Thanks for your prompt answers. Here are some further thoughts about time.
-_Changes_ are the measure of time -they impose the rythm. We (in our
western cultures) perceive the passing of time in the movement of a clock's
hands. Without changes, there is no time -and if changes are delayed, so is
our sensation of time. We use to think that time makes the things change;
we could all the same believe that changes build our perception of time.
And maybe time and changes are so interweaved on the basis of our
construction systems that we can't just separate them!
-Memory, of course, is memory of changes (just another version of the
Dichotomy Corollary: we must have a contrast and a likeness to form a
construct). Devi: can you explain further your suggestions about
cybernetics and memory? You say (07/06/97)
>>
The present system is a product of history: the
product of previous input-state combinations, and so memory exists in the
sense that the system _is in_ a particular state now, rather than in any
of the other states it might have been with different prior state-input
combinations.
>>
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. Must we give our automaton an
>>
internal representations of bits of its previous states among its elements
>> ?
I don't know. I fear this is becoming a little bit Gödelian! Then we'll
have to give our model the ability to compare its states, and we may start
an infinite progression of levels of autoperception.
That's it. I feel it's getting clearer, but I'll be grateful for any
contribution. Feel free to comment on this hazy ideas!
Esteban Laso
eslaso@ibm.net
----------
> From: Devi Jankowicz <anima@devi.demon.co.uk>
> To: pcp@mailbase.ac.uk
> Subject: Re: The meaning of time
> Date: Sábado 7 de Junio de 1997 04:03 PM
>
> Esteban Laso raises the fascinating question:
>
> <snip>
> >My question is: what do we mean when talking about _time_?
> >I have something in mind. We talk about _the_ past, not the past
_events_;
> >and _the_ past is a far more comprehensive construct that _the things
that
> >already had been_. Are we using a construct based on _what we can
remember_
> >against _what we cannot remember_? Or is it _what is likely to have been
> >(because it would explain the things we see now) _ vs. _what is
> >unbelievable to have happened_?
> >These are only vague ideas. I would like to get your answers to throw
> >lights on the subject. (It's a very dark and cloudy one, I know!)
>
> Here's a few thoughts, in no particular order.
>
> Since we create events by snipping out those bits of the phenomenal flow
> which our construct systems indicate might be useful, the present would
> seem to involve a more precise and in-the-near-future testing out of our
> construing than does the past. The entities which lie in the past are
> somehow more fuzzy and open to testing in a less precise way than the
> present, I guess. (Notice I'm not distinguishing between the
> present-out-there, and my memory of the past-in-here, when I think about
> the testing which construing, by definition, involves.)
>
> Talk of time involves talk of memory. In this context, you might find a
> cybernetic model of memory useful. Psychological models of the memory
> process use a locational-storage metaphor, and explore the properties of
> the input process, repository mechanisms, and retrieval process. In
> contrast, there's a cybernetic model based on the idea of the person as a
> Finite Automaton (pardon the last word, it's just a handy label with no
> particular value statements intended).
>
> The finite automaton is an information-processing system with the
> property that its present state is specified by its immediately previous
> state, together with an intervening input. (A state is defined as a
> particular set of values taken by the set of all elements which
> constitute the system; if one or more of the elements were to take
> different values, we'd have a different state.) At the simplest level, if
> we apply this model to a human being, we can begin to talk about the way
> in which a construct system at time (t) is the product of a construct
> system at time ( t minus 1) together with what has happened in between.
>
> Now, in Finite Automata theory it is possible to understand memory, not
> as a location for storage, but in terms of the history of the
> input-state relationship. The present system is a product of history: the
> product of previous input-state combinations, and so memory exists in the
> sense that the system _is in_ a particular state now, rather than in any
> of the other states it might have been with different prior state-input
> combinations.
>
> A very simple model, I grant you, but it does make it easy to model the
> "alternate-worlds" debate nicely. Another problem is that it doesn't
> account for our experience of being able to recall, i.e. address
> representations of previous states from within our present state. To do
> so, you'd need to model the present system as including internal
> representations of bits of its previous states among its elements. ( It's
> a long time since I learnt any cybernetics actively, but I would guess
> that someone in that field should have developed this by now. ) At any
> rate, it does allow us to get away from the locational model of memory as
> a box in one's head...
>
> Kind regards,
>
> Devi Jankowicz
>
>
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%