More about gentleman in Dickens

Introduction The aim of this paper is to present a corpus-stylistic
study of the collocation of gentleman in Dickens.
The word gentleman is among the most frequent ‘content’
words in Dickens. In fact, as Fig. 1 shows, gentleman
appears more frequently in Dickens than in any other
author examined, and thus is a key word in Dickens in
the sense that it ‘appear[s] in a text or a part of a text with
a frequency greater than chance occurrence alone would
suggest’ (Henry and Rooseberry, 2001: 110).
Fig. 1 Normalised frequency of gentleman per million
words
Building on my pilot study of the collocation of gentleman
in Dickens and Smollett (Tabata, 2008), this study
expands the scope of analysis by comparing Dickens
texts with a larger reference
corpus covering major eighteenth-
and nineteenth-century authors as well as by
combining quantitative techniques with qualitative interpretation
of statistical fi ndings. The corpus, upon which
the present study is based, is made up of three components:
1) 24 Dickens text sets (4,835,158 words), 2) a
set of 23 eighteenth-century texts (4,163,353 words: Defoe,
Fielding, Goldsmith, Richardson, Smollett, Sterne,
and Swift), and 3) a set of 31 nineteenth-century texts
(5,118,346 words: Austen, the Brontë sisters, Collins,
George Eliot, Gaskell, Thackeray, and Trollope). The total
of running words amount to 14,116,857. One might
find that female authors
outnumber male authors in the
nineteenth-century set. However, the female set is not
so overpopulated as to imbalance the population of the
subcorpus. The total of the tokens by male authors accounts
for as high as 45 % of the running words due to
comparatively thick volumes produced by male authors. A few methodological issues
To investigate Dickens’ stylistic features associated with
the use of gentleman, it will be appropriate
to analyze
the word in collocation, rather than in isolation since semantics
of a word is extended to the surrounding words,
or co-text (Firth, 1957; Sinclair, 1991; Stubbs, 2001).
The fi rst issue to be discussed is a collocational span.
Although there is not a total agreement between
linguists
regarding an optimal range of collocational span, a generally
accepted practice is to examine collocation in a
span of four words to the each side of the node (Jones
and Sinclair, 1974). This is based on the fi nding that it
will become increasingly diffcult to fi nd meaningful collocational
pattens beyond a span of four words (Sinclair
et al., 2004), as shown in Fig. 2.
Fig. 2 Graph showing average node predictions over span
positions 1–10*
* From Sinclair (1969) as reprinted in Sinclair et al.
(2004)
Next comes the issue of how to measure collocational
strength between words. If we use raw frequency (or
normalised frequency) counts, the predominance of
function words such as the, a, and, of, etc. (as evident
in Table 1) would overcrowd subtler, more meaningful
patterns. My proposed solution is to use a statistical
measure. Among a number of techniques for fi ltering out
unimportant neighbouring words, a Mutual Information
(MI) score measures collocational strength, by a logarithmic
compression of the frequency of collocates. It
therefore is likely to spotlight semantic relationships
rather than syntactic relationships between the node and
its collocates. Mutual information score for the collocation
of the word x and the word y (I(x,y)) is obtained from
the formula (1).
I(x,y) = log2
f(x,y) × N
f(x) × f(y)
The third issue is how we set a threshold for variables.
Church and Hanks (1990: 24) state that the association
ratio becomes unstable when the counts are very small
(for example, when fx.y ≤ 5). My tentative proposal is
to base analysis upon collocates occurring 10 times or
more, and with MI scores higher than 3.0, drawing on
Church and Hanks’ account that “pairs with I(x,y) > 3 tend
to be interesting” (24). Table 2 list 100 strongest collocates
with f(x,y) ≥ 10. A close look at Table 2, however,
reveals that there are a few proper nouns that occur only
in a particular text (Brass’s, Greystock, and Oliver). As
a results of eliminating twelve such proper nouns from
the candidates for variables, 378 collocates of gentleman
were found qulified as variables. The initial number of
texts in the corpus was 78, but one text, Smollett’s The
History and Adventures of an Atom (1769) did not meet
the requirements and, therefore, was eliminated from the
data set.
The sheer volume of data to be analysed is daunting: a
collocational strength matrix for 378 collocates across
77 texts (77 rows by 378 columns). Since it would be
extremely diffcult for a human eye to detect meaningful
patterns from the data of such dimensions, correspondence
analysis is employed to visualize complex
interrelationships among gentleman’s collocates, interrelationships
among texts, and the association patterns
between the gentleman’s collocates and texts in multidimensional
spaces.
