Re: Thing and Class

I hate to get involved, but here we go-

Richard H. McCullough wrote:
> 
> RE: Thing and ClassDavid
> I don't know much about the OWL DL and nothing about ISO 15926.
> Perhaps you could explain a little of them to me/us?
> 
> I don't think your statements 1) and 2) are correct.
> 
> 1) A "Class" is not a set of things.  It is a strange sort
> of group.  If it is a "plural" class, e.g. "dogs", then
> it may be considered to be a set.  But if it is a "single"
> class, e.g. "dog", then it may be considered to be an
> enumeration (OneOf).

This has been going on for a while. Let's just be frank. Dick, this is 
wrong because you don't understand what's going on fundamentally in the 
world of semantics in the Semantic Web.

The type of semantics going on on this list is a type of formal 
semantics. In formal semantics (much like the SW in general), the names 
of things don't matter because they aren't meant to correspond to the 
real world notions. The fact that OWL uses the word "class" shouldn't be 
taken to mean that we can do some introspection about how we feel about 
the notion of classes philosophically, how we feel about the English 
word "class", or about how we use the English word "class" in everyday 
life to draw any conclusions about OWL. It's just irrelevant. It's not 
the point of OWL.

But besides, if you actually did any linguistic research into how people 
use plurals, since you brought it up, you would find that it is orders 
of magnitude murkier than OWL. There is little to gain from looking at 
English plurals like "dogs" in order to understand OWL "class". You'll 
just get more confused.

*The bottom line here is that if you are justifying things by talking 
about your intuitions about these notions or by referencing facts 
contingent on human language, like plurals, then you are not on the same 
page as everyone else.* This isn't about being right or wrong per se, 
it's about being relevant to the people you're writing to on this list.

As was suggested, learning some formal semantics will help. The best I 
can suggest are the two books I learned from:

   Language, Proof, and Logic by Barwise and Etchemendy
   Semantics in Generative Grammar by Heim and Kratzer

They were quite good books, though I credit my professors more than the 
books for what I got out of them. The second one is specific to 
linguistics. I don't know how other ways of learning formal semantics 
works, but doing it from a linguistic semantics perspective seems like 
it would be useful even if you aren't interested in linguistics because 
it gives it a nice purpose.

> 2) Classes are not members of Thing (Class type Thing).
> They are subclasses of Thing (Class subClassOf Thing).

Dick, for reasons that would be clear if you understood formal semantics 
well, what I've quoted above from you is incoherent to the rest of us. 
(The bits outside of the parens don't agree with the bits inside.) It 
may make sense in the way you interpret these things, but it doesn't 
make sense to us --- and since you're writing *to us*, I think it would 
be good to choose a language we can understand. That is, learn formal 
semantics.

While I'm taking the pedantic tone that I've adopted for this email 
(shudder), I might as well continue in the spirit: I think you should 
take a break from the semantics in the semantic web to read the books 
above carefully and open-mindedly, and treat as a homework assignment 
figuring out why what I quoted above is incoherent to the rest of us. 
The books won't agree with any intuitive notions you have about class, 
individual, etc., but you can rest assured that the rest of us think 
that the notions in the books *are* still very important.

(Okay, no more pedantic tone for a while.)

-- 
- Josh Tauberer

http://razor.occams.info

"Yields falsehood when preceded by its quotation!  Yields
falsehood when preceded by its quotation!" Achilles to
Tortoise (in "Godel, Escher, Bach" by Douglas Hofstadter)

Received on Thursday, 28 August 2008 15:08:20 UTC