Make a Donation

Consider making a donation to support SHAKSPER.

Subscribe to Our Feeds

Current Postings RSS

Announcements RSS

Home :: Archive :: 1996 :: March ::
Re: Funeral Elegy
Shakespeare Electronic Conference, Vol. 7, No. 0238.  Tuesday, 26 March 1996.

(1)     From:   Jim Helfers <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
        Date:   Monday, 25 Mar 1996 14:07:31 -0700 (MST)
        Subj:   The Funeral Elegy

(2)     From:   Richard J Kennedy <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
        Date:   Monday, 25 Mar 1996 11:54:27 -0800
        Subj:   Re: Funeral Elegy

(3)     From:   Roger D. Gross <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
        Date:   Tuesday, 26 Mar 1996 13:01:05 -0600 (CST)
        Subj:   Funeral Elegy


(1)----------------------------------------------------------------------------
From:           Jim Helfers <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
Date:           Monday, 25 Mar 1996 14:07:31 -0700 (MST)
Subject:        The Funeral Elegy

I've been following with interest the ongoing discussion concerning _The
Funeral Elegy_.  Although I begin to believe that whatever else needs to be
said should be said in careful scholarly studies, I have my crumb to add to the
discussion.  I especially find interesting Professors Kennedy and Godshalk's
comments on stylistic questions, specifically the general quantification of
function words and feminine line endings.  Having just waded painfully through
the swamp of stylometry for a presentation, however, I'd like to inject a
cautionary note concerning modes of proof which use quantitative study
(statistical descriptions of linguistic items and patterns in texts).

First a bit of background.  In 1978, Andrew Q. Morton, a classical and biblical
scholar, published the book _Literary Detection:  How to prove authorship and
fraud in literature and documents_.  In it he proposed a new approach to
studies of attribution.  Instead of examining texts for special terms and
deducing authorship from their use, as has been done in biblical studies since
the nineteenth century (and which Donald Foster is doing with SHAXICON, in
part), Morton asserted that the statistical description of universally
occurring characteristics of text (such as sentence length, or the rate of
occurrence and collocation of syntactic function words like articles,
conjunctions and prepositions) would show patterns of usage which would be
unique to a particular author.

One would, theoretically, using multiple samples of different function words or
collocations of words, be able to come up with a unique statistical description
of a given author's style, using such statistical measures as mean, median,
mode, standard deviation, and other calculations.  One needs statistical
measures to assure that differences between samples result from something other
than random variation.

This theory is immediately attractive.  It indicates that one could identify
unique features of styles in the same way that one distinguishes fingerprints,
by looking for discrete patterns at particular points.

Unfortunately, a number of scholars, notably M. W. A. Smith and Barbara
Stevenson (Smith, _Language and The Law_ 374-413, Stevenson, _Literary
Computing and Literary Criticism_ 61-74), have pointed out grave difficulties
with Morton's methodology and reporting.  Their critiques center on three
areas: Morton's initial framing of his hypotheses, his questionable
assumptions, and his handling and analysis of the data.  Among the several
assumptions they question, the most important involves Morton's methodology.
Morton uses statistical tools which assume the independence of the variables
that they statistically graph.  The linguistic elements that Morton graphs are
not, in fact, independent of each other.

A question arises: Is Morton's stylometric method a valid way to analyze texts,
in light of Smith's and Stevenson's criticisms?  Morton himself has proposed
several modifications of his procedures, but his new methodologies seem to be
under attack for the same reasons as his original ones (Smith (1994) 412-413).
However, both Stevenson and Smith set out further parameters which could
conceivably validate stylometry.  Stevenson suggests a simulation study,
specifically, resampling.  However, this resampling "requires hundreds of
complicated chi-square manipulations, not just one simple test" (Stevenson 71).
Smith suggests a revised frequency test, in which certain words are chosen for
analysis based both on their syntactic function and their frequency of
occurrence in authors' texts.  The sampled words and collocations are then
manipulated in a variety of complex ways, to achieve a valid result (411-12).

This discussion probably, in Barbara Stevenson's words, "portrays perfectly the
current status of computational stylistics:  the experts cannot agree on the
ways statistics should be adapted to literary criticism, and statistical
novices are unable to understand the jargon of the experts"  (61).

