Re: [MODS] Mark Twain Rides Again (was Re: [MODS] MADS/RDF for review)

On Mon, Jan 3, 2011 at 4:06 PM, Simon Spero <ses@unc.edu> wrote:
> In Mon, Jan 3, 2011 at 9:26 AM, Dan Brickley <danbri@danbri.org> wrote:
>
> As you mention, this is a part of a bigger problem in RDF - it lacks both
> default reasoning and Micro theories, which means you have to do a lot of
> extra work when dealing with fictional, hypothetical, or inconsistent
> things.  Ontology is supposed to be On What There Is; in these cases it's
> easy to end up in WTF...
> In Cyc, predicates are true (or default true) only relative to a specified
> context, or Micro Theory.  Different micro theories make incompatible
> assertions, and theories can be nested with multiple inheritance. For
> example, rules stating that supernatural things are fictional is not made in
> the base knowledge base, but it is in the Naturalism microtheory.  Different
> cultural groups have different assumptions as to which supernatural things
> exist.

I've run across these same exact issues while working on a project to
model fictional objects, and it's gratifying to see the approaches
I've come up with on an ad hoc basis (e.g. nested inheritance)
validated by their use elsewhere.  The only difference is that I use
the term "epistemology" rather than "microtheory", and that I've found
this approach necessary even when dealing with real-world objects.
For example, what caused the extinction of the dinosaurs?  When did
Homer live?  Was he a real individual who created the works attributed
to him, or is he the composite of several unnamed poets?  Note each
theory is eminently defensible by current academic standards, and that
we have not even touched upon the politically contentious issues yet
(climate change, intelligent design, 9/11 conspiracy theories, etc.)

Received on Tuesday, 4 January 2011 22:27:31 UTC