SUO: Re: Re: Time, context, and relations
Pat, John and others,
So I propose that we partition things into three fundamental *metaclasses*,
which, if the CSPers will pardon me, I will call 3rdness, 2ndness and
1stness. And that we assume the existence of two embedding metafunctions
that reify 3rdness into 1stness, and at the same time transform the linking
2ndness into a 3rdness. I also assume that 2ndness is relative -- there is
no absolutely required 2ndness. This framework has already aided me in
defining category-based metatheories for logical languages and models.
_______________________
III. 3rdness (nexus) (the central thing)
Represented idea: verb sense, situation, eventuality, state of affairs,
relation, CG conceptual relation
II. 2ndness (prehension) (the spokes or connecting links, possibly
partitioned into *obligatory case* and *optional case*)
Represented idea: case (thematic role), argument index, CG arc
Examples:
[who?] -- agent -- (agentive case)
[using?] -- instrument -- (instrumental case)
[what?] -- them(atic) object -- (objective case)
[to(for) whom?] -- recipient (patient) -- (dative case)
[where?] -- location -- (locative case)
[when?] -- time (tense) -- (temporal case)
[why1?] -- purpose
[why2?] -- cause
[how?] -- manner
-- epistemic source
-- type
I. 1stness (actual entities) (the peripheral things)
Represented idea: noun, (argument) entity, CG concept
_______________________
Robert E. Kent
rekent@ontologos.org
----- Original Message -----
From: "pat hayes" <phayes@ai.uwf.edu>
To: <sowa@bestweb.net>
Cc: <standard-upper-ontology@ieee.org>
Sent: Thursday, January 11, 2001 2:15 PM
Subject: SUO: Re: Time, context, and relations
>
> John Sowa wrote:
> >Graham Horn wrote:
> >
> > >> However, I still don't understand why relations
> > >> do not also have a time dimension.
> >
> >Jon Awbrey responded:
> >
> > >A k-adic (-ary, -valent) relation with a time dimension adjoined
> > >is just a special kind of a (k+1)-adic (-ary, -valent) relation.
> >
> >Time is a very important aspect of "context", but it's only
> >one of many ways of classifying contexts. Place is another
> >important aspect. Combining place and time in a 4-D system
> >is useful, but it still doesn't capture all the possible ways
> >of classifying contexts.
> >
> >Other useful classifications include the people involved, the
> >general subject matter, the ontology used to classify the
> >subject, the reporter who reported the information, etc., etc.
> >
> >Example:
> >
> >The sentence "A cat is on a mat" could be represented by
> >the following conceptual graph (which could be translated
> >to KIF or any other logic-based notation):
> >
> > [Cat]->(On)->[Mat].
> >
> >The dyadic relation (On) could be augmented with arbitrarily
> >many additional arguments, coordinates, attributes, or whatever
> >you want to call them:
> >
> > (On)-
> > -1->[Cat]
> > -2->[Mat]
> > -3->[Time: <2001, Jan, 11, 16:12:05 UCT>]
> > -4->[Location: <73W 53' 35", 41N 12' 27">]
> > -5->[Reporter: 'John F. Sowa']
> >
> >Instead of stuffing all that information into the relation,
> >it is possible to wrap the original graph (or formula or
> >sentence or statement) in a context wrapper of type Situation
> >and then to add additional links to the context box:
> >
> > [Situation: [Cat]->(On)->[Mat] ]-
> > ->(Ptim)->[Time: <2001, Jan, 11, 16:12:05 UCT>]
> > ->(Loc)->[Location: <73W 53' 35", 41N 12' 27">]
> > ->(ReportedBy)->[Person: 'John F. Sowa']
> >
> >This CG could be read as the English sentence "There existed
> >a situation of a cat on a mat at the point in time 2001 Jan 11
> >16:12:05 UCT at the latitude 73W 53' 35" and longitude 41N 12' 27"
> >as reported by the person John F. Sowa."
