Love In Celestina

Lang. Soc. 13, 145-204.

Printed in the United Stales of America

Language style as audience design*
ALLAN BELL
Wellington, New Zealand
and

Department of Linguistic Science
University of Reading
ABSTRACT

The style dimension of language variation has not been adequately explained in sociolinguistic theory. Stylistic or intraspeaker variation derives
from and.mirrors interspeakervariation. Style is essentially speakers' response to their audience. In audience design, speakers accommodate primarily to their addressee. Third persons - auditors and overhearers - affect
style to a lesser but regular degree. Audience design also accounts for
bilingual or bidialectal code choices. Nonaudience factors like topic and
setting derive their effect by association with addresseetypes. These style
shifts are mainly responsive - caused by a situational change. Speakers can
also use style as initiative, to redefine the existing situation. Initiative style
is primarily referee design: divergence from the addressee and towards an
absent reference group. Referee design is especially prevalent in mass communication. (Sociolinguistic variation, code-switching, bilingualism,accommodation theory, ethnography of communication, mass communication)
In the quantitative study of language variation over the past fifteen years, style is
one dimension that has often been measured but seldom explained. Sociolinguists have commonly distinguished two initial categories of factors which correlate with linguistic variation (Figure 1). First are the linguistic factors - phonological,morphological, and syntactic constraints which promote or inhibit the
application of a variable rule. Thus, following consonant promotes consonant
cluster reduction - wes' side - and following vowel inhibits it - west end. Then
there are the extralinguistic factors, also divided into two categories. Since
Labov's (1966) pioneer study in New York City, these have been labelled the
"social"and "stylistic" axes of linguistic variation. The social dimension
denotes differences between the speech of different speakers, and the stylistic
denotes differences within the speech of a single speaker. The social axis has
been subjected to considerable examination, which convincingly shows that
linguistic variation correlates with variation in a speaker's class, gender, social
network, andso forth.
0047-4045/84/020145-60 $2.50 © 1984 Cambridge University Press

145

ALLAN BELL

Phonological

Syntactic

Class

Network

Age

FIGURE

Attention

Addressee

Topic

i: Some factors influencing language variation.

The other axis - the stylistic - has not been analysed so rigorously, however,
although Labov (19723:245) has maintained that "the most immediateproblem
to be solved in the attack on sociolinguistic structure is the quantification of the
dimension of style." Researchers have sometimes treated style as if it were an
independent, quantifiable variable - like age or class - which affects and explains linguistic variation (e.g., Wolfram & Fasold 1974:73). To say that an
individual speaker's language varies according to style, however, is tosay merely that a speaker's language varies - period. Language does not co-vary with
style, rather style is itself one axis of sociolinguistic variation. We must not
confuse the linguistic code with extralinguistic factors which may affect the
code. Just as the so-called social axis is correlated with certain extralinguistic
factors, so the "style" axis should be correlated with genuinelyindependent
variables.
The style dimension has been related to factors ranging from the narrow and
mechanistic to the broad and comprehensive. On the one hand, Labov (e.g.,
1972a) has proposed that style varies according to the amount of attention paid to
speech - which Brown and Levinson (1979:304) characterize as an "impoverished" view of style. At the other end of the scale are sprawling...