Charles Brockden Brown: Quantitative Analysis and Narrative Voice

Charles Brockden Brown: Quantitative Analysis and
Narrative Voice

Larry
Stewart

The College of Wooster
LStewart@wooster.edu

2002

University of Tübingen

Tübingen

ALLC/ACH 2002

editor

Harald
Fuchs

encoder

Sara
A.
Schmidt

In "Stephen Crane and the 'New York Tribune': a Case Study in Traditional and
Non-Traditional Authorship Attribution," a paper delivered at ALLC/ACH 2000,
David Holmes and Michael Robertson argue for the use of "objective, stylometric
evidence" to support "the 'traditional' scholarship on the problem of
authorship" and suggest that "this joint interdisciplinary approach should be
the way in which attributional research is conducted" (175) This paper presents
a case study to support the argument that such an interdisciplinary approach may
prove as important in literary analysis as in attribution studies.
Of course, in one sense, this recognition of the symbiotic relationship between
traditional and quantitative analysis is not new. J. F. Burrows' groundbreaking
Computation into Criticism brilliantly and seamlessly moves between statistical
and traditional literary analysis and clearly recognizes the dialogue between
the two. As well, any number of papers coming out of this conference use
quantitative analysis to test or illustrate insights from literary criticism.
(See, for example, Opas and Tweedie, "The Magic Carpet Ride" and their "Come
into my World: Styles of Stance in Detective and Romantic Fiction" or McKenna's
"The Statistical Analysis of Style: How Language Means in Beckett.") However,
the value of using statistical analysis to augment traditional criticism appears
not to have been widely recognized and it seems important to continue to insist
on the insights that each approach may bring to the other.
This case study concerns texts of Charles Brockden Brown, usually recognized as
the first professional novelist in the United States. For many years, Brown was
considered a relatively imprecise and careless stylist, who gave little
attention to character creation and particularly to narrative voice. However, a
revival of interest in Brown during the last fifteen to twenty years has brought
with it increased admiration for Brown as a stylist but with little close
examination of that style. The purpose of this study was to look more closely at
that style, particularly by examining the narrative voice in two texts: Wieland
or The Transformation (1798) and Carwin, The Biloquist (1803-1805). Wieland is a
novel of twenty-seven chapters, most of which are narrated by Clara Wieland, the
novel's Gothic heroine, but with three chapters narrated by each of three other
characters. Carwin is an unfinished novel, the first ten chapters of which were
published serially. Carwin, the Gothic villain of Wieland, narrates one chapter
of that novel and the whole of Carwin. The question with which this study began
was whether Brown had created in Carwin a character and narrator with a
distinctive voice, whether what we call the character Carwin is a distinct
literary or linguistic entity. If so, the evidence would seem to suggest greater
precision in the creation of narrative voice than has been traditionally
recognized.
To determine differences or similarities in narrative voice, the study utilized a
variation on what is sometimes called the Burrow's technique, using multivariate
analysis to discover patterns in the occurrence rate of the text's most common
function words. However, along with the use of the fifty most common words, this
analysis included as variables the occurrence rate of punctuation marks
(periods, commas, question marks, exclamation points, semicolons, colons, and
ellipses) and of other stylistic markers such as sentence and paragraph length.
These variables included the rate of short and long sentences (defined as five
or fewer words and twenty-five or more words), short and long paragraphs
(defined as fifty or fewer words and one hundred and twenty-five or more words),
average words per sentence, average sentences per paragraph, and percentage of
unique words. The latter variables were added because in the analysis of a
sample of fifteen late-eighteenth- and early-nineteenth-century novels, the
addition of these variables appeared substantially to increase the sensitivity
of the analysis. The presentation will discuss the variables which were most
effective in making discriminations. With the addition of these variables, a
principle components analysis was made for each chapter in Wieland and Carwin,
and the results of the two most significant factors were displayed on a scatter
graph. These results show a clear distinction between most of the chapters in
Wieland and those in Carwin. (See Figure 1.)

