Why Our Minds Extend Further Than We Think — Epoche B2
Otto's notebook shows that a belief can sit outside the head, not that thinking does Why ask about Otto again Andy Clark and David Chalmers's 'The Extended Mind' (1998 [1] ) has one famous example and one famous principle, and they are usually taken to support each other. The example is Otto, who has Alzheimer's disease and writes what he needs to remember in a notebook, consulting it as reliably as anyone else consults their memory. The principle is parity: if part of the world does something which, done in the head, we would count without hesitation as part of a cognitive process, then that part of the world is part of the cognitive process. This column asks a single question about the pair. Does the notebook case, argued in the way Clark and Chalmers argue it [2] , show that any process of thinking extends beyond the skull? My answer is that it does not. It shows at most that a standing state — a belief one has while asleep, unconsulted — can be realised outside the body. That is a real and surprising conclusion, but it is not the conclusion the paper is remembered for, and the parity principle cannot carry us from the one to the other. The claim is contestable, and Clark contests it. The internalist starting point, stated fairly The view being challenged is that cognition happens inside the organism. On this view the brain, or the brain and body, does the thinking; everything else is input, output or apparatus. Notebooks, calculators and diagrams help, in the way that spectacles help seeing, without being parts of the seeing. It is worth noticing why this is not a silly view. It comes with a natural explanation of where to draw the line. Cognitive processes are the ones implemented by the machinery whose organisation explains the systematic patterns psychology studies — the ones that show recency effects, that interfere with each other under load, that degrade in characteristic ways under damage. That is not a boundary drawn at the skin for its own sake; it is a boundary drawn where the explanatory generalisations happen to hold. This is Robert Rupert's objection [3] , and it is the sharpest version of the internalist case: a notebook exhibits none of the effects on which the psychology of memory is built, so counting it as memory costs the generalisations that made memory a scientific kind in the first place. Anyone who wants to move the boundary has to say why the generalisations should be redrawn too. The parity principle and what it compares Clark and Chalmers's principle is stated in one sentence that is worth having in front of us, because everything turns on what it compares with what: If, as we confront some task, a part of the world functions as a process which, were it done in the head, we would have no hesitation in recognizing as part of the cognitive process, then that part of the world is (so we claim) part of the cognitive process. The comparison is between an external process and an internal process . Its force comes from a charge of prejudice: if we would count the internal doing of a thing as cognitive, then refusing to count the external doing of the same thing is discrimination by location rather than by function. Their opening illustration fits this exactly. Three people have to decide whether a shape fits a socket. One rotates the shape in imagination; one rotates it on a screen with a button; one, in some future case, rotates it with a neural implant. The rotating is a process, it is the same process, and it is hard to see what principle counts the first as thinking and the second as not. Where the notebook is a different kind of case Now put Otto beside that. Inga hears about an exhibition and walks to the museum on Fifty-third Street, because she believes it is there. Otto hears about the exhibition, consults his notebook, and walks to Fifty-third Street. Clark and Chalmers argue that the notebook plays for Otto the role that memory plays for Inga, and conclude that Otto believed the museum was on Fifty-third Street before he opened the book. Notice what the argument has done. It is no longer comparing an external process with an internal process. It is comparing the notebook with Inga's dispositional belief — the belief she had while she was asleep, before anyone asked. A dispositional belief is not a process. It is a standing state, and it does nothing at all until something calls on it. So the parity principle, which asks whether an external process does what an internal process would do, is not the principle being applied. The principle actually applied is a different one: if an external state plays the role an internal standing state would play, it is a state of the same kind. That substitution is what the conditions Clark and Chalmers attach are for. They require that the notebook be constantly available, that Otto endorse what he finds in it more or less automatically, and that the information have been consciously endorsed at some past time. These are conditions on standing states . They say nothing about processing, and no analogous set of conditions is offered for the rotation case, because none is needed there. So there are two arguments in the paper, not one, and they reach two conclusions. The rotation argument shows that cognitive processing can be spread across brain and world. The notebook argument shows that a standing belief can be realised in an external medium. Neither yields the other. A person could accept the first and deny the second, holding that processing spreads but that belief requires an integrated internal store; or accept the second and deny the first. They are usually presented as one thesis with one example, and that is the confusion this column is aimed at. The critique that misses, and why Fred Adams and Kenneth Aizawa's The Bounds of Cognition (2008 [4] ) is the standard opposition, and it is worth seeing why their argument does not reach the point just made. They charge extended-mind theorists with the coupling-constituti