Re: SUO: Re: KIF Syntax & Semantics & A Basic Ontology
> > I have a little (ok, a lot) of trouble reading this due to
> > the lack of care or maybe the different conventions in the
> > use of quotation marks. I know it aint easy in ASCII, what
> > with needing quotes for emphasis and all. But when you write
> > something like this:
> >
> > > There is really only one complication that needs attention,
> > > namely, that an expression of the form (P T1 ... Tn) can occur
> > > both as atomic sentences and as a term in an atomic sentence.
> > > ...
> >
> > Do you mean:
> >
> > > There is really only one complication that needs attention,
> > > namely, that an expression of the form "(P T1 ... Tn)" can occur
> > > both as atomic sentences and as a term in an atomic sentence.
>
>Well, actually, no, though I admit that the way I put it is not
>excruciatingly correct. The problem with your suggested correction is
>that ordinary quotes turn the expression in question into a name for
>that very expression, i.e., the expression left paren - upper case P -
>space - upper case T - the numeral "1" - and so on. And that is not
>what I mean, as "P", "T1", and "Tn" are *metavariables* that can take
>any terms of the language as value; I don't mean to be talking about
>that very expression. Hence, what are really needed here are
>so-called Quine corners, or corner quotes, which work sort of like the
>LISP backquote operator, in whose scope variables (properly marked)
>can continue to function as variables that can take values. But we
>have no corner quote ASCII characters, and rather than invent them, I
>simply used the expression "of the form" as a sort of corner quoting
>operator to tip off the reader that the expression to follow is a
>general schema, and that "P" stands for any term and "T1 ... Tn" any
>row of terms. (Though there may be other places where I end up using
>quotes as corner-quotes where they might have been a danger of
>ambiguity -- I'm sure I could have been more careful still.)
>
> > As far as the substance of this suggestion goes, I know you are
> > working in some other flavor, so maybe what I say will not apply,
> > but I used to think that I could get away with this very thing --
> > loving polymorphism as much as I do -- and I discovered to my
> > considerable grief that I cannot, and so there is now a whole
> > subsection of my dissertation that is devoted to saying why and to
> > building a work-around. Basically, I had to create two different
> > types of B domains, there distinguished by single and double
> > underlines, one for the NP type (or noun phrase grammatical
> > category) and one for the S type (or sentence grammatical category).
> > Again, this may be a side-effect of my needing to preserve the
> > option of a functional interpretation at all times, but you may want
> > to think about it.
>
>I agree that quotation is a delicate matter indeed. But I must say,
>Jon, that, in this case, I am quite puzzled as to what you could
>possibly find problematic here. I really don't see any reasonable
>interpretation of the passage in question other than the one I
>intended. I'm clearly not *using* the expression, and it's also clear
>that I don't intend to be talking about just that one single
>expression (with the ellipsis, it's not even a legal KIF expression).
>That the expression is a schema for all KIF sentences of that form,
>i.e., all atomic sentences) seems to me to be the only game in town!
I might add that the row-variables (those whose first character is
'@') provide just the expressive power one needs here in the language
itself, since the use of a universally quantified row variable
corresponds almost exactly to the informal but common usage of the
three-dots convention to indicate some arbitrary sequence of marks,
as in Chris' introductory discussion. The soon-to-be-proposed-new-KIF
expression
(P @x)
has as its instances all expressions of the form indicated, ie any
expression consisting of a sequence of lexical items and
subexpressions comprising, from the left: an opening parenthesis,
followed by the letter 'P' , then by a finite number of KIF terms,
then by a closing parenthesis.
However, Jon's second extract seems to be referring not to the
subtleties of quotation, but to the need, as he sees it, to maintain
a sharp distinction between expressions which are thought of as
referring to things (NP type) and expressions which are thought of as
making assertions (S type). This is traditional in most versions of
logic, but Menzelian ingenuity seems to have made this distinction
unnecessary. One can always adopt it, if that feels more comfortable,
of course, but there seems to be no pressing semantic need to do so,
as far as we can determine. Jon, in case we have missed something,
can you tell us what was the source of your grief?
Pat Hayes
---------------------------------------------------------------------
IHMC (850)434 8903 home
40 South Alcaniz St. (850)202 4416 office
Pensacola, FL 32501 (850)202 4440 fax
phayes@ai.uwf.edu
http://www.coginst.uwf.edu/~phayes