RE: ACTION-219: review of CORE (more)

Bijan and Gerd,

I am glad you were able to bring a more positive atmosphere to this
interchange.  As Obama said the other day, we can "disagree without
being disagreeable." However, I would like to say a few things about
the remarks I made that might help to see where I was coming from and
that no "bombastic claims" were intended.  

When I said that "logic tells me what follows from what" and  I
referenced a distinction between logic and reasoning, I was talking
about logic in the sense of the whatever it is "in reality" that
accounts for a conclusion's following with necessity from a set of
premises. I should have used the term "deductive logic" to reference
this phenomenon.  

I do not identify that phenomenon with predicate calculus (FOL) or any
other theory that seeks to account for it either in whole or in part.
Moreover, as a theory, FOL is probably far from perfect.  It is
doubtful that FOL accounts for all our intuitions concerning deductive
validity.  Moreover, there are arguments sanctioned by FOL that are
arguably invalid: because standard FOL presupposes a non-empty domain
of discourse the statement (Exists x) F(x) "follows from" (Forall x)
F(x).  Such an inference is invalid in so-called "Free logic." 

I recognize that people use the term "logic" to characterize other
phenomena involving "warrented" or "justifiable" inference than
deductively valid inference.   There is a very long philosophical
tradition of skepticism  with regard to the possibility of establishing
a "science" of such non-deductive inference.  I am not saying that I
put myself in that camp, but I am influenced by it to a certain extent.
However, that is a philosophical matter and it does not translate into
any specific position with regard to RIF core.  In particular I
recognize that logics based on closed-world assumptions make a lot of
sense in logic programming applications such as deductive databases. 

But if someone claims that "Tweety can fly"  "follows from"  1) "Tweety
is a bird" and 2) "If something is a bird and it is consistent to
assume it can fly, then (it is reasonable to) assume it can fly", then
I think they are mistaken because it is possible for the premises to be
true and the conclusion to be false.  Perhaps other people think it is
acceptable to use "follows from" to characterize the relationship
between premises and conclusion in this case.  I suppose as long as the
distinction between deductive validity and this type of inference is
recognized no harm is done, but it does make for confusion.

Allen




-----Original Message-----
From: Bijan Parsia [mailto:bparsia@cs.man.ac.uk] 
Sent: Wednesday, February 14, 2007 8:45 AM
To: Gerd Wagner
Cc: Ginsberg, Allen; kifer@cs.sunysb.edu; public-rif-wg@w3.org
Subject: Re: ACTION-219: review of CORE (more) 

On 14 Feb 2007, at 13:07, Gerd Wagner wrote:

>>>>> Not being "consistent with classical semantics" doesn't mean
>>>>> much for what we are trying to do. No important computational
>>>>> formalism I'm aware of is "consistent with classical semantics".
>
>> ...
>
>> So I take this to be similar to Allen's arrogation of the term
>> "logic" to mean only classical logic.
>
>> ...
>
>> So I respectively request that people refrain from making such
>> bombastic claims, regardless of their personal perceptions of the
>> merits of those claims.
>
> Yes, Bijan, you are right.
>
> But sometimes it is tempting making a strong claim to
> counter a misconception (namely that true logic is
> classical logic and therefore that's what we should
> try to use/impose in/to the world of computation).

I agree it is tempting. I was so tempted wrt to *your* claim :) But  
as fun as sarcasm and bombast are, sometimes it's just more  
productive to refrain.

(I am v. bad at resisting this temptation, or the sarcasm temptation,  
in a lot of circumstances.)

Glad we're on the same page!

(FWIW, I don't think there's bad faith anywhere in here! I just think  
a little more generosity on terminological matters and acknowledgment  
of other terms and perspectives would be good.)

Taking such disputes offline until one reaches a resolution is  
sometimes helpful as well.

Cheers,
Bijan.

Received on Wednesday, 14 February 2007 23:06:36 UTC