SUO: Re: Mor-i-arty Relations
Patrick Cassidy wrote:
>
> I agree with Adam's view on higher-arity relations,
> and in general believe that the same criterion should
> apply to any notation that can make use of the SUO simpler,
> or make it easier to understand:
>
> > Jon,
> >
> > I think we're agreed that relations of higher arity
> > are not strictly required. The issue is whether they
> > are advantageous for reasons of clarity or convenience
> > and whether that facility would have any downside.
> > I believe they are advantageous for convenience and
> > clarity and I'm unaware of any negatives that may
> > result from their being employed.
> >
> > Adam
> >
>
> i.e. if there is some notation that is consistent
> with the fundamental representation, but makes use
> or understanding easier, it should be included as
> an optional notation. In particular, there should
> be some frame-like notation which can be translated
> automatically into SUO-KIF. However, I think it
> would be best to fix on only one frame notation
> as part of the SUO.
>
>
> Pat Cassidy
>
<...>
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤
Pat,
I am still in the process of trying to figure out
what Adam's view on "higher-arity relations" is,
and so I will reserve my comments on that until
such time as I come to understand what it is.
I had thought that my analogy with different
numeral systems for the same set of numbers
was pretty apt -- or so I reckoned at the
time. I am at a loss what to say next,
so I will wait for furher information.
Just as a hypothesis -- you understand --
I am beginning to think that I detect a slight
difference in attitude toward notational systems,
representational frameworks, and syntax in general.
For me, when the syntax gets in the way
of thinking clearly about the objects,
then the syntax has to change.
I used to be a fan of frame'n'slot systems
back in the days when I could recognize them
as revivals of some really good ideas from
both Frege and Peirce -- anything that these
two could agree on has just gotta be Good! --
namely Frege's "poly-unsaturated sentences"
and Peirce's "rhemes or rhemata". However,
the ways these representations have fallen
into lately I just no longer see much trace
of the originally good ideas left in them.
Some of the earliest projects that led me into AI
had to do with finding default fillers for frames --
but the many years that I spent on this led me into
other approaches, other directions, other solutions.
From my present point of view, one of the main and
the most adverse factors that just keeps on killing
"artificial intelligence based on logic" (AIBOL) is
the "circumstantial and inescapable nuisance" (CAIN)
of the fact that the real world just has to keep on
changing, and, partly as a consequence, that agents
of interpretation just have to keep on enterprising
to view it from many diverse perspectives. Hence,
the main theme of the work that marked my transition
into Systems Engineering, "the problem of dealing with
change and diversity in logic-based intelligent systems".
Having spent the last decade or so working to resurrect AIBOL
from that early grave, the one to which the "frame-problem" --
no relation, or is it? -- seemed to have consigned it long ago,
and, incidentally, having seen the necessity of working out the
form of this geometry whose points are views, I feel that I have
learned one or two things in the meantime, if nothing else, about
a host of approaches that I have tried before and found wanting.
I will not pass that way again,
End of Lecture (EOL),
Jon
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