I’m giving a talk at AECT in a couple of weeks on embodied cognition and education, as well as working on some related writings (and here are slides from my previous AERA talk on the subject). One related book I recently picked up is Embodied Cognition by Lawrence Shapiro. It’s a brand new book, and I’ve seen some positive reviews of it. It’s supposed to be a balanced perspective on embodied cognition research and theory.
But I flipped to the few pages on sensorimotor contingency theory (Noe, O’Regan), and Shapiro repeatedly says that a problem for the theory is that it can’t show that a brain in a vat doesn’t have sensory experiences (the “Argument from Envatment”).
I think even a 3 year old can tell you that a brain in a vat doesn’t have sensory experiences, no more than a head of lettuce.
This is a clear case of paradigm shifts. Shapiro is trying to talk about one paradigm from the perspective of another, older one (what he asserts is “standard cognitive science”). And according to Shapiro, it is the burden of the new paradigm to “distinguish itself” from the old one and “prove” itself. Take for example his assertion that “the burden that the sensorimotor theory of perception carries is to show that the brain alone is not constitutive of perceptual experience.” He most frequently cites work by Adams and Aizawa, who wrote a book critical of embodied cognition.
That’s not how paradigm shifts work (see Kuhn’s Structure of Scientific Revolutions from 1962). We shift to a new paradigm when the old one starts sounding ridiculous (brains in vats), or when the new paradigm is more useful or more parsimonious, or more consistent in its framework and so forth.
Shapiro’s book takes the traditional point of view on cognition, and of a computer-like, disembodied brain (he himself calls this “standard cognitive science”), and analyzes embodied cognition theories from that viewpoint.
He keeps using the term “knowledge”, for example, as something in the head that has nothing to do with action or physical experience or the environment. For example a fully paralyzed person is only capable of having “knowledge”, not actually “doing” anything embodied. I think paralyzed people can still try to do things (phantom limb, etc.), and they know how to do things (not to mention they can actually still do many things such as move their eyes and so forth). You can call their attempted actions a mental simulation if you like, but FMR studies show that mental simulation activates the same brain regions as the real actions.