Strong CA Hypothesis

The strong CA hypothesis is that all cognitive processing is based around CAs.
I don't buy this.
However, a lot of the cognitive architecture stuff focuses on humans as symbol processors.
Roughly, all (or almost) all cognition is symbol processing.
That's why rules work so well for them.
They frequently (and quietly) discount all processing in the brain below this as sensing, or low level motor action.
I would argue that each symbol is represented by a CA.
However, I think a lot of early neural processing (like vision) is not CA based. None the less, line recognition (the first cortical stage of vision) does have influences from top down effects. So, a CA might influence which lines are activated in V1.
I wouldn't say that central pattern generators for muscle control are CAs.
So, I go with the broad range of a cognitive architecture representing all neural processing, but on a weaker CA hypothesis, that it is not everything cognitive.
I'm still thinking about it, and of course these are just hypotheses. The actual neuropsychological behaviour will tell us, and models can suggest the answer.