RE01 Rice Brian T. EM2
Wed, 25 Nov 1998 18:19:10 +0900
>From: RE01 Rice Brian T. EM2
>Sent: Tuesday, November 24, 1998 9:32 PM
>To: 'Codrin Nichitiu'
>Subject: RE: tunes-0.0.1.c
>I'd like to develop a little more the "arrow paradigm" w/ u, if u like...
>Once upon a time I had a similar idea, though I didn't follow
>its leads to a "clear" and "shining" end, I was only tossing
>I guess we should focus more on the "axiomatic" frame first,
>before really asking how to implement the whole thing. Of
>course, the implementation will soon follow, otherwise it'd all
>be pointless, but before, IMHO, we should clear things out, clean
>our ideas and have a neat frame.
>well, i could interpret the arrow-frame argument in terms of predicate
>reasoning as we go along. for instance, i could restate my question of "how
>to develop the machine model to the extent that infinitary arrow structures
>could be constructed" into "how to place a finite-based predicate system of
>logic in a context where it could encapsulate the idea of, for instance, the
>number of natural numbers". the theorems of church and goedel and their
>related fellows provide an ample structure proving the inadequacy of standard
>logics alone to cover this limitation. if this helps you to understand the
>nature of my question, then please respond. as well, i can continue to
>explain things in this way, although i'd rather not, given the semantical
>ambiguity of arrow-structures in their homo-iconic austerity.
>My burgeon of idea (and I do mean to be modest) was that
>sentences (even in natural language sense... as much as
>we can formalize it :) should be such paths in a (multi-)graph.
>I couldn't come up with something very meaningful for the nodes...
>either the equivalent of words, or sentences themselves...
>well, if you understand the nature of the homo-iconic system which i wish to
>build, you will understand then that the nodes will never have any definite
>semantical meaning, and never _should_ have. the point is to create multiple
>contexts constructively to build the idea of complete knowledge-level
>reflective computing. the context, itself will be nothing at first, with
>nodes being, say, a single "reserved arrow" at location 0x0000, for instance,
>whose source and destination are itself. (ok, so that was quite a hack. but
>it might just serve the purpose nicely). semantics (as ontologies, for
>instance) are relative, as i've stated repeatedly. the arrow structure and
>its infinite capacity for semantical interpretations. more later...
>This is very rough, intuitive, incomplete and blurry,
>I'm aware of it, but if u'r willing, we could
>stretch it and clean it up step by step.
>My hunch is somehow we shouldn't get too far from
>natural languages. Of course, I won't discuss anything which
>is not supposed to end up completely formalized and mathematically
>written, but at the beginning maybe we can play around a little... :)
>natural languages are just around the corner, i assure you. my intent is to
>build up these things as structures from a common arrow-frame world, as it
>were, and to make interfaces out of these.
>(PhD student, theoretical computer science, 2nd year, ENS Lyon, France.
>I'm doin' cellular and graph automata, and some formal languages stuff :)
>p.s. However, most of all what I would like to see is some results of the
>chewing all of you I am sure have done with the outflowing of my words over
>the last month or so. The more interpretations I get of my words thrown back
>at me, the more I can in turn create for us. Trust me when I say that many
>of my ideas still lay below the surface of this interface, waiting for the
>light of day when I am able to express them to you. To the future!!