[Pellet-users] Inference problem I don't understand

[Bijan Parsia]

On Tue, September 4, 2007 6:35 am, Alejandro de Campos <acamposr at gmail.com> said:
[snip...it's most helpful to provide a full ontology at a public uri...otherwise I have to read through and sometimes recreate the ontology which was relatively simple in this case, but still]

 
> I expected the reasoner to tell me that Test2 is a subclass of Test1.

Your expectation is wrong, I think. Here's an explanation from OWLSight:

Axiom: 	Test2 equivalentClass Test1
Explanation:
	Test1 equivalentClass (hasX some X1)
	Test2 equivalentClass (hasX some X1)

> But
> what it really says is that Test1 and Test2 are equivalent classes!

Having something as a necessary *and sufficient* condition for a class is the same as saying that that thing *defines* the class (or is equivalent to the class). So you said that Test1 is equivalent to (hasX some X1), as is Test2. Two things equivalent to the same thing are also equivlent to each other.

You *also* said that Test2 was equivalent to (hasX some X2), but this is the same as saying (hasX some X1) is equivalent to (hasX some X2).

> And then
> if I try to get the individuals belonging to Test1 or Test2 the reasoner
> will return the same in both cases, even if the individuals "hasX x2" and
> not "hasX x1"

But this is no surprise, yes? If I'm an instance of Test2, then I *must* hasX X1 (that's a necessary condition) even if I don't know what X1 I hasX.

> What's the reason for that behavior?

The semantics of the language ;)

Cheers,
Bijan.