This is a nice statement of traditional cognitive psychology ideas. It is certainly worth reading as a short course.
But, it is a functional description and not a computational (formal) description of the things it talks about.
A factoid from NLP (natural language processing) AI has been that attempts to take such functional descriptions converting them into artificial systems has largely failed. NLP can give astounding performance (for example answering questions from descriptions of game play) without much reliance on such functional attributes.
Some of us believe that a simpler computational system will yield, if 'wired properly,' such functional behavior. Indeed, we would say that natural language and intelligence are inseparable functionally, but separable in computation. Communication like any human action or perception is hierarchically organized ephemeral serialization and deserialization while 'intelligence' is heterarchical and massively parallel.
Here is a more modern approach to the relationship between language and neocortical knowledge which looks at a computational description of how all this works:
We are still a very long way from understanding how any (sufficiently advanced) brain computes, let alone the human brain. But we are a lot closer than we were with the old standard model.
