Pompe-Alama calls attention to the “feeling of what it is like to think”, which she identifies as the experience of our thoughts as inner speech. There is of course debate about whether it feels like anything at all to think. However, regardless of whether our recognition of inner speech is a feeling or a cognitive introspective conclusion, this phenomenon certainly plays a role in the general tendency to and perhaps our willingness to identify thought with language. But Pompe-Alama’s easy identification of the phenomenology of inner speech with Davidson’s denial of animal thought threatens to trivialize what I take to be a fairly sophisticated, if incorrect, view about the nature of animal thought. Davidson’s interpretationism is the root of his denial, and his target is specifically propositional thoughts and related attitudes, not cognitive processing more generally. Pompe-Alama cites Vygotsky’s claim that lots of thought is not verbal thought, and she suggests that pictorial or imagistic thought should be possible for non-linguistic creatures. I don’t suppose Davidson would refuse to recognize that animals have complex representations and even some relatively high-level cognitive capacities. But he would deny that these forms of thought had propositional contents. So the real question at issue is whether the representational power afforded by representations in nonlinguistic animals allows them to represent propositions.
That said, Pompe-Alama’s claim that the restriction of thought to verbal vehicles may be a “theory-induced illusion” is well taken. The tendency to think that only language-like formulations allow propositional content to be captured or delineated seems ungrounded, especially since philosophy has supplied us with non-linguistic means of representing propositions (Stalnaker 1987), or alternatives to propositional attitudes (Churchland 1992). Undoubtedly, propositional content requires some kind of framework that permits complex structural relationships between representations, but there is no a priori reason to think that such structure can only be achieved with linguistic implementation. Pompe-Alama is correct to point out that in our own interpretation of others, we often privilege behavior over self-report, and much social science has suggested that words, and indeed even one’s own introspective thoughts, are not a reliable window into higher cognitive processes. She also mentions that our own interpretational skills, applied to animals, yields attributions of cognitive processes that are in many ways akin to our own. Indeed, we easily attribute to them propositional attitudes. These observations put pressure on Davidson’s view, and raise the question of what our own propositional attitudes may endow us with, cognitively speaking, that the presumptively propositional-attitudeless animals are missing, if in fact he turns out to be right.
Pompe-Alama doubts whether language really plays a key role in human higher cognitive functions. We know it certainly does in one of them: Linguistic cognition. Whether it plays a fundamental role in other aspects of higher cognition is yet unknown. Davidson himself is not clear about whether he thinks language is necessary as a vehicle for thought. This distinguishes him from Fodor, who also thinks language is central to thought, but posits a mental language to serve as the vehicle of thought, and that is available to linguistic and non-linguistic creatures alike. Davidson’s view is more subtle, and seems to depend more on social/interpersonal factors and abilities or dispositions than on vehicles per se. Thus, for Davidson, the fact that we can identify instances of non-linguistic symbol use in high-level thought is not telling, since it is the fact that we are language-using creatures that is of prime importance. It is within Davidson’s purview to claim that our mastery of language makes possible thoughts that rely on non-linguistic (yet symbolic) properties.