questions about "Languages with Self-Reference I" (Perlis85)

I recently followed the refences from KIF3 [2] back to your 1985 paper
[5] which appears to solve a lot of important problems.  I'm working
on Semantic Web development at W3C / MIT-LCS [3], and we're very
concerned with reasoning about belief and trust.  My colleagues and I
are generally trained more as software developers than logicians, but
we're learning.  (I've been helped immeasurably by occasionally
discussions with Pat Hayes, Drew McDermott, Jim Hendler, Lynn Andrea
Stein, and a few other kind souls you may know.)  Anyway, your work
looks tremendously useful, but I have a few questions I hope you might
help me with.

1.  In 1995, the KIF standards committee decided [4] to stop using a
truth predicate like yours in favor of the simpler "wtr", mostly on
the grounds it and easier to implement (and, I think, reasoning about
beliefs wasn't important to them).  Do you have any implementation
experience with your proposed system?  Do you know of systems which
use it?  I can imagine axiomatizing it in FOL, and using a generic
theorem-prover with it, but that's perhaps not very tractable.  Do you
know of such axioms written out somewhere?

2.  Calling the truth predicate "true" seems misleading; do you know
of a more precise name, something which evokes the Gilmore's star
which is part of the definition?  My best attempt so far is
"neutrallyTrue", which suggests to me the kind of slight disempowering
the star seems to perform.

3.  Do you know of any important follow-up work, supporting or
undercutting your work?  How about explanations that are less
technical than yours, but still detailed enough for a potential
implementor to use?

4.  Do you have rights to publish your paper on the web?  I have a
scanned copy, but of course by law I can't share that widely.

5.  In KIF's LISP-like usage, and even more so in the RDF-based usage
I imagine, the truth predicate is applied to an abstraction of a
sentence (some tree/graph of logical connectives, predicates, and
terms) rather than a simple sequence of characters.  This might
introduce additional dangerous self-reference possibilities if not
done carefully.  Do you know of any work on this issue?

Thank you so much for any information you can provide.  I greatly
appreciate your time in this.


     -- Sandro Hawke                http://www.w3.org/People/Sandro/

Note: I have CC'd a public archive [1].  If you CC it on your reply,
it will be permanently archived on the web and thus much easier for me
to refer to and share with people.  Feel free to omit it if you're
more comfortable answering privately.  Don't worry about addressing a
wider audience in a CC'd reply; the archive is understood to simply
contain e-mail that people are willing to have kept around for future
reference.

[1] http://lists.w3.org/Archives/Public/www-archive/
[2] http://meta2.stanford.edu/kif/Hypertext/node35.html
[3] http://www.w3.org/2000/01/sw/
[4] http://logic.stanford.edu/kif/decisions.html
[5] @article{Perlis85,
  AUTHOR = {D. Perlis},
  TITLE = {Languages with Self-Reference I: Foundations. (or: We Can Have Everything in First-Other Logic!)},
  YEAR = 1985,
  JOURNAL = {Artificial Intelligence},
  VOLUME = {25},
  PAGES = {301-322},
  KEYWORDS = {}}

Received on Friday, 20 September 2002 15:36:54 UTC