Re: ONT RE: Ontology case study
On Thu, May 30, 2002 at 09:02:36AM +0200, Chris Partridge wrote:
> But what is the logical status of "Drinks(john)"?
Depends on the semantics we've given to "Drinks" and "john".
> It seems to me to need some more work. Surely John needs to be
> indexical to work.
Well, of course we can't have ambiguous names, but that's not a problem
with FOL. You just have to set up your system right.
> If I am John then this may be true, if John is someone else it does
> not seem to logically follow.
It does if "John" is not ambiguous, which is an assumption built into
the semantics of FOL.
> I am not sure that "e.g., operators that can form proposition-denoting
> terms from sentences. " buys you this -
It is not being tendered for that purchase.
> but would be pleased if it did. Can you explain?
The point of such operators in the example in question was to provide
terms to serve as the objects of intentional attitudes like belief and
desire. The second argument to a belief statement is, on the face of
it, a proposition: "John believes that if he drinks, his thirst will be
slaked" appears to express a relation between John and the proposition
[if he (John) drinks, his thirst will be slaked].
> It seems to me that databases systems can do 'practical reasoning', but
> cannot explain how (a serious crime in academic circles?),
I don't know why not. In effect, we program them to do what they ought
in the given circumstances. We can look at the code that does that. I
don't see why that's not a complete explanation of their ability to do
"practical reasoning".
> whereas FOL does not -
It doesn't dance or do sums either, but we don't consider that a
shortcoming.
> but can explain what it does.
No it can't do that either. *We* offer explanations of what FOL "does"
-- better, what its properties are. Logic doesn't *do* anything, though
*we* can use it to do lots of interesting stuff.
-chris