> >
> >To represent this CG in KIF, one might introduce a dyadic
> >relation Dscr (description), whose first argument is some
> >entity that is described by a quoted proposition as its second
> >argument:
> >
> > (exists ((?s Situation))
> > (and (Dscr ?s '(exists ((?x Cat) (?y Mat)) (On ?x ?y)))
> > (Ptim... )
> > (Loc... )
> > (ReportedBy... ) ))
> >
> >The dots indicate some conventions that the SUO might specify
> >for representing times, locations, and other information.
>
> Fine with all this, John. But a few questions....
> 1. Why not go the whole hog and have the cat, and the mat and even
> the sitting, as aspects of the situations? After all, it seems that
> the choice of these to put into a sentence representing content, as
> opposed to being kind of apppendages, is rather arbitrary: the text
> could have started :"John Sowa saw it happen one afternoon...". If
> one pursues that line, one gets to a representation very similar to
> case grammar, where there is just a central 'thing', which might be
> called variously a situation, a proposition, an eventuality, or a
> state of affairs, or whatever, (it is whatever is basically expressed
> by a simple sentence, is the idea) and it has a largish class of
> known relations with various categories of other thing: the type (a
> sitting, in this case), an agent, a subject, a time and a place, an
> epistemic source, whatever. (Extra ones can be added later, which is
> a handly contrast to the representational technique which insists on
> putting them in as arguments to a relation.) Some of these other
> things can themselves be situations, allowing us to talk (with some
> care) about various kinds of aspect or intentionality relationships.
> I suspect you will agree with all this, in fact.
>
> 2. On the other hand, quite what ARE these central things to which
> the other things are related? This is where I tend to get a little
> lost. I know what propositions are, more or less (I think of them as
> equivalence classes of sentences under logical equivalence, more or
> less), and I know what times and places are, and I know what people
> are, and so on. More or less. But I really don't know what
> 'situations' ARE: and without some idea about that, I don't have any
> way to evaluate anything that anyone might say about them. To follow
> the line that Nicola has always emphasised, I don't know what the
> identity conditions for situations would be.
>
> I have met several extremal views on this, and they all make a kind
> of sense. On one view, a situation is basically a fragment of the
> world: some selected aspects of the way the world actually is.
> (Leaving aside issues about imaginary things, etc., for now.) On that
> view, for example, a fast walking and a slow travelling might well be
> the same situation, described in two different ways. This is the
> usual view among linguists, oddly enough, as far as I can determine.
> On another view, a situation is more like a piece of information
> about the world, and the 'same' physical situation might be a
> different information-situation depending on how it is described. It
> is impossible to describe the same situation in different ways: the
> differences in the descriptions are already sufficient to distinguish
> the described situations from one another. This is what Barwise
> apparently meant in 'Situation Theory', and what Hobbs means in his
> promiscuous ontology. The intuitions get even odder when one thinks
> of more complicated descriptions. If there are two cats sitting on
> the same mat, is this one or two situations? How about the same cat
> sitting on two different mats? (Why the disparity, if there is one?)
> Can situations be characterised by conditionals or disjunctions or
> negations?
>
> Now, both of these seem to me to be reasonable answers to my basic
> question (though I prefer the first one, myself.) But Im really not
> sure what your answer is: sometimes you seem to prefer the 'physical'
> way of thinking and sometimes the 'information' way, so there seems
> to be a kind of hybrid view in there somewhere, but I can't quite
> discern what it is. (Maybe, perish the thought, I need to understand
> Peirce first...?)
>
> Pat Hayes
>
> ---------------------------------------------------------------------
> IHMC (850)434 8903 home
> 40 South Alcaniz St. (850)202 4416 office
> Pensacola, FL 32501 (850)202 4440 fax
> phayes@ai.uwf.edu
> http://www.coginst.uwf.edu/~phayes
>