Gary Richmond wrote:
[...]
> I agree with the need for abstracting out the commonalities. For that
> reason I would stress this item from the list I recently posted:
>
>> Use visual as well as verbal representations.
>
That is an interesting thought. (01)
>
> For example, Peirce's existential graphs, Sowa's conceptual graphs and
> my own trikonic diagrams are meant to "abstract out the commonalities"
> in as diagrammatic form as is possible. I think that many of us are sick
> of words, words, words. . . (02)
I have thought about this option now and again. I think there is an
issue that the combined subtlety and expressive power of words mean that
they have a very high 'efficiency' (there's probably a better way to
express that based on Shannon's information theory work) that is
difficult to compete with.
There is also the issue that two sentences may be 'isomers', containing
the same words in different order for (only) marginally different
intent, and that symbolic 're-presentations' should capture this. (03)
However, I also have no doubt that of the many papers written about, for
example, RDF, a single (reference to a) description of the set
properties of the triples would have sufficed across the majority of
their introductions. And that presently most of these differ only in style. (04)
So perhaps visual capture at the sentence level is no good and a higher
level might be where it belongs. I wonder about geometric
representations at that higher level... say, the geometrical form of the
intent of a page's content. (05)
An idea I have been chewing on (but haven't built yet) is a sort of 3-D
(2 plus time) graph of (for want of a better description) the activation
patterns that sentences produce. (Think of symbolic chemistry diagrams.)
So if it reads a page, there is one sentence pattern per unit time, and
a page is represented by a cascade of these geometric patterns. (06)
I'm then left wondering whether 'iso-semantic' intent can be captured
under (simple?!) geometric mapping rules - a bit like having a 3-D form
collapse down to aggregate minima under gravity (or similar optimising
constraint). The original organisation of the lexicon in the 2-D plane
might matter a lot. (07)
Searching on Google Images revealed some similarities with:
Topologies for Artificial Gene Networks
http://mendes.vbi.vt.edu/AGN/topologies.html
Chemical diagrams
http://www.imr.salford.ac.uk/groups/polymers/images/diels-alder_reaction2.gif
Lattice diagrams
http://www.cyberartsweb.org/cpace/cpace/poetry/lattice/images/lattice4.gif (08)
--
Peter (09)
--
This message is archived at: (010)
http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=yak&i=434F9005.1020809@concept67.net (011)
|
