This brings me to the chief difference between these two examples [Kanzi and Rush Hour] that I want to draw explicit attention to. This is that whereas the first involves the causal integration of explicit symbols located in an organism's environment into that organism's cognitive regime, the second appeals to the cognitive incorporation of nonsymbolic aspects of that environment. Much of the discussion of the extended mind has focused exclusively on cases of just the former kind. (Wilson, 2010, p. 181).I think there is something to what Wilson is saying here. Clark, it seems to me, focuses on instances of being causally integrated to informational resources as cases of EC, where others, such as Wilson, it seems to me, have been happy to let causal integration with any sort of thing count as instances of EC. So, there is (another) division among those who advance EC.
(That said, there does seem to me to be some room for saying that those who play Rush Hour by moving the pieces around as aids are using the pieces as an informational resource, even if the pieces themselves are not representations.)