Bounded rationality, distributed cognition, and the computational modeling of complex systems

Computational modeling of complex systems dominates much of contemporary science. This chapter sketches a cognitive account of model-building in one computational field in particular: computational systems biology. This field aims to produce large-scale models of complex biological systems. Importantly, the field exemplifies the distributed nature of much contemporary computational modeling, which relies on coupled interactions between computers and human agents. The results of research into these practices show that modeling practices can be understood, and indeed rationalized, using the concept of bounded rationality and the procedures of problem search. However, accounting for these practices must go beyond purely “internal symbol system” accounts of cognition and recognize both the distributed nature of cognition and the role of simulative mental modeling. Both notions help provide clarity in this case as to why problem-solving is “bounded” and why the limited modeling strategies chosen by modelers as a result are “rational.”