Representing cultural change in agent-based models of landscape scale socio-ecological systems

Many agent-based models (ABMs) use typologies to classify diverse actors into few simplified conceptual categories with uniform decision-making strategies. This approach usually assumes that a representative agent can belong to only a single conceptual category at any one time. For example, Millington et al. (2008) used this approach in an ABM of an agricultural Mediterranean landscape, classifying agents as either ‘traditional’ or ‘commercial’ farmers.

            However, social psychological theory asserts that individual actors’ identities are constructed of multiple hierarchical self-concepts which motivate decisions and behaviour (Stryker and Burke 2000). For example, within their broader identity as a farmer, some actors may hold ‘agricultural producer’ and ‘agribusiness-person’ self-concepts as salient to their identity but reject ‘diversifier’ and ‘conservationist’ self-concepts (e.g., Burton and Wilson 2006). Other actors may hold all four self-concepts as salient to their farmer identity but prioritise some over others. Furthermore, identity may change slowly through time if an individual cannot exhibit behaviours that express their identity, for example because of land resource or other constraints. Identity may also change in response to others’ expressed identities within the individual’s social network (Stryker and Burke 2000). 

            To investigate the potential of this approach for agent-based models of landscape scale socio-ecological systems, we present an ABM that explicitly represents multiple and hierarchical agent self-concepts of actors (farmers) in an agricultural landscape. Using the model we explore dynamics of change in farmer self-concepts and agricultural land use for different macro-structural conditions, including spatial configurations of land resources. Initial results indicate that productivist farmer identities are stubborn to change, spatial distribution of land resources influences identity change, and rules for social network formation influence the likelihood of agents' identity change. We discuss the potential of approaches like this which incorporate representation of cultural change into agent-based models of landscape scale socio-ecological systems.