This is an excellent article on the idea of extended cognition. There is one area which I think needs clarification. Joshua Sokol outlines a distinction between extended cognition, which Kevin Laland and others see as the equivalent of niche construction, on the one hand, and extended phenotype on the other. But it seems to me that while extended cognition and extended phenotype are compatible, indeed interdependent, the notion of niche construction, with its unavoidable teleology, is something quite different. So the disagreement is between niche construction on the one hand, and extended cognition and and extended phenotype on the other.
Implicit in the article is also a suggestion of a distinction between knowledge (conceived as “abstract” representations in the “mind”) and information, as when information is what a spider’s neuro-muscular system receives at an intersection in the web. Thus information is categorically differentiated from the spider having a total representation of a web in its brain, knowledge. But that in turn suggests the existence of an exclusive and inclusive, that is discrete, representation in a brain, as it might be in a human brain, of “breakfast” or “Roman Catholicism” or an internal combustion engine.
That does not seem to be how the brain works. My visual cortex will produce many images for an internal combustion engine, a black and white illustration of a certain shape and surface detail nearly all of which is made up of Dennet’s figment, or what I see when I lift the bonnet of our car, or, in a sequence of visual images, the throughput of a turbofan jet engine; but I don’t have a single, discrete holistic totality of an ICE in my head, merely an interacting community of reomes.
This suggests to me that the distinction between information and knowledge is fuzzy. And as to information exchange between we humans and our extended phenotype, it is clearly continual. Without it, we would be merely apes with a big expensive brain and nothing to do with it.