Table 1 100 most common collocates of gentleman
Various multivariate analyses of texts have been successful in elucidating linguistic variation
over time, variation
across registers, variation across oceans, to say nothing
of linguistic
differences between authors (Brainerd,
1980; Burrows, 1987 & 1996; Biber and Finegan,
1992;
Craig, 1999a, b, & c; Hoover, 2003a, b, & c; Rudman,
2005). My earlier attempts
used correspondence analysis
to accommodate low frequency variables (words) in profiling
authorial/chronological/cross-register variations
in Dickens and Smollett (Tabata, 2005; 2007a/b; 2008).
Given the fact that most collocates of content words tend
to be low in frequency, my methodology based on correspondence
analysis would usefully be applied to a macroscopic
analysis of collocation of gentleman.
Table 2 100 strongest collocates of gentleman based on
MI-score
Results
Fig. 3 visualises interrelationships among 77 texts. Data
points (texts) closer to each other in the diagram tend to
have similar collocates in common. The greater the distance
between texts, the less they have in common. Fig.
4 indicates interrelationships among 378 collocates. The
proportion accounted for by Dimensions 1 and 2 is only
4.21 % and 3.03 %, respectively, of the total variance in
the data, indicating the relationships among the matrix of
77 rows by 378 columns are extremely complex.
Fig. 3 provides an interesting overview of similarity or
contrast between texts: the horizontal axis, the strongest
variance in the data set, differentiates Dickens versus the
eighteenth-and nineteenth-century authors. One seeming
anomaly as far as Dickens texts are concerned is the
position of 1851_CHE, A Child’s History of England
(1851), which fi nds itself between the eighteenth and the
nineteenth century text groups. This history book written
for children is considerably different in style from
other Dickensian works. Therefore it is not unexpected
for this piece to be found least Dickensian as indicated
by the vertical axis, a phenomenon in keeping with previous
multivariate studies based on other linguistic variables,
such as –ly adverbs (Tabata, 2005: 231) and partof-
speech distribution (Tabata, 2002: 173). In addition,
early Dickens texts are found in the lower half of the
Dickens cluster, again in consistent with my other works
(Tabata, 2008; 2009a; 2009b).
Fig. 3 Correspondence Analysis of the collocates of
gentleman (378 collocates across 77 texts)
Fig. 4 A galaxy of gentleman’s collocates: Word-map of
378 collocates
The vertical axis, furthermore, shows the eighteenthcentury
texts to wards the bottom and the nineteenth centhe two sets are not in two distinct clusters. Fig. 4 corresponds
to Fig. 3 and thus tells us what words tend to cooccur
with gentleman in Dickens, in the eighteenth- and
nineteenth-century texts.
Concordance
A close examination of Fig. 4 leads to an awareness that
gentleman in Dickens are characterised
by (uncommon)
adjectives while, in the eighteenth-century texts, verbs
(of past tense) are prominent collocates. The nineteenthcentury
texts do not display a particular pattern apart
from words related to family or position. I would rather
interpret this result as suggesting the nineteenth-century
texts are negatively characterised both against the Dickens
set and the eighteenth-century set.
Fig. 5 Concordance: egotistical
Fig. 6 Concordance: censorious
A close study of gentleman in collocation with such
Dickensian adjectives, egotistical, censorious,
throwingoff
makes us realize Dicken’s ironical use of gentleman.
Moreover, these concordance lines make us realize that
in Dickens more than one modifiers are likely to precede
gentleman. In fact, Fig. 8 illustrates Dickens’s tendency
to use multiple adjectives. The increase in the proportion
which Dickens instances occupy in ‘the multi-adjective
gentleman’ from ‘a | an ADJ ADJ gentleman’, as well as
gentleman in total indicates that Dickens uses ‘the ADJ
ADJ gentleman’ formula as character appellations, instead
of simply referring to a character as a gentleman.
Fig. 7 Concordance: throwing-off
Fig. 8 Proportion of instances accounted for by Dickens
Fig. 9 Concordance: the X X gentleman
Fig. 10 Concordance: gentleman of X X
Conclusion
What has emerged from this survey can be summarised
as follows:
1. The most powerful solution obtained from multivariate
analysis can be interpreted as demonstrating
that Dickens has distinctive style in collocation of
the word gentleman.
2. Dickens is more likely to use adjectives in collocation
with gentleman than the control set. Adjectives
with higher MI scores often strike oxymoronic humour
when collocating with gentleman in Dickens
texts.
3. In Dickens, modifier-collocates (adjectives) tend to occur in succession (juxtaposition or concatenation)
typically in early works. They are often employed as
character appellation (variations of character-name) with a negative semantic prosody.
References
Adolphs, S. and R. A. Carter (2003) ‘Corpus stylistics:
point of view and semantic prosodies in To The Lighthouse’,
Poetica, 58: 7–20.
Church K. W. and Hunks P. (1990) “Word Association
Norms, Mutual Information and Lexicography,” Computational
Linguistics, 16/1: 22-29.
