Cognitio 2011

Abstract: In everyday discourse, people tend to reason about inanimate objects and animals in a way that implies they possess human-like mental states. An intentional stance (Dennett, 1987) is a generative explanatory theory that individuals use to impute intentionality to objects and entities. Dennett’s basic proposal requires that we regard an entity or object as a rational goal-directed agent irrespective of whether it is so or not. This approach reduces the burden of constructing more complex theories based on physical forces (a physical stance) or function (a design stance). Consequently, when adopting an intentional stance, an anthropomorphic analogy is used to more readily encode or decode the relations of the parts within a system. An individual, however, must still reason about what actions will result from the corresponding beliefs.

Considerable evidence suggests that specialized, domain-specific cognitive processes are used for face recognition, folk biology, naïve physics and naïve psychology (Hirschfield & Gelman, 1994). It seems reasonable to assume that individuals might be more or less inclined to use this information based on the apparent animacy of the entity. For instance, people’s willingness to reason using an intentional stance is likely to differ when reasoning about atomic particles, migrating animals, or human groups. Although an understanding of naïve psychology might be available from an early age (e.g., Carpenter, Call, Tomasello, 2002; Song & Baillargeon, 2008), the early development of an animate-inanimate distinction (e.g., Bonatti, Frot, Zangl, & Mehler, 2002; Carey, 1985; Poulin-Dubois, 1999) might limit the use of the intentional stance for inanimate objects (cf. Csibra, 2008) or entities that are highly dissimilar to humans (e.g., microorganisms).

In the present study, we used a belief-bias task (Evans, Barston, & Pollard, 1983) to consider whether the degree of animacy of an entity places constraints on the willingness of an individual to attribute intentionality in the context of syllogistic reasoning. In this task, participants are presented with logical syllogisms that vary in their validity and their relationship to prior knowledge. Results demonstrate that valid statements are accepted more than invalid statements, believable statements are accepted more than unbelievable statements, and that believability has a greater effect on invalid syllogisms than valid syllogisms (Evans, 1989; Newstead & Evans 1993). In the present study, we presented our participants with descriptions of the behaviour of entities and participants were required to indicate whether the explanations that followed were valid or invalid. A set of entities was selected varying in terms of their animacy and the extent to which they had human-like properties. The categories included particles (e.g., atoms), microorganisms (e.g., viruses), insects (e.g., ants), reptiles (e.g., iguanas), mammals (e.g., mice) and humans (e.g., families). Importantly, explanations could use either mechanistic (e.g., particles attracted by a force) or anthropomorphic terms (e.g., particles like one another) that were irrelevant to the statements logical validity.

Our results support earlier findings that the believability of a statement and its logical validity interact. More importantly, we found that participants perceived entities along a cognitive continuum defined by the extent to which their mental states are related to humans. Regardless of the logical validity of a statement, however, participants generally believed that nonhuman animals had similar mental states to humans. Participants’ subjective confidence ratings also suggested a lack of awareness in the bias they maintain towards these statements. Thus, although people might adopt the intentional stance in casual conversation it seems that they are disinclined to adopt this perspective when reasoning formally about most nonhuman entities.