Language as a complex adaptive system

ASLAN’s scientific program for the project’s prolongation continues within a framework of language as a complex adaptive system. ASLAN accounts for linguistic phenomena in all of their complexity, within an integrative, multidimensional and non-reductionist approach. All levels of linguistic analysis are covered, ranging from the processing of phonetic information in the brain to mutual adjustment of linguistic forms in everyday conversation. Using language necessitates both cognitive processes and social abilities and is characterized by a large diversity in terms of the vast number of languages as well as the differing contexts in which language is used.

In the face of such diversity and ambition, how does ASLAN organize its research program? We are inspired by a leading figure in linguistic anthropology, Levinson (2005), who suggests that in order to build up a theory about a complex system, one needs to study subsystems independently, while hypothesizing about the relations between separate subsystems (1). This is very similar to what Greeno (1998) suggests in educational psychology where the system includes the context in which an individual’s cognition and behavior is studied (2). And In the heyday of artificial intelligence, Clarkson & Simon (1960) used simulations to reproduce part or all of the output of a behaving system where the system was an aggregate of units, each with a behavior (3). In Levinson’s words:
“The model suggested is one of three distinct levels of analysis, or three different kinds of systems — sociocultural systems, interaction systems, and language systems — interlocked in various ways. One doesn't have to be a realist about these entities — one can treat them as analytical fictions, whereby one gets a better model of the whole shebang by finding relatively differentiated subsystems which seem to have organizing principles of their own” (Levinson, 2005, p. 449).

ASLAN therefore promotes an open thematic structure in which different subsystems (e.g. cognitive, neural, biological, linguistic, social, interactional…) can be studied separately, but in which hypotheses relating subsystems (e.g. grammatical, semiotic, conceptual…) are encouraged. The interest of this approach is that it fosters collaboration between researchers from different traditions because systems and subsystems are viewed as boundary objects (4) that can be instantiated according to the epistemological assumptions and worldview of each collaborator. If researchers are open to accepting the definitions and contours of a system or subsystem, as viewed by a colleague from another tradition, then this sets the stage for shared methodological approaches and productive comparisons of analytic constructs, either about a particular (sub)system or regarding relations between them. Such an approach aims for broader scientific understanding through the collaborative work of researchers from language sciences, psychology, education, and computer science. But it also specifically fosters this collaboration through an empirical understanding of how scientific teams can work productively together (5).

(1) Levinson, S.C. (2005). Living with Manny’s Dangerous Idea. Discourse Studies. 7, 431-453.
(2) Greeno, J. G. (1998). The situativity of knowing, learning, and research. American Psychologist, 53, 5-17.
(3) Clarkson, G.P.E., & Simon, H.A. (1960). Simulation of individual and group behavior. American Economic Review, 50, 920-930.
(4) Star, S. L. & Griesemer, J. R. (1989). Institional ecology, “translation” and boundary objects: Amateurs and professionals in Berkley’s Museum of Vertebrate zoology, 1907-1939. Soc. Stud. of Science, 19 (387-42).
(5) Suthers, D. D., Lund, K., Rosé, C. P., Teplovs, C. & Law, N. (Eds.). (2013). Productive Multivocality in the Analysis of Group Interactions. In C. Hoadley & N. Miyake (Series Eds.), Computer Supported Collaborative Learning Series: Vol. 15. New York: Springer.