[Avery Andrews 960418]
I've put a fairly scrappy and exiguous draft on:
http://www-csli.stanford.edu/users/andrews/pctsem.txt
(just a text file). If any wants it but can't get it off the web,
I'll email it to them.
Fire when ready...
Avery.Andrews@anu.edu.au
[Avery Andrews 960418]
I've put a fairly scrappy and exiguous draft on:
http://www-csli.stanford.edu/users/andrews/pctsem.txt
(just a text file). If any wants it but can't get it off the web,
I'll email it to them.Fire when ready...
Avery.Andrews@anu.edu.au
Most excellent start. I'll look at it tomorrow and over the weekend, I
hope. Meantime, a couple of quick knee-jerks off the top half.
But the perceived beliefs and intentions of other people differ
significantly from the perceptions most often studied in PCT, in
a variety of ways. One of the differences might be described,
somewhat inaccurately, I believe, by saying
that they are `indirect'. I.e. you can't tell what someone knows
just by looking at them, but only by what they do in various situations
(answering questions, trying to open a virtual door in a 3d action
game, etc.), and even then the resulting perception might be
inaccurate: they might be lying, or playing dumb.
The PCT truism that you can't tell what someone is doing by observing what
they are doing could strengthen your point that this is not the distinction
you're after. How about "inferred" rather than "indirect"? That leads into
your next point, that it is inferred and built up from disjoint clues.
The real
difference, I suggest, is intermittency: under ordinary circumstances,
it is easy to tell whether someone you are with is tall, angry, etc.,
but what they know can only be observed occasionally, when they do
something that relies on the knowledge. This intermittency entails
that alongside the usual scalar values for a perception (how tall,
how smart, etc.) there has to be an additional `unknown' value: if
you meet somebody on a University Campus where you have both been
for a while, for example, you initially don't know how much they know
about the place that you don't, and vice-versa. You might acquire
various reference percepts for what you want them to know, but what
you actually tell them depends on what perception you manage to
develop of what they actually do know.
The key point seems to be accumulation, the building up of a story or model
out of categorical "hits". The fidelity (beggared term, maybe "viability"?)
of the story or model depends upon the granularity and aptness of the net
of categories with which one attempts to catch the fish of non-categorical
perceptions. Here comes presupposition, etc. That there is a net of
categories or something of the sort I have no doubt--absurd image of check
boxes waiting to be ticked off. Your thoughts?
BTW, "smart" strikes me as more akin to "knowledge" than to "tall".
Properties, such as DOG and LARGE, are generally thought of as being
`true' or `false' of some single entity, while Relationships, such
as CHASING, are thought of as being true or false of pairs, triples,
etc. of entities. Or, conversely [check this, may should be `dually'],
one can describe entites as instantiating or not instantiating the
property DOG, pairs of entities as instantiating CHASING, etc.
Smells like a "notational difference" to me.
Gotta go for now. Today is moving day, and all the computers shut down at
noon EDT.
Bruce
[Avery Andrews 960420]
(Bruce Nevin 960419)>
The PCT truism that you can't tell what someone is doing by observing what
they are doing could strengthen your point that this is not the distinction
you're after. How about "inferred" rather than "indirect"? That leads into
your next point, that it is inferred and built up from disjoint clues.
`Inferred' vs. `Direct' perception is a hot potato in psychology
(or used to be at least), and I also think it is content-free,
so I'm trying to describe it in observationally meaningful terms
like `intermittent', etc.
The key point seems to be accumulation, the building up of a story or model
out of categorical "hits". The fidelity (beggared term, maybe "viability"?)
of the story or model depends upon the granularity and aptness of the net
of categories with which one attempts to catch the fish of non-categorical
perceptions. Here comes presupposition, etc. That there is a net of
categories or something of the sort I have no doubt--absurd image of check
boxes waiting to be ticked off. Your thoughts?
Not sure I get this, but certainly the catogories associated
with words, such as `tall', underspecify the corresponding
perceptions.
BTW, "smart" strikes me as more akin to "knowledge" than to "tall".
Certainly, but in that context I wanted to group them together
just as potential perceptions.
>Properties, such as DOG and LARGE, are generally thought of as being
>`true' or `false' of some single entity, while Relationships, such
>as CHASING, are thought of as being true or false of pairs, triples,
>etc. of entities. Or, conversely [check this, may should be `dually'],
>one can describe entites as instantiating or not instantiating the
>property DOG, pairs of entities as instantiating CHASING, etc.Smells like a "notational difference" to me.
Yes, but there are philosphical problems about individuals (can
there be individuals with no properties, two distinct individuals
with the same properties, etc.) that might be avoided if you think
of the properties and relations as basic and the individuals as
derivative.
Avery.Andrews@anu.edu.au