Agent-based models of language competition

Abstract

This article uses computer simulations as a means of assessing two different models for the competition between two languages from the interdisciplinary perspective of complex systems. These models allow for the analysis of the role of bilingual speakers and they consider two of the basic factors determining the use of each language: their relative prestige and the language loyalty of their speakers. Here we assess the effect of these parameters in different social networks, with the aim of contributing to the understanding of the role of bilingualism and social networks in processes of language contact. We first present a qualitative analysis based on a simulation tool that we have designed, in order to visualize the dynamics of the models. Secondly, we consider a qualitative analysis of the role of complex networks, identifying mechanisms for the rapid extinction of a language and for situations of language segregation. Finally, we present a quantitative analysis which shows, counter-intuitively, a transition from a final scenario of language maintenance to one of language shift as language loyalty increases. Other findings are: (i) that the scenario of language coexistence is reduced in the presence of bilingual agents; and (ii) that networks which allow for local effects reduce further the situations for which language maintenance is possible.

Is the coexistence of Catalan and Spanish possible in Catalonia?

“…explicit flow between groups, see [17,30,31,33].) The parameter a (which has been referred to as volatility [21]) affects those speakers that promote language shift (termed attracting population [33,39]). It confers an idea of how persistent the linguistic groups are: the lower a the easier it is for all groups to lose speakers, thus rendering the system more volatile [32].…”

Modeling morphological learning, typology, and change: What can the neural sequence-to-sequence framework contribute?

“…[ 82 ] training data for another, observing how the system changes over time Hurford 1997, 2002). Such iterated learning experiments have been used to study the emergence and disappearance of irregular forms in both simulated (Ackerman and Malouf 2015;Parker et al 2019) and real (Hare and Elman 1995;Cotterell et al 2018b) datasets, and to study the spread and loss of different languages or linguistic features in a social network (Abrams and Strogatz 2003;Castelló et al 2013).…”

“…In this kind of model, users make both conscious and unconscious decisions about what language features to use (Milroy 2007). For instance, in a model of language change in Spain (Castelló et al 2013), agents in a social network speak either prestigious Castilian or stigmatized Galician. Speakers may switch languages in either direction, based on how many of their neighbors in the network they will be able to communicate with and how socially prestigious they will become.…”

Language policy and planning as an interdisciplinary field: towards a complexity approach

“…In this text, we start from a perspective of complexity that is more epistemological and theoretical in nature, applying this perspective to sociocultural phenomena. Our aim, however, is not meant to disregard the potential of other approaches, such as the study of so-called 'complex adaptive [or evolving] systems', which may contribute new advances to the sociocultural sciences by means of computer modelling and simulation (Castelló, Porto, & Miguel, 2013;Ellis & Larsen-Freeman, 2009;Helbing, 2012;Holland, 1995Holland, , 1998. 4.…”