Current Research
Representation:
I have long investigated alternatives to language-like schemes of mental representation. It is motivated by the idea that perception and cognition require representations that are subject to meaningful structural transformations--isomorphs, essentially, of representationl targets rather than terms that refer to them. The work is constrained also by the idea that representation in the brain is primarily in the form of activation patterns (short term) and patterns of synaptic weights (long term).
Representation, as I am construing it, should be distinguished from indication. Indication--the sort of thing…
Current Research
Representation:
I have long investigated alternatives to language-like schemes of mental representation. It is motivated by the idea that perception and cognition require representations that are subject to meaningful structural transformations--isomorphs, essentially, of representationl targets rather than terms that refer to them. The work is constrained also by the idea that representation in the brain is primarily in the form of activation patterns (short term) and patterns of synaptic weights (long term).
Representation, as I am construing it, should be distinguished from indication. Indication--the sort of thing "idiot lights" do in your car--is arbitrary, transitive and source dependent. The information is all in who shouts, not in what is shouted. Indicator signals are therefore not subject to meaningful structural transformations. They say, "my target is in my receptive field" but say nothing about what the target is like. Representations re-present structural features of their targets, but say nothing about whether they are present or absent.
Functions:
After a long absence from the functions literature ("Functional Analysis" was written in 1968 and published in 1975), I have returned to a consideration of the relationship between function and evolution.
Learning, Development and Evolution:
The emerging field of evolutionary robotics provides a promissing venue for investigating the puzzling relationship between selection, development and learning, calling into question the traditional distinction between development and learning, and suggesting some promising alternatives to the doctrines of massive modularity and nativism that have dominated cognitive evolutionary psychology to date.
Recently, I have become convinced that the distinction between learning and growth is problematic. To learn, the brain must be altered. In particular, synapses must change. This requires protein synthesis, which, on turn, requires gene expression. Gene expression is environmentally triggered. When the triggers are very local, we call this growth. When, as in the case of the brain, the triggers can be distal, e.g., because of sense organs, we call it learning, especially if we can model the changes as "inference". At bottom, the processes are the same.