Hi Jo, 

I noticed that you did not offer any information as to why the Marr analysis was evidence that the location of the synapse on the neuron was important (rather than strength of input for example). Which I found slightly strange considering you were considering it to be the "best clue" that that was the case.  

You wrote:

The cell's output will be one bit but that will be enough to encode two to the power 5,000 possible experiences because all it needs to mean is 'yes, that pattern you have stored over there that you are asking me if it is that one'. That is basically how communication works - by deixis or pointing to a referent. The cunning part of language is pointing in a virtually infinite number of ways by the combinatorial use of signs. Within a brain all that is needed is a sign for 'yes, what you just signalled fits'. 

I am not sure how you think a referent can be pointed in any communication. I am not suggesting it is not, I am just not sure how in your story you think it happens. I assume you could not explain it for a computer system. Anyway, I will ignore that for now.

So consider cell A which gives an output to cell B. 

1) Cell A receives inputs which describe one of possibly 5000 experiences, not simply through the value of the inputs but also their location on cell A. 

2) a "yes" or "no" message is then sent to cell B.

I am not clear how subject cell B has information as to which of the 5000 possibilities the "yes" or "no" from cell A refers to, but then I am assuming that it cannot be encoded into the synapse location as the "yes" and "no" to all 5000 possibilities arrives at the same location, and I am also assuming that the 5000 possibilities would not encoded any more efficiently for cell B than cell A, and I am assuming that there is no evolutionary advantage to having the same inputs as cell A plus the "yes" or "no" input that would result form those inputs anyway. So how are you suggesting subject cell B has information as to which of the 5000 possibilities the "yes" or "no" refers to such that its experience would involve the correct content?

Yours sincerely, 

Glenn