In any case, stylometry is a method of adducing internal evidence, and internal
evidence is only a single factor in an overall effort to identify an author.
This effort is a complex one involving both internal and external evidence.  It
is also true that, for many reasons, some kind of stylometric analysis will
remain a factor in overall efforts to identify authors of medieval and
Renaissance texts.  One of the most interesting recent projects, which makes
use of both "traditional" stylometric methods and some new techniques is The
Shakespeare Clinic, a project undertaken by undergraduates and professors since
1988 at the Claremont Colleges (Elliott, "Touchstone" 199).  The project is
unique in two ways:  first, some of the major professorial contributors are not
literary scholars.  Second, Ward Elliott and Robert Valenza (two of the
professorial leaders of the project) have pioneered a further refinement of
statistical measurement for stylometric analysis, which they call modal
analysis.

Finding that another method the clinic used was strongly sensitive to genre,
and required quite large sample sizes for statistical validity, Elliott and
Valenza took a new tack.  They used a technique called the Karhunen-Loeve
transform (KLT) "to determine the principal modes by which an author deviates
from his or her average usage of selected keywords.  These modes do not
directly represent keyword occurrences, but instead measure complex patterns of
deviation from the average rates" ("Touchstone" 201).  This they found to be an
effective way to discriminate between the style of two compared poets.

The preceding information impacts the _Elegy_ discussion in two ways:  first,
it strongly suggests that the simple comparison of word occurrence numbers, or
even numbers of feminine-ending lines, is a potentially flawed way to argue for
or against authorship.  Second, it leads me to ask a question:  Is any member
of the Shakespeare Clinic out there?  I'd be interested in reading the results
of the Clinic's analysis of the _Funeral Elegy_, which, I hear, cast doubt on a
Shakespearean authorship from a stylometric perspective.

For anyone who is interested in stylometrics as a tool for resolving authorship
disputes, I've appended a short list of references on stylometry in general,
and Shakespearean stylometry in particular.

Elliott, Ward E. Y.  "Glass Houses and Glass Slippers:  The Shakespeare
     Clinic and Its Critics."  _The Shakespeare Newsletter_ 40.4 (1990
     Winter): 59.

Elliott, Ward E. Y. and Robert J. Valenza.  "A Touchstone for the Bard."
     _Computers and the Humanities_ 25 (1991): 199-209.

Elliott, Ward E. Y. and Robert J. Valenza.  "Was the Earl of Oxford the True
     Shakespeare?  A Computer-Aided Analysis."  _Notes and Queries_ December
     1991: 501-506.

Foster, Donald W.  _Elegy by W. S.: A Study in Attribution_.  Newark: U of
     Delaware, 1989.

_______________.  "Reconstructing Shakespeare 1: The Roles that Shakespeare
     Performed."  _The Shakespeare Newsletter_ 41.1-2 (1991 Spring/Summer):
     16-17.

_______________.  "Reconstructing Shakespeare Part 2:  The Sonnets."  _The
     Shakespeare Newsletter_ 41.3 (1991 Fall): 26-27.

_______________.  "Reconstructing Shakespeare Part 3 of 3: New Directions in
     Textual Analysis and Stage History."  _The Shakespeare Newsletter_ 41.4
     (1991 Winter): 58-59.

_______________.  "Re: SHAXICON."  The Shakespeare Electronic Conference
     (
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 ) 6.0533 (Thursday 6 July 1995).

_______________.  "Re: SHK 6.0874, Re: Julius Caesar, Re: SHAXICON."  The
     Shakespeare Electronic Conference (
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 ) 6.0891
     (Thursday 10 November 1995).

_______________.  "Stylometry and Quantitative Stylistic Analysis."  The
     Shakespeare Electronic Conference (
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 ) 7.0031
     (Friday 12 January 1996).

Morton, Andrew Q.  _Literary Detection:  How to prove authorship and fraud in
     literature and documents_.  New York:  Charles Scribners Sons, 1978.

Smith, M. W. A.  "Computers, Statistics and Disputed Authorship."  In John
     Gibbons, ed.  _Language and the Law_. London : Longman, 1994.  374-413.

_______________.  "Counting Wilkins In:  Stylometry Reveals Who Wrote Acts I
     and II of 'Pericles.'"    _The Shakespeare Newsletter_ 40.4 (1990
     Winter): 60.

