Dr Marcus, I will admit that I have not read them all your writings - what I have read has not inspired me to read more so I may have missed where you addressed my concerns.
This is the core of my problems with what I have read - the brain does not seem to work like you are describing. You seem to base your explanations as if the brain processes things like we see them out here in sensory space.
Objects, relations between objects, words, letters, sounds, sequences of sounds. The brain sees this as clouds of impulses with some time and space correlation.
The very first thing it does is extract additional features out of time and space correlations, with more complicated comparisons as you ascend the sensory hierarchy.
The information is roughly topological throughout these hierarchies and never do combine into a form you may call objects. Further, the areas where this information is processed has all the other information mixed together in a wicked palimpsest.
The closest this ever get to an object is clouds of features extracted and spread throughout the areas around the temporal lobe. The only connection seems to be that the bits happened with coincidence.
It is as though the information was processed in the compressed zip format. It’s actually has some of the features of a zip file as the data is decomposed into features distributed over each area/map with connections between these maps/areas.
Trying to fit “operations on variables” to feature clouds and coincidence detections is a real stretch. Since the features are distributed the operations on these features is likewise distribute. The “operations” are still coincide detection in space.time.
This soup of digested features is fed to the much older subcortical structures and they do what they have always done for the last 400 million years – they decide on some action.
This is passed to the frontal lobe and it expands these to actions, external or internal, that are then perceived and start the process again in a continuous loop. This is what I call my dumb boss/smart advisor model.
So – unless I missed it – I see you as barking up the wrong tree with “symbolic” operations.
