Re: Types v. monadic relations

• Mail folder: Interlingua Mail
• Next message: sowa@watson.ibm.com: "Requirements for a standard"
• Previous message: YOUR NAME HERE: "Re: Types v. monadic relations"

Message-id: <9202032155.AA10451@quark.isi.edu>
To: YOUR NAME HERE <hayes@sumex-aim.stanford.edu>
Cc: sowa@watson.ibm.com, interlingua@ISI.EDU
Reply-To: macgregor@ISI.EDU
Subject: Re: Types v. monadic relations 
Date: Mon, 03 Feb 92 13:55:49 PST
From: Robert MacGregor <macgreg@ISI.EDU>

MacGregor:
> > ...the semantic difference between` a "type" and a "monadic relation"?  
> > Why should we care?
Hayes: 
> Good question. But apparently some people do, so why not allow predicates to
> have a limited range of properties ( as in being-a-type-predicate ). 

If there ought to be a distinction between types and monadic
relations, then I agree that asserting something like
   (being-a-type-predicate FOO)
looks like the simplest way to do it.  So I endorse that
much of Pat's suggestion.

However, my feeling is that each feature in a KR language ought to
have one or more *inferences* associated with it (or constraints
on usage, or something).  Therefore, I would prefer not
to introduce a "being-a-type-predicate" predicate into KIF or
wherever unless it was accompanied by some indication of
the semantics or style of use implied by the predicate.

If anyone wants to propose that types and monadic relations be
distinguishable, then I suggest that their suggestion should only be
taken seriously after they indicate what the difference is.

- Bob