Re: SUO: Re: Ballot Comment
Pat,
At 05:20 PM 8/14/2001 -0700, pat hayes wrote:
>>Folks,
>> Pat sets up a strawman and then knocks him down. I disagree with the
>> premise that by finding one ontology that one is saying that others are wrong.
>
>That is not what I said. What I said is that different people use
>different ontologies, and sometimes indeed people use different ontologies
>for different purposes. The idea that any of them is wrong or right is
>what I think is mistaken.
In your message you stated "The proposal to find a single coherent
upper-level ontology then amounts to an insistence that all but one of
these alternative ways of thinking are wrong." I do not believe that it
amounts to such an insistence. If you've changed your mind on this, that's
great.
>>I certainly don't intend that other ontologies are incorrect, impossible
>>or inconsistent. Merely that we can create a useful single upper
>>ontology and that has value as a standard.
>
>Calling it a 'standard' with the level of imprimateur associated with the
>IEEE is not just marking it as 'useful'. It is putting it before the world
>as a recommendation for a wide, possibly universal, set of applications
>and uses. (If this is not true, then what is the point of the
>standardisation process, and where is the claimed benefit of
>interoperability? ). To me, that is a very high standard indeed. It is not
>sufficient to simply be 'useful'; to reach that standard, it has to be
>clearly better than any other alternative, or have some clear advantage
>which singles it out from other alternatives. (One such advantage might be
>that it represents an accepted standard of usage in some community, as
>with many other IEEE standards; but there is no such community for an
>upper ontology.)
I disagree. Standards don't have to be technically superior to be useful
(not that we're not striving for excellence!). The well-worn Beta-VHS
example is a case in point. A standard just have to be good enough, and
shared by many.
>>Many other ontologies could be created with the same coverage that would
>>be just as valid.
>
>Then there is no reason to choose one of them as a 'standard'. Let me ask
>you, Adam, what you think this term 'standard' means, and why you think
>there will be any utility in having a standard, if the world goes on using
>many other ontologies? Presumably the idea of the standardisation process
>is to somehow encourage many people to conform to the standard. So whether
>or not one says that other ontologies could be created, the intended aim
>of a standardization process is to discourage or thwart such activity. If
>we are talking about the sizes of electrical plugs, that makes sense. If
>we are talking about the conceptual framework of human thought, however,
>(which we are, in effect, here) then the idea of choosing a standard is
>both more pernicious and less obviously useful.
An important distinction in my comment above is that other useful
ontologies *could* be created. The value in having a standard is that we
pick one that is good enough. Yes, we need to pick one, no, it need not be
the one true correct ontology in order to be very useful.
>>The value in having just one is that much like that in having a shared
>>human language for communication between people.
>
>You might find (if you look around on the Web, for example) that there is,
>in fact, no shