Embodied cognition and abstract concepts: Do concept empiricists leave anything out?

ABSTRACT

According to the embodied cognition hypothesis, the mental symbols used for higher cognitive reasoning, such as the making of deductive and inductive inferences, both originate and reside in our sensory-motor-introspective and emotional systems. The main objection to this view is that it cannot explain concepts that are, by definition, detached from perception and action, i.e., abstract concepts such as TRUTH or DEMOCRACY. This objection is usually merely taken for granted and has yet to be spelled out in detail. In this paper, I distinguish three different versions of this objection (one semantic and two epistemic versions). Once these distinctions are in place, we can begin to see the solutions offered in the literature in a new, more positive, light.

KEYWORDS:

• Concept empiricism
• abstract concepts
• scope objection
• embodied cognition
• situated cognition
• concepts

Acknowledgements

I thank Dimitri Mollo, Juan R. Loaiza, Markus Werning and the participants of the 2017 workshop “Getting Real on Words and Numbers” for their helpful comments.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. I use capital letters to denote concepts, single quotation marks to denote words, and italics to denote properties.

2. This is what Dove (2016) calls disembodiment. Abstractedness lies on the same continuum as concreteness.

3. Both ways of defining the notion of an abstract concept in psychology are not ideal, but since it is the way the notion is defined in the relevant scientific literatures, I stick to this use for this paper’s purpose. Note that nothing hinges on the question of whether this definition is adequate here.

4. Frege also used the term mode of determination (Art des Gegebenseins), which is supposed to be mind-independent. However, in many contemporary theories of concepts, Frege’s notion of sense is mentalized. Nothing here hinges on this distinction.

5. Note that a categorization device could, in principle, be composed of other mechanisms besides beliefs as long as this explains how we in fact classify our environment.

6. Note that some authors (e.g., Prinz, 2002) give a unified account of all three notions. However, this is not necessarily advantageous, as I show in the following.

7. For example, a blind person uses very different descriptions or categorization devices to recognize something as a tree than a sighted person.

8. See Malt (1994).

9. Semantic definitionism has become unpopular mostly due to Quine’s (1951) arguments against the analytic–synthetic distinction and the lack of examples of successful definitions (Fodor, 1998; Prinz, 2002). But see Peacocke (1992) for such an account.

10. Causal-historical accounts of concepts have been applied to many concepts that are abstract, including natural kind terms like ‘atom’ (Millikan, 1998; Putnam, 1975) but also social kind terms like ‘race’ or ‘sex’ (Hacking, 1999; Haslanger, 2005; Mallon, 2015). Such accounts are yet to be fully developed, but they are a highly promising alternative to internalist semantics and there is so far no reason to rule them out.

Additional information

Funding

This work was supported by the Deutsche Forschungsgemeinschaft: [Grant Number GRK-2185/1].

Notes on contributors

Guido Löhr

Guido Löhr is at the Ruhr University Bochum and the Institut Jean Nicod at ENS Paris. He has recent publications on concepts and experimental philosophy in Synthese and Philosophical Psychology.