From: Emil Gilliam (emil@emilgilliam.com)
Date: Sat Nov 02 2002 - 22:15:12 MST
Quoting LOGI:
[Matching imagery against stored memories is going to be very
computationally expensive,
but...] One hopeful sign is the phenomenon of cognitive priming on
related concepts
[Meyer71], which suggests that humans, despite their parallelism,
are not using pure brute
force.
I would not be so hopeful here; the phenomenon might only be shifting
the moment in time when the human is doing the brute force.
The human might "cheat" by taking the test as follows: When she is
given the first word (the priming word), she immediately does the brute
force work of coming up with as many related words as possible, and
then keeps as many in her short-term memory as she can -- in the hopes
that some of these will show up later in the test.
Then, when a word comes up later on, if it's in her short-term memory
she recognizes the relationship immediately and thus her reaction time
is reduced. Like a computer's cache, if a word is not in her short-term
memory she must then resort to a hard lookup to determine whether there
is a relationship or not.
Obviously this caching is better than nothing as it reduces the number
of times brute force is involved, but brute force is still ultimately
involved at some point.
- Emil
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:41 MDT