This is a very useful article. Thanks, John. It partly answers a question I wrote in a comment on one of your other articles, about class/type hierarchies. However, I'm still wrestling with how to represent meanings usefully in a networked way like our brains do -- one that doesn't involve language explicitly but has language networks as an adjunct. For example, in your Figure 2 diagram, how would the links labeled "ACTOR" and "UNDERGOER" actually be represented? Would they be unidirectional or bidirectional? Would "eat" have other attributes? And what class hierarchy of actions would "eat" be part of?
I have been skeptical of Chomsky's approach since the 1960s and have found it very discouraging that so many linguists followed it. In fact, I wrote a paper critiquing that approach in 1967 and published it with my thesis advisor as coauthor. It appeared in the journal Language and Speech, "On the Problem of Describing the Grammar of Natural Languages, April 1967:
https://journals.sagepub.com/doi/abs/10.1177/002383096701000204?journalCode=lasa
Since then I have worked on building working models of human language behavior incorporating syntax, semantics, and pragmatics. To extend your argument that meaning, not syntax, is the basis of language, I believe that the use (i.e. pragmatics) of utterances in non-linguistic context is as important as the meaning (semantics). You can find more of my work at my York University web site. See
http://www.yorku.ca/jmason/asdindex.htm
and
