More often than not, great divides have been postulated between humans and other animals: it has variously been maintained that only humans have souls; that only humans laugh; that only humans play; that only humans are rational. The status of these claims is not merely of theoretical interest: human exceptionalism has long been used to justify or discount arbitrary and often inhumane treatment of animals, including the abuses perpetrated in factory farms and the devastation of habitats for human gain. While the issue of the soul is beyond empirical confirmation or disconfirmation, many other claims about the uniqueness of humans have been shown to be untrue or only half-true. Recently, in response to philosophical and empirical work, there has there been significant political pushback. For example, the Great Ape Project (http://www.projetogap.org.br/en/) aims to establish great apes as persons with recognized legal rights. Whether we should stand behind such a project or other less ambitious efforts to treat animals as entities with moral worth depends at least in part on what kind of capacities they have, both cognitive and affective.
Here I combat a philosophically prominent claim of human uniqueness: Donald Davidson’s famous argument that only humans can think. In the light of the complex cognitive activities of which animals are clearly capable, one might think this patently untrue. However, Davidson means by this not that animals have no cognitive capacities at all, but that nonhuman animals cannot have beliefs, desires, and other propositional attitudes. What the thesis does is set animal cognition apart from human cognition as a different natural kind, due to radically different representation schemes (see also arguments in Malcolm 1972). This is not a straw man, but an interesting and challenging thesis. In critically evaluating the arguments Davidson provides in light of empirical evidence from developmental psychology and ethology, insight can be gained into the nature of the relationship between thought and language. Despite its prima facie plausibility, I conclude that in light of contemporary studies from human and animal cognition, arguments for restricting propositional attitudes to humans fail.[1] The implications of this result could be far-reaching. Language as a cognitive ability has held a special status in analytic philosophy, where it is often assumed to be foundational to thought and cognition. Rethinking the role of language as a cognitive newcomer and possibly in large part a cognitive overlay resting atop a toolbox of already-powerful cognitive abilities may lead us to rethink a number of fundamental issues in philosophy, as well as to reconsider our cognitive and ethical relationship to the rest of the natural world. This critique of Davidson is illustrative of Dennett’s caution:
[p]hilosophy of psychology driven by the concerns of philosophy of language does not fall happily into place. (Dennett 1987b, p. 204)
The various arguments Davidson supplies for thinking that humans are unique in having propositional attitudes all rest upon the idea that having language is an enabling condition for having propositional attitudes. Since only humans have language, it follows that only humans have propositional attitudes.[2] Thus, his main argument against animal thought is:
P1 If something has propositional attitudes, then it has language.
P2 Animals don’t have language.
C Animals don’t have propositional attitudes.
The logic here is unassailable: if P1 and P2 can be established then the conclusion that animals lack propositional attitudes follows. For Davidson, having language is having the ability to speak (1984, p. 167, 2001a, p. 99), to express one’s thoughts, and to understand the speech and propositional attitudes of others. It is generally accepted that nonhuman animals don’t have this ability, despite some evidence that certain birds and higher mammals have some nontrivial linguistic abilities (Kaminski et al. 2004; Pepperberg 2000; Savage-Rumbaugh 1986). Therefore, we will grant P2.[3] The success of this argument denying propositional attitudes to nonhumans therefore rests on the ability to establish P1, namely the claim that having propositional attitudes requires language. In this paper I consider the various avenues by which Davidson tries to establish P1, for his arguments make contact with a broad range of research concerning mind and language, and serve as a good guide to attempts to link propositional attitudes to language. I begin by situating Davidson’s arguments in his larger theoretical context, and raise a few methodological worries about his approach. I then briefly consider some of his minor arguments, before turning to his strongest argument linking propositional attitudes to language. I argue that on the most plausible consistent readings of his arguments, one or another premise can be empirically falsified. I conclude by considering how to proceed to better understand the nature and limitations of animal thought.