What are the defining differences between human and animal cognizers? This concern has driven philosophers and scientists for a long time,[1] well before Darwin’s (1871) theory of evolution and its inherent claim of developmental continuity between the species. The prevailing intuition has been, and often still is, that even though we stand in a direct developmental line with other mammals in a physiological sense, our cognitive and affective abilities far exceed theirs, not only in a quantitative, but also in a qualitative sense. Criteria to support this notion are frequently sought in an array of special cognitive abilities, such as the ability to speak (e.g., Savage-Rumbaugh et al. 1985), the possession of concepts (e.g., Newen & Bartels 2007), or behavioral traits like altruism or cooperation (Hamann et al. 2011; Warneken 2013; Warneken & Tomasello 2009). All of these are to varying degrees attributed to humans, but are either to a much lesser degree or not at all ascribed to animals, thus representing the cornerstones of the critical divide between “us” and “them” (Hare 2007). The problem raised by Davidson and discussed by Roskies concerns the special case of beliefs and the general case of the attribution of propositional attitudes to nonlinguistic creatures.
According to Davidson, it is only in the domain of human cognition that we can sensibly apply the notion of thinking. His reasons for holding this conviction are manifold, as Roskies uncovers beautifully in her treatment of Davidson. The general line of argument will be sketched out and discussed below. Roskies refutes Davidson’s arguments mainly on empirical grounds, with the aim of establishing that nonlinguistic animals can be cognitive agents with beliefs and mental representations, which function as kinds of propositional attitudes. In this commentary, I would like to complement this line of reasoning by questioning what it takes to credit human cognizers with thoughts; or rather, what we consider to be the prerequisites for attributing thoughts and beliefs to humans. Davidson puts much weight on the possession of language. Here, I want to argue that focusing on language as a necessary cognitive instrument for being able to think poses a methodological barrier for examining what the human ability to think actually amounts to. Stressing the point that the introspectively experienced properties of thinking, a term that requires careful consideration in itself, should not be identified with and reduced to experiencing inner speech, I want to show that our understanding of what thought is needs to be complemented by a bottom-up investigation into the neural processes and mechanisms that produce higher cognitive states, such as thoughts. I argue, therefore, that our introspective access to the way thinking presents itself to us as thinkers is only one part that needs to be considered. What is required in order to understand the phenomenon of thinking is first a suitable conceptual framework of the notions “thought” and “thinking”, which distinguishes between their intentional and phenomenological aspects, i.e., between the content of propositional attitudes and the phenomenal states of subjects making use of these attitudes. Second, we need to show how the sub-personal and personal levels of these factors can be distinguished from each other in order to show if and how they are interconnected. These considerations will be discussed in detail after a review of Roskies’s discussion of Davidson’s account of language and belief.