The case for protoconcepts
Joel Parthemore (Centre for Cognitive Semiotics)
Jerry Fodor is famous – or infamous – for the proposal that (all) concepts are innate, even if what he means by that is not always clear. John McDowell is likewise famous for proposing that concepts extend “all the way out” into the world and arguing against what he calls (borrowing a phrase from Wilfrid Sellars) The Myth of the Given: the idea that non-conceptual percepts can justify conceptual frameworks. In a now famous turn of phrase, he writes, “the idea of the Given offers exculpations when we wanted justifications.”
These two issues interrelate. Logically, it might seem – as it does to McDowell – that only concepts can beget concepts: a truly non-conceptual realm cannot serve as foundation to a conceptual one. Likewise it might seem – as it does to Fodor – that we are predisposed, from the beginning of our lives, to look at the world in certain ways and not in others. One need not go so far as either Fodor or McDowell, however, in allowing merit to their arguments. The world need not be fully conceptual to be never entirely free – for the conceptually minded agent – of conceptual taint. Conceptual and non-conceptual do not cleanly pull apart, either in experience (when our thoughts are directed “inward”) or in the world we experience (when they are directed “outward”). The world seems structured remarkably like our concepts because our concepts present it that way, and our concepts present it that way because of predispositions that are, in some substantive fashion, innate: in one way or another, they are with us from the beginning, and they carry us through to the end.
In keeping with the so-called animal-concepts philosophers – people like Albert Newen, Andreas Bartels, and Colin Allen – I hold that concepts, in the sense of systematically, productively, compositionally, and intentionally structured thought, are prior to (human) language, ontogenetically and phylogenetically. Rather than making conceptually structured thought possible, language transforms and extends conceptual abilities: allowing concepts to be more abstract, more richly compositional than ever before (at the same time, making them much easier to share). If so, any account of protolanguage requires an account of concepts to ground it; and any account of concepts requires – to take the right lesson, I believe, from Fodor and McDowell – an account of protoconcepts to ground that.
Protoconcepts are not concepts because they lack several of the usual desiderata of concepts: in particular, they are too few in number (I propose three) to be, of themselves, productive; and, being innate, they are not under the agent’s endogenous (active, intentional) control. Given the appropriate environment and interactions with that environment, a small set of protoconcepts can give rise to the most richly structured of conceptual frameworks.
I explain what protoconcepts are and why we need them: why, contra Vittorio Gallese and George Lakoff, sensorimotor engagement is not enough. I describe what those protoconcepts most likely are and outline the process by which they give rise – through sensorimotor engagement – to fully-fledged concepts. Finally, I try to ask the unaskable question: what would our conceptual frameworks and our world be like if our protoconcepts were different?