaksper.net>

DISCLAIMER: Although SHAKSPER is a moderated discussion list, the
opinions expressed on it are the sole property of the poster, and the
editor assumes no responsibility for them.
 

©2011 Hardy Cook. All rights reserved.