Chris and David, (01)
JFS>> To support modal logics and other kinds of things, there is
>> only one additional operator needed: the that-operator of IKL. (02)
CP> Does this support deontic logic? (03)
Short answer: Yes. (04)
Longer answer with a bit of history: (05)
The old KIF had a powerful backquote operator, which very few
people used because they had no idea what it could do, but which
some people found extremely powerful for doing wondrous things. (06)
I used the backquote of KIF to support the metalevel features
of conceptual graphs, and I wanted a similar feature in CL.
But Pat H. and Chris M. didn't want to include it in the CL
semantics because they didn't have a clean way to handle it. (07)
Some people, including Doug Lenat and me, kept lobbying for
something that had similar functionality. Lenat said that
he could use the context mechanism of conceptual graphs to
represent CycL, but that mechanism required backquote. (08)
The IKRIS project, which included a number of participants
who were involved with a variety of knowledge representation
issues, included a subcommittee to develop an IKRIS Knowledge
Language (IKL). See (09)
http://www.ihmc.us/users/phayes/IKL/SPEC/SPEC.html (010)
The only feature added to CL was the that-operator: (011)
(exists (p) (= p (that (blue sky))) (012)
This says that there exists a proposition p, which is equal
to the proposition that the sky is blue. This operator
also allows quantifiers outside the that-proposition to
bind variables inside the proposition: (013)
(exists (q x y z) (= q (+ x y z))) (014)
This feature gives Lenat and me the ability to define the
metalevel operators and other kinds of definitions we would
like to have. In my summary of conceptual graphs, I summarize
how I use this operator in CGs and their translation to IKL: (015)
http://www.jfsowa.com/cg/cg_hbook.pdf (016)
See Section 4, pp. 16 to 22. The following note (which is
adapted from an excerpt in my KR book) has more detail
about propositions, and we used it during our discussions
about the IKL semantics: (017)
http://www.jfsowa.com/logic/proposit.htm (018)
The following two papers (which I wrote before the IKL project)
describes how metalevel operators about propositions can be
used for a very flexible semantics, including the ability to
define modal logics (including deontic): (019)
http://www.jfsowa.com/pubs/laws.htm
Laws, Facts, and Contexts:
Foundations for Multimodal Reasoning (020)
http://www.jfsowa.com/pubs/worlds.pdf
Worlds, Models, and Descriptions (021)
In the laws.htm paper, I defined Nested Graph Models (NGMs),
which were inspired by conceptual graphs. But they can also
be represented by IKL. (022)
CP>> My impression is the deontic aspect of models in commercial
>> systems is often not handled formally, but rather by informal
>> craft and convention. (023)
DW> Perhaps I'm being a bit hard nosed, but if you know that
> the modeling language has a convoluted intended interpretation,
> why would you not define the Common Logic internal logic to be
> exactly that interpretation?
>
> If you are going to use logic, I would think you would want
> to be exact. (024)
The advantages and the difficulties of any formal definition
are that they force vague and ambiguous informal practices to
be stated in a single, unambiguous way. The people who write
the formal definitions must do a lot of hard work to make it
precise while keeping it as close as possible to the informal
intuitions. (025)
CP> Where a modelling language has an unclear on convoluted intended
> interpretation (not uncommon) merely translating it into CL as
> best one can will not resolve this. Though you could argue it is
> a good first step to sorting things out. (026)
Vague, loose, and unclear specifications are usually implemented
in ways that the designers were never clear about. Furthermore,
different implementers will interpret it in different ways. (027)
That is what happened to JavaScript, and users rarely wrote
complex programs in JavaScript because they had no idea how
different browsers (even different versions by the same vendor)
would interpret the language. (028)
Note that JavaScript made a major leap in acceptability *after*
ECMA defined the precisely specified ECMAScript and the major
vendors implemented it. That definition supported the huge
increase in AJAX applications after Google Maps adopted it. (029)
I admit that the definitional work may be nontrivial, but
languages above the IKL level will look exactly the same as
they did before they were formally defined. The important
advantage is that different implementations that enforce the
formal definitions will be guaranteed to be interoperable. (030)
CP> In the academic sphere there is also... research on
> using Searlean performatives - which are, I believe, not
> in CL. (031)
Yes, but they are supported by IKL. By the way, I'd call
them the Peirce-Wittgenstein-Austin-Searle performatives.
At our VivoMind company, we use them in the Flexible Modular
Framework: (032)
http://www.jfsowa.com/pubs/arch.htm
Architectures for Intelligent Systems (033)
http://www.jfsowa.com/pubs/paradigm.pdf
Two Paradigms Are Better Than One,
And Multiple Paradigms Are Even Better (034)
In the FMF, the pragmatics field (which specifies a
performative) is separate from the language field. That
allows languages without a metalevel ability to be used
in messages that have an attached performative, which
indicates the purpose of the statement. (035)
CP> If we assume for a moment that we have formalised this,
> then my guess is this revised modelling language would not
> be a sub-set of CL as CL cannot (I believe) handle deontic
> logic. (036)
Yes. But note the use of metalevels in the laws.htm and
worlds.pdf papers. In the FMF, by the way, the metalevel
doesn't go beyond a single proposition. That alone is
a big step for many applications. But the laws and worlds
papers show that a hierarchy of multiple metalevels can
give you an enormous amount of expressive power. (037)
The that-operator of IKL can support those hierarchies. (038)
John (039)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/sio-dev/
Join Community: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/sio-dev/
Unsubscribe: mailto:sio-dev-leave@xxxxxxxxxxxxxxxx
Community Shared Files: http://ontolog.cim3.net/file/work/SIO/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?SharingIntegratingOntologies (040)
|