Haugeland, for instance, does not claim that since the road to San Jose is tightly coupled to him it becomes part of him. Rather, he claims that the he and the road together make up a coupled cognitive system. Schematically, here is Haugeland’s actual claim:
X is coupled to Y.So, X and Y make up Z.The idea is that when two things are tightly coupled, they constitute some third thing, of which they are both parts. A&A do not seem to recognize that.
I'm not sure that things are so simple as the when two things are tightly coupled, they constitute some third kind of thing. Does a piece of chewing gum stuck to my shoe make some third kind of thing? Does a cannon ball lodged in a castle wall make some third kind of thing? (Bear in mind that for Haugeland an HBI is just interacts intensely, as do the parts of a camshaft. So, just have a case where the two parts interact intensely, as gum stuck to a shoe.)
But, that aside, I think A&A are on to this idea of cognitive systems. We talk about this in section 7.2, "Haugeland’s Theory of Systems and the Coupling of Components," in The Bounds of Cognition. We draw a distinction between the hypothesis of extended cognitive systems and the hypothesis of extended cognition. We are willing to accept that there can be cognitive systems that extend beyond the boundary of the brain and body, but we resist the idea that cognitive processes extend beyond the boundary of the brain and body. How is this possible? We think that a type of process can be used to label a type of system without having that process pervade the whole of the system. Think of computing. That is a type of process that enables one to speak of a computing system, but the computational process does not pervade the whole of the computing system. So, we think that coupling might give you extended cognitive systems, but not extended cognitive processes.