Firth, J. R. (1935). The Technique of Semantics. Transactions
of the Philological Society, 36– 72 (Reprinted
in Firth (1957) Papers in Linguistics. London: Oxford
University Press, 7–33).
Firth, J. R. (1957) ‘Modes of Meaning’, in Papers in Linguistics
1934–51. London: OUP. 191-215.
Greenbaum, S. (1970) Verb-Intensifier Collocations in
English: An experimental approach. The Hague: Mouton.
Henry. A. and R. L. Rooseberry (2001). Using a small
corpus to obtain data for teaching a genrein M. Ghadessy,
A. Henry and R. L. Roseberry (eds.) Small Corpus and
ELT. Amsterdam/Philadelphia, Pa.: John Benjamins.
93–133.
Hori, M. (2004) Investigating Dickens’ Style: A Collocational
Analysis. New York: Palgrave Macmillan.
Hunston, S. and G. Framcis (1999) Pattern Grammar:
A Corpus-Driven Approach to the Lexical Grammar
of
English. Amsterdam: John Benjamins.
Jones, S. and J. Sinclair (1974) “English Lecical Collocation”,
Cahiers de Lexicologie, 24: 15–61.
Kjellmer, G. (1994) A Dictionary of English Collocations:
Based on the Brown Corpus (3 vols.). Oxford:
Clarendon Press.
Louw, W. (1993) ‘Irony in the text or insincerity in the
writer? The diagnostic potential of semantic prosodies’,
reprinted in G. Sampson and D. McCarthy (eds.) (2004)
Corpus Linguistics: Readings in a Widening Discipline.
London and New York: Continuum. 229–241.
Partington, A. (2003) The Linguistics of Political Argument:
The Spin-doctor and the Wolf-pack at the White
House. London/New York: Routledge.
Partington, A. (2006) The Linguistics of Laughter: A
Corpus-Assisted Study of Laughter-Talk. Lon-don/New
York: Routledge.
Sinclair, J. (1991) Corpus, Concordance, Collocation.
Oxford: OUP.
Sinclair, J. M., Jones, S. and R. Daley (2004) English
Collocation Studies: The OSTI Report. Continuum.
Stubbs, M. (1995) ‘Corpus evidence for norms of lexical
collocation’, in G. Cook and B. Seidlhofer (eds.)
Principle and Practice in Applied Linguistics: Studies in
Honour of H. G. Widdowson. Oxford: OUP. 243–256.
Stubbs, M. (2001) Words and Phrases: Corpus studies of
lexical semantics. Oxford: Blackwell.
Tabata, T. (1995) ‘Narrative Style and the Frequencies
of Very Common Words: A Corpus-based Approach
to
Dickens’s First-person and Third-person Narratives’,
English Corpus Studies, 2: 91–109.
Tabata, T. (2002) ‘Investigating Stylistic Variation in
Dickens through Correspondence Analysis of Word-
Class Distribution’, in T. Saito, J. Nakamura and S.
Yamasaki (eds.) English Corpus Linguistics
in Japan.
Amsterdam: Rodopi. 165–182.
Tabata, T. (2004) ‘Differentiation of Idiolects in Fictional
Discourse: A Stylo-Statistical Approach to Dickens’s
Artistry’, in R. Hiltunen and S. Watanabe (eds.)
Approaches to Style and Discourse in English. Osaka:
Osaka University Press. 79–106.
Tabata, T. (2005) ‘Profiling stylistic variations in Dickens
and Smollett through correspondence analysis of
low frequency words’, ACH/ALLC 2005 Conference Abstracts,
Humanities Computing and Media Centre, University
of Victoria, Canada, 229–232.
Tabata, T. (2008) Gentleman in Dickens: A multivariate
stylometric approach to its collocation, Digital Humanities
2008, Book of Abstracts, The 20th Joint International
Conference of the Association for Literary and Linguistic
Computing and the Association for Computers
and the Humanities, University of Oulu, Finland, June
24–June 29, 2008, 199–202.
Tabata, T. (2009a) ‘“Wickedly, Falsely, Traitorously,
and otherwise Evil-adverbiously, Revealing” the Author’s
Style: Correspondence Analysis of –ly adverbs
in Dickens and Smollett’, Stylistic Studies of Literature:
In Honour of Professor Hiroyuki Ito. Bern: Peter Lang.
113–134.Tabata, T. (2009b) ‘“The Cunningest, Rummest, Superlativest
Old Fox”: A multivariate approach to superlatives
in Dickens and Smollett’, English Philology and
Corpus Studies: A Festschrift in Honor of Professor Mitsunori
Imai. Tokyo: Eihosha. 225–240.
Watanabe, S. (2009) ‘Gentleman in Oliver Twist: A Linguistic
Approach to Literature’, English Philology
and
Corpus Studies: A Festschrift in Honor of Professor Mitsunori
Imai. Tokyo: Eihosha. 273–286.