> This is just a commonly held belief among memory
> researchers that memory is essentially unlimited. Various models posit
> that the networks that store memory have an essentially unlimited
> capacity.
'Essentially' since it would be practically impossible to test! A project called
"Robokeno" [Robot child cat in Japanese] is being undertaken that attempts to
simulate the brain's many neurons and their thousands of dendritic connections with
other neurons (insanely enormous number of total connections). Personally, I feel
this project is outside the realm of human capability at this point. But the
article about this project can be found:
http://news.bbc.co.uk/hi/english/sci/tech/newsid_250000/250343.stm
> This is usually achieved because elemental representations are
> re-used in many episodic representations (or something to this effect).
> So whereas a computerized memory of you seeing a cat and then the next day
> seeing it again would have complete copies of the representation of the
> cat and yourself (and the room, etc.) which are unconnected, the brains'
> memory would use the same representation of the cat and yourself and room
> for both memories; these would be seperated by some additional time
> representation.
I don't know if this explanation addresses the space problem but I think that as
humans are extremely adaptable, if the need for more storage ever was introduced,
the body would deal with it.
> The only problem with this kind of storage is that it is subject to
> massive interference (of course, this isn't necessarily a problem... since
> it would lead to "spreading of activation" which isnt necessarily a bad
> thing).
Hence, the memory of the scent of a woman's perfume, the taste of a entree, wetness
of a cild's kiss, etc. can be triggered by other sensory stimuli originally present.
Peter