_______________.  "An Investigation of the Basis of Morton's Method for the
     Determination of Authorship."  _Style_ 19.3 (1985 Fall):  341-368.

_______________.  "Stylometry:  Will the Computer Finally End Authorship
     Controversies?"  _The Shakespeare Newsletter_ 41.1-2 (1991
     Spring/Summer): 14-17.

Stevenson, Barbara.  "Adapting Hypothesis Testing to a Literary Problem."  In
     Rosanne G. Potter, ed.  _Literary Computing and Literary Criticism:
     Theoretical and Practical Essays on Theme and Rhetoric_.  Philadelphia :
     U of Pennsylvania P, 1989.  61-74.

(2)----------------------------------------------------------------------------
From:           Richard J Kennedy <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
Date:           Monday, 25 Mar 1996 11:54:27 -0800
Subject:        Re: Funeral Elegy

I have found that Doctor Dodypoll (1600) was printed in Robert Dodsley's "A
select collection of Old English Plays", which was first published in 1744.
A.H. Bullen edited a 4 vol. edition of Dodsley's selections, published in 1964.
 He has this to say of Dodypoll:

"The writer, whoever he was, scatters his gold with a lavish hand.  In the fine
panegyric on painting, there is a freedom of fancy that lifts us into the
higher regions of poetry..."

However, Bullen finds that "In the conduct of the complicated plot no great
dexterity is shown.  There is a want of fusion and coherence."  It's too true,
the plot is difficult to follow, but that's very like Shakespeare.  Bullen
passes over this fault to recommend the poetry in the play.  He directs us to a
love scene, and says:  "The beauty of that scene is beyond the reach of any
ordinary poet.  And what shall be said of that exquisite description of the
cameo in ii,I?:"

FLORES:  See then (my Lord) this aggat that contains
        The image of that Goddesse and her sonne,
        Whom auncients held the Soveraignes of Love;
        See, naturally wrought out of the stone
        (Besides the perfect shape of every limme,
        Besides the wondrous life of her bright haire)
        A waving mantle of celestial blew
        Imbroydering it selfe with flaming Starres.
ALBER:  Most excellent:  and see besides (my Lords)
        How Cupids wings do spring out of the stone
        As if they needed not the help of Art.

Bullen comments:  "Is there in the whole Greek Anthology anything so absolutely
flawless?"  But Bullen will not dare to say Dodypoll might have been written by
Shakespeare.  "As to the authorship of Dr. Dodypoll I am unable to form a
conjecture."

And so once again we are reminded to check a too hasty attribution of a poem or
a play to Shakespeare, and I may well be wrong in thinking that Shakespeare
wrote Doctor Dodypoll. No doubt it has been looked at many times since
Dodsley's 1744 printing.  No doubt many scholars have looked at the Funeral
Elegy, enticed towards some great treasure by the W.S. initials. But the
chances that Shakespeare wrote such poor poetry are small indeed.  The chances
are a good deal better that Shakespeare wrote those several beautiful passages
in Dodypoll.

(3)----------------------------------------------------------------------------
From:           Roger D. Gross <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
Date:           Tuesday, 26 Mar 1996 13:01:05 -0600 (CST)
Subject:        Funeral Elegy

Re. 4-syllable words in FE:

I get somewhat different numbers from R. Kennedy:  he finds 75 four-syllable
words in FE.  I count 62.   I can guess that he may have counted "oblivion",
"experience", "melodious", as four-syllable words but they are here (and
normally in Shakespeare) given three syllables.  Also, he may have counted
"contemplation" as four syllables but it is here (and usually is when in the
last position in a verse line) a five-syllable word. But these possibilities
don't account for the difference in our numbers.

I do note one interesting thing:  of the 62 four-syllable words in FE, 17
appear nowhere else in Shakespeare.  I draw no inference...yet.

Also, "thank" appears in FE as a noun.  Odd.  Shakespeare's standard works have
309 uses of the word, always as a verb.  Perhaps this is a typographical error;
might it be "thanks"?

Onward.
Roger Gross
U. of Arkansas
 

Other Messages In This Thread

©2011 Hardy Cook. All rights reserved.