Language change and social networks

Jinyun Ke

Language Engineering Laboratory, City University of Hong Kong

jyke@ee.cityu.edu.hk

Sociolinguistic studies have shown that social networks play an
important role in stratification of language use, and language
maintenance and shift (Milroy, 1980; de Bot and Stoessel, 2002). 
However, these empirical studies only focus on small scale social
networks of local communities, and examine the situation over a short
time span.  The effect of different topologies of social networks on
language change in the long run has been little discussed.  While it
seems difficult or even impossible to deal with large scale social
networks over a long period of time for empirical studies, simulation
models serve as an effective alternative methodology. 

In recent years, there has been a dramatic growth of interest in complex
large-scale networks in various areas (Watts, 1999, 2003; Barab¨¢si,
2002; Buchanan, 2002), triggered by the discovery of several new types
of networks, such as small-world networks (Watts and Strogatz, 1998),
and scale-free networks (Barab¨¢si and Albert, 1999).  It has been found
that many complex networks in reality demonstrate several
self-organizing characteristics which can not be captured by random or
regular networks.  Social networks, such as friendship and cooperation,
are found to exhibit similar characteristics, which suggests linguistic
social networks may be modeled adequately by scale-free or small-world
networks. 

In this study, we argue that it is important to take into account social
networks in simulation models of language change, and to choose an
appropriate type of social structure closer to reality.  We compare
different types of social networks regarding their effect on language
change, based on a diffusion-through-learning model proposed by Nettle
(1999a).  In Nettle¡¯s model the social structure of the population can
be considered as a regular network, which is very unlikely to be true in
real situations.  Our simulation examines more realistic social
structures, and finds that the dynamics of language change in
populations of small-world or scale-free networks are closer to
empirical data than regular or random networks.  Nettle (1999b) argues
that the rate of language change is unlikely to be constant in
populations of different sizes.  We re-examine this argument using our
simulation model and suggest that a constant rate of change is possible
under some conditions.  Furthermore, we use the model to study how a
regular language change is possible if the change progresses in a
lexical diffusion manner. 

In the current network models of general interest, the individual nodes
are often assumed to have the same internal properties, except for their
different connections and different states.  We propose that it is
necessary to consider the heterogeneity of the individuals in the
language community.  The data from an empirical study of an on-going
sound change in Cantonese show that individuals exhibit a large degree
of variation in their language behavior which can not be explained by
idiosyncratic linguistic experiences, but rather requires individuals to
have different learning styles.  We hypothesize that there are at least
two types of learning styles, that is, some individuals learn to use
both competing variants of a change, while others learn only one variant
without accommodating the coexistence of the two.  The empirical data
show that the former type of learner (called a ¡°probabilistic
learner¡±) is much more frequent than the latter (called a ¡°categorical
learner¡±), and the simulation further suggests that the existence of
the probabilistic learners provides a much higher chance for language
change to complete than what is suggested by Nettle¡¯s (1999a) model in
which only categorical learners are assumed. 
 
Barab¨¢si, Albert-Laszlo.  Linked: the New Science of Networks. 
Cambridge, Mass.  : Perseus Pub., 2002. 

Barab¨¢si, Albert-L¨¢szl¨® and R¨¦ka Albert.  Emergence of scaling in
random networks Science 286, 509-512, 1999. 

Buchanan, Mark.  Nexus : Small Worlds and the Groundbreaking Science of
Networks.  New York ; London : W.W.  Norton, 2002. 

de Bot, Kees and Saskia Stoessel.  Special Issue on Language Change and
Social Networks, International Journal of the Sociology of Language, no. 
153, 2002. 

Milroy, Lesley.  Language and Social Networks.  Oxford: B.  Blackwell,
1980/1987. 

Nettle, D.  1999a.  Using social impact theory to simulate language
change, Lingua, 108:95-117. 

Nettle, D.  1999b.  Is the rate of linguistic change constant? Lingua,
108:119-136. 

Watts, Duncan J.  and S.  H.  Strogatz.  Collective dynamics of
small-world networks.  Nature, 393:440-442, 1998. 

Watts, Duncan J.  Small Worlds : the Dynamics of Networks Between Order
and Randomness.  Princeton, N.J.  : Princeton University Press, 1999. 

Watts, Duncan J.  Six Degrees : the Science of a Connected Age.  New
York: W.W.  Norton, 2003.