Embodied Bounded Rationality

There is little doubt that one of Simon’s key contributions throughout his scientific career - if not the main one - was rooting the notion of bounded rationality in cognitive psychology (Simon, 1976). He called his notion of bounded rationality in the cognitive psychology realm ‘cognitivism’ (Haugeland, 1978), an approach, also known as the ‘information-processing’ approach, which he contributed to affirming together with his colleague Allen Newell starting in the mid-1950s. According to cognitivism, cognition works through the internal (i.e. mental) manipulation of representations of the external environment accomplished through referential ‘symbols’ (e.g., Newell & Simon, 1972). Connecting Simon’s theory of cognition to his theory of rationality is the notion that cognition works in a way that is necessary and sufficient for intelligent behavior (what is known as the ‘physical symbol system hypothesis’, see Newell & Simon, 1976). The idea that results from integrating Simon’s view of cognition with his view of bounded rationality is that rationality is a “process and product of thought” (Simon, 1978) in which the internal bounds of reason (Simon, 1955) adapt to the external bounds of the environment (Simon, 1956) in a disembodied fashion. In this picture, in fact, there is no place for flesh and blood. Patokorpi (2008) has emphasized that there is an inner tension in Simon’s thought, as on the one hand he was a strong advocate for a realistic approach to the bounds of rationality while, on the other hand, he represented them through the ‘unbounded’ power of digital computation and the metaphor of computers. This inner tension, which did not simply concern Simon’s thought but an entire generation of cognitive scientists, would have huge consequences in the history of cognitive psychology. In the early 1990s, Newell and Simon’s physical symbol system hypothesis was questioned when the ‘embodied robots’ designed by Rodney Brooks proved able to simulate simple forms of intelligent behavior by externalizing most of cognition onto the physical properties of environments, thus dispensing with abstract symbolic processing (Brooks, 1991). This is just one instance from the recent history of cognitive science pointing to the fact that while bounded rationality remains a pivotal notion in behavioral economics and economic psychology, new theoretical views and massive experimental evidence in cognitive science have led to supersede cognitivism and its abstract representation of cognition (Wallace et al., 2007). Contemporary cognitive psychology emphasizes that cognition is ‘embodied’, as it constitutively depends on body states, on the morphological traits of the human body, and on the sensory-motor system (see, e.g., Wilson, 2002). As such, it can be said to introduce another ‘bound’ to human cognition, able to integrate the internal bounds represented by cognitive limitations and the external bounds of task environments: the human body. In this paper, we argue that, in so far as the human body represents a new bound for human cognition, it can also have an important role in the re-conceptualization of bounded rationality. Since the new approach of embodied cognition is so recent, it is still rather plural and variegated, and as such far from a stable synthesis (for reviews on the issue of conceptual pluralism in embodied cognition see the classic Wilson, 2002; for more recent reviews, see Gallese & Lakoff, 2005; Clark, 2008; Kiverstein & Clark, 2009). Without the pretension to be exhaustive, here is a list of classic books on the idea that the body is a constitutive part of cognition: Varela, Thompson, & Rosch (1991); Clancey (1997); Clark, (1997); Lakoff & Johnson (1999); Rowlands (1999); Shapiro (2004); Gallagher (2005); Pfeifer & Bongard (2006). As a matter of terminology, although different labels have been used to identify and distinguish different views of embodiment, we will refer to them all by means of the common synthetic label ‘embodied cognition’ (Calvo & Gomila, 2008; Shapiro, 2014).