The Architecture of Cognition: Rethinking Fodor and Pylyshyn's Systematicity Challenge

13. Neo-Empiricism and the Structure of Thoughts

Neo-empiricism is one of the most exciting theories of concepts developed in the last twenty years in philosophy and in psychology. According to this theory, the vehicles of tokened concepts are not different in kind from the vehicles of perceptual representations. Proponents of neo-empiricism have touted the virtues of their theory (Barsalou 1999, 2010; Prinz 2002; Gallese and Lakoff 2005), including its alleged empirical support, while critics have raised various concerns (e.g., Markman and Stilwell 2004; Machery 2006, 2007; Mahon and Caramazza 2008; Dove 2009, 2011; McCaffrey and Machery 2012). In this chapter, I will criticize neoempiricists ’ views about the nature of thoughts, that is, about those representations that express propositions. In substance, if neo-empiricism were right, our capacity to think—to move from thought to thought— would be either mysterious or a matter of a contingent history of learning. Fodor and Pylyshyn’s classic article against connectionism will be useful to develop this criticism, since neo-empiricists’ views about the nature of thoughts suffer from problems similar to those Fodor and Pylyshyn diagnosed twenty-five years ago.1 Here is how I will proceed. In section 1, I will review the central tenets of neo-empiricism. In section 2, I will review how neo-empiricists characterize occurrent and non-occurrent thoughts. In section 3, I will argue that amodal symbols seem to be needed to have thoughts in longterm memory (i.e., non-occurrent thoughts). In section 4, I will argue that neo-empiricists’ views about occurrent thoughts are unacceptable. 1 Neo-Empiricism Neo-empiricism in contemporary philosophy and psychology is not a reductionist semantic theory: it does not state that the content of thoughts can be reduced to perceptual properties such as red, square, loud, and so on. 13 Neo-Empiricism and the Structure of Thoughts Edouard Machery 336 Edouard Machery Prinz (2002) argues at length that the types of representations posited by neo-empiricists (in his terminology, “proxytypes,” or in Barsalou’s terminology , “perceptual symbols”) can be about three-dimensional objects and their physical properties. Nor is neo-empiricism a developmental theory about concepts: it does not state that all concepts are learned. Prinz (2002) clearly argues that neo-empiricism is consistent with the claim that some conceptual representations are innate, whereas Barsalou (1999) holds that some conceptual representations are in fact innate. Instead, neo-empiricism is a theory about the vehicles of concepts (Prinz 2002, 109; Machery 2007; McCaffrey and Machery 2012), that is, about the nonsemantic properties of the physical (presumably neural) states that realize concepts. Although neo-empiricists disagree about various points, all of them accept the following two theses about the vehicles of concepts: (1) The knowledge that is stored in a concept is encoded in several perceptual representational systems. (2) Conceptual processing involves essentially reenacting some perceptual states and manipulating these perceptual states. The first thesis describes what it is to token, or entertain, a concept: according to Thesis 1, to entertain a concept is to entertain some perceptual representations. Thus, to think about dogs consists in entertaining some perceptual (visual, auditory, olfactory, proprioceptive, etc.) representations of dogs. These perceptual representations are of the same kind as the perceptual representations one would entertain if one were to perceive dogs. Thus, thinking consists in reenacting, or simulating, the perception of the objects of one’s thoughts. The reenacted perceptual representations need not be conscious: in line with modern cognitive science, neo-empiricists hold that one can entertain a perceptual representation unconsciously. This first thesis denies the view held by amodal theorists that the vehicles of thoughts are distinct in kind from the vehicles of perceptual representations (Fodor 1975, 2008; Pylyshyn 1984). Neo-empiricists have been crystal clear in their embrace of Thesis 1. Barsalou writes (1999, 577–578): Once a perceptual state arises, a subset of it is extracted via selective attention and stored permanently in long-term memory. On later retrievals, this perceptual memory can function symbolically, standing for referents in the world, and entering into symbol manipulation. As collections of perceptual symbols develop, they constitute the representations that underlie cognition. Neo-Empiricism and the Structure of Thoughts 337 Similarly, Prinz proposes that “concepts are couched in representational codes that are specific to our perceptual systems” (2002, 119), and that “tokening a proxytype is generally tantamount to entering a perceptual state of the kind one would be in if one were to experience the thing it...