With the addition of these variables, a principle components analysis was made
for each chapter in Wieland and Carwin, and the results of the two most
significant factors were displayed on a scatter graph. These results show a
clear distinction between most of the chapters in Wieland and those in Carwin.
(See Figure 1.) However, chapter twenty-three of Wieland, the chapter narrated
by the character Carwin, is clearly situated among the chapters from Carwin.
Thus, it would seem that the narrative voice of Carwin is distinctive, that
Brown has created a character whose voice in one novel is recognizable as his
voice in the other.
Those results might have concluded the study were it not for the fact that two
other chapters from Wieland are situated in the constellation of Carwin
chapters, chapter twenty-seven, the final chapter of the novel, and chapter
thirteen, a chapter narrated by another character, Pleyel, a romantic interest
of the protagonist, Clara Wieland. It is when considering and attempting to
explain the location of these two chapters among those from Carwin that the
strongest case can be made for the joint use of quantitative and traditional
literary analysis. It is not just that quantitative analysis may support and
supply evidence for traditional approaches but that traditional analysis may
help us understand what we are seeing in the results of quantitative analysis.
Upon reflection, the analyst should probably not be surprised to find the
narration of Clara in the final chapter to have much in common with the
narrative voice of Carwin. Critics have long noted that Clara's voice in the
final chapter undergoes a transformation, and some have even suggested that
Clara has come ultimately under the influence of Carwin, whose villainy
throughout has been associated with his powers of ventriloquism. That he has
thrown his voice a final time and taken over the voice of the dominant narrator
has been at least hinted by critics. Quantitative analysis supports such an
interpretation, and the interpretation helps the quantitative analyst to
understand his or her results.
The initially more problematic finding is the location of chapter thirteen of
Wieland among the chapters from Carwin. Since it is a chapter narrated by
Pleyel, one might speculate that the narrator Clara has one voice and all other
narrators have another. However, such an explanation does not account for
chapter nineteen, narrated by Theodore Wieland, which differs both from the
chapters narrated by Carwin and from those narrated by Clara. The more likely
explanation comes from Steven Watts, who, in his book The Romance of Real Life:
Charles Brockden Brown and the Origins of American Culture, comments on the
usually unrecognized similarities of Pleyel and Carwin and, in fact, finds
Carwin to be the alter ego of Pleyel. Although this interpretation has not
gained wide currency, the similarities in voice indicated by the statistical
analysis suggest that it might be reconsidered.
These findings seem to support what J. F. Burrows points out: "that exact
evidence, often couched in the unfamiliar language of statistics, does have a
distinct bearing on questions of importance in the territory of literary
interpretation and judgement" (2). However, the findings also suggest that
traditional critical interpretation has a real bearing on how we understand the
meaning of those statistics.

J.
F.
Burrows

Computation into Criticism: A Study of Jane Austen and
an Experiment in Method

Oxford
Clarendon Press
1987

Bill
Christophersen

The Apparition in the Glass: Charles Brockden Brown's
American Gothic

Athens and London
U of Georgia Press
1993

David
I.
Holmes

Michael
Robertson

Stephen Crane and the 'New-York Tribune': a Case Study
in Traditional and Non-traditional Authorship Attribution

Paper delivered at ALLC/ACH 2000

2000

C.
W.
F.
McKenna

The Statistical Analysis of Style: How Language Means
in Beckett

Abstracts of the ALLC/ACH Conference 2000, Glasgow

2000

L.
L.
Opas

F.
J.
Tweedie

The Magic Carpet Ride: Reader Involvement n Romantic
Fiction

Literary & Linguistic Computing

14
1
89-101
1999

L.
L.
Opas

F.
J.
Tweedie

Come into my World: Styles of Stance in Detective and
Romantic Fiction

Abstracts of the ALLC/ACH Conference 1999, Virginia

1999
247

Donald
A.
Ringe

Charles Brockden Brown

New York
Twayne
1966

Steven
Watts

The Romance of Real Life: Charles Brockden Brown and
the Origins of American Culture

Baltimore and London
Johns Hopkins
1994