Gosh, how many issues it raises!
To pick two:
>And maybe time and changes are so interweaved on the basis of our
>construction systems that we can't just separate them!
Yes. Actually, there used to be an active field of research on time in
the 1970s laboratory-experiment, information-processing tradition: I
haven't kept up with it since, but I can remember doing stuff on the
accuracy of peoples' perception of brief temporal intervals. Here's a
reference that would open up this particular strand of research if you're
interested; there were several information-processing models of one's
experience of temporal duration around at the time, and you could always
use the Social Sciences Citation Index to see how the research developed.
I did no follow-ups as I got into pcp shortly after, and my research
methodologies changed!
Jankowicz A.D. "Ornstein's 'storage size' metaphor: a cautionary note"
Perceptual and Motor Skills 1977, 45, 284-286.
Ornstein himself is worth reading: I don't recall the full reference, but
it was a Penguin entitled something like "On the Experience of Time" and
came out around 1972 ish.
>But if we are to have any sensation of change, we have to compare the
>actual state of the system with at least two previous and different states.
>I don't see how can this be done without a _memory_ consistent of the
>representations of past and present states.
Well, your way of looking at things still uses the memory-as-a-box (or,
perhaps, memory-as-a-process) metaphor, in which you're conceptualising
memory as a "picture of the past located somewhere in one's head". The
point about the cybernetic metaphor is that it _isn't_ about locations,
but about a set of variables which _could_ have taken different values
depending on what had happened before, but didn't, because of what _did_
happen!
Now, you might argue that this accounts for _history_ quite well, and
even for duration. (Ornstein pointed out that one's perception of
duration is related reciprocally to the amount of information processed
in a given interval; and so one could say something like perceived
duration is t - [ 1/(n x t)], where n is the number of elements that
differed from before and t is our time interval counter).
But I admit, it doesn't account for one's construing of memory as the
recall of internal representations of previous states of the system, and
I had to posit a set of elements which bore representations of previous
states of the automaton to account for that. You say that this would
involve an infinite regress, but I don't see that that _is_ necessarily
implied. If you introspect your ability to construct representations of
representations of representations etc., you'll have to admit that there
is something about the way our brains work that prevents us from doing
this sort of thing over more than three or four steps.
Try it: visualise yourself thinking about someone else, and try to
apperceive:
"I said that she said...": no problem.
"I said that she said that I said...": no difficulty.
But I'll bet you can't apperceive (="hold in your head" the percept)
"I said that she said that I said that she said..."!
(Substitute the word "know" for the word "said" if it makes the exercise
easier.)
and I imagine (very vaguely!) that this sort of thing places a limit on
the number of regresses that one would need to model.
If this interests you, please note that Jack Adams-Webber is
well-informed about the works of Lefebvre, who has a very well-developed
theory and model which according to Jack accounts nicely for this
limitation on peoples' ability to represent things recursively. (It also
provides an explanation for the Golden Section phenomenon with which Jack
is closely associated.)
Are you there Jack and could you join in to this discussion? (Yes, I know
you posted the Lefebvre reference before and I was going to read it and
then a million other tasks intervened.... Can you help Esteban?)
Cheers,
Devi Jankowicz
***********************************************
Aren't mailing lists wonderful!
They legitimise the expression of half-baked ideas
***********************************************
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%