On Jun 14, 2007, at 9:57 PM, John F. Sowa wrote:
> Kathy and Pat,
>
> I agree with Pat's explanation, but I think it could be made
> somewhat clearer by distinguishing the base domain D from
> the domain D' of *all* relations over D for second-order logic
> (and then a domain D'' of *all* relations over D', etc.).
>
> PH> The key semantic difference between the other logics is
>> that they all impose conditions on the domain, requiring
>> it to contain some entities as a result of containing others.
>> So for example, classical second-order logic semantics
>> requires that the domain is contain all relations
>> over the base domain.
>
> I would rephrase the last sentence in the following way:
>
> So for example, classical second-order logic semantics
> starts with the given base domain D and introduces
> another domain D' of *all* relations over D.
>
> I just wanted to give different names D, D', D'', etc.
> to distinguish the base domain D from any domains that
> may be introduced by implicit assumptions.
>
> CL allows the domain D to contain relations, but it doesn't
> require D to contain *all possible* relations. (01)
Indeed the mathematical facts require that it *not* contain them.
There are (as of course John and Pat know) 2^card(D) relations over
any set (taking relations here to be sets of n-tuples). If (per
impossibile) D were to contain all possible relations over D then
(even ignoring issues of non-well-foundedness) we'd have card(D) =
2^card(D), which of course contradicts Cantor's Theorem. (02)
-chris (03)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (04)
|