Make a Donation

Consider making a donation to support SHAKSPER.

Subscribe to Our Feeds

Current Postings RSS

Announcements RSS

Home :: Archive :: 2005 :: September ::
Two New Authorship Studies
The Shakespeare Conference: SHK 16.1563  Tuesday, 20 September 2005

From: 		Ward Elliott <
 This e-mail address is being protected from spambots. You need JavaScript enabled to view it
 >
Date: 		Monday, 19 Sep 2005 18:59:28 -0700
Subject: 	Two New Authorship Studies

Many SHAKSPERians have wanted a look at the long, documented case 
statements underlying our most recent postings. Valenza and I have just 
put two of our latest articles on the web:  our long Tennessee Law 
Review Article, "Oxford by the Numbers," and a shorter one on the 
"Shakespeare" sections of Sir Thomas More and Edward III, "Two Tough 
Nuts to Crack," to be published in the Shakespeare Yearbook next year. 
The web addresses are:

http://govt.claremontmckenna.edu/welliott/UTConference/Oxford_by_Numbers.pdf
http://govt.claremontmckenna.edu/welliott/UTConference/2ToughNuts.pdf

"Oxford by the Numbers" gives a detailed explanation of how we calculate 
the odds of Shakespeare authorship of a sample text, relative to those 
of the least Shakespearean play or verse block in our core Shakespeare 
baseline. It describes the distinctive features of our methodology: 
quantitative evidence; clean, commonized baseline; negative evidence; 
and our use of blocks and profiles.  It discusses cautions about some 
tests for genre, date of composition, prosody, and editing; it considers 
internal consistency, replicability; correspondence with available 
documents, margins of error, holding up under adverse criticism, and 
presents our old, $1,000 wager.  It also explains scientific notation 
and tells how to calculate it with a $15 scientific calculator.  Table 
2.4, extracted below, gives a summary of some of our results:

Table 2.4.  Shakespeare's Farthest Outlier 3,000-Word Poem Blocks 
Compared with Oxford, Marlowe, and Funeral Elegy

Poem		Tests	Rejections	Discrete 	Continuous
Block					Composite   	Composite
					Probability	Probability

Shakespeare	14	1		3.1E-01		9.0E-02
Oxford		14	6*		7.7E-07*	<1.0E-15*
Bacon		14	7*		2.3E-08*	<1.0E-15*
Marlowe 1	14	3*		5.2E-03*	1.6E-02*
Marlowe 2	14	4*		3.7E-04*	3.1E-04*
Elegy		14	6*		7.7E-07*	<1.0E-15*

Table 2.4.  * = outside Shakespeare range.  3.1E-01 means 3.1 times ten 
to the minus one = 31%.raw composite Shakespeare probability for the 
Shakespeare block.  7.7E-07 means 7.7 times ten to the minus seven = 
.00000077 raw composite Shakespeare probability for the Elegy. Relative 
composite Shakespeare odds are computed by dividing the raw probability 
score of the least probable baseline Shakespeare block (top line) by the 
probabilities observed for comparison samples (next 5 lines).

The three leading Shakespeare claimants and the Funeral Elegy are all 
far less likely than Shakespeare's own farthest outlier baseline blocks 
to have come from Shakespeare by chance.  The closest block, Marlowe 1, 
is 5-60 times less probable than Shakespeare's own outliers.  The most 
distant block, Bacon, is between 13 million and 90 trillion times less 
probable than Shakespeare's outliers.  Marlowe 1 would be a much closer 
call than Bacon, Oxford, or the Elegy, but Marlowe himself is hardly a 
likely Shakespeare claimant, thanks to his very distant second block and 
his long list of plays which don't come close to matching Shakespeare. 
Neither he, nor Oxford and Bacon, nor any of the 34 other Shakespeare 
"claimants" we could test are close calls.

All the claimants are on different stylistic planets from Shakespeare. 
So is the author of the Funeral Elegy, which, however, fits comfortably 
within our stylometric profile of John Ford.

For whole plays, some numbers of interest from "Two Tough Nuts" would be:

Play		Tests	Rejections	Discrete 	Continuous
					Composite   Composite
					Probability	Probability

The Tempest	48	2		2.3E-01	3.7E-03
Sir T More 	48	7*		3.3E-05*	<1.0E-15*
Edward III 	48	13*		4.4E-12*	2.6E-12*
Woodstock	48	20*		<1.0E-15*	<1.0E-15*

The Tempest falls just within the Shakespeare range and marks the outer 
threshold of our core Shakespeare baseline.  None of the three others 
listed are on the same stylistic planet with Shakespeare, even after 
appropriate cautions and discounts for genre, date, editing, and 
prosody.  Both versions of Woodstock's Shakespeare probability, Discrete 
and Continuous, are too low to compute on a standard PC.

These are figures for whole plays and do not rule out partial 
Shakespeare authorship.  The "Shakespeare" parts of Sir Thomas More and 
Edward III are much closer calls, but the odds in both cases still do 
not favor Shakespeare.  "Two Tough Nuts" also discusses which plays 
which appeared during Shakespeare's lifetime raise the most interesting 
questions of authorship; it also raises the ante of our wager from 
$1,000 to its present amount of 

 

©2011 Hardy Cook. All rights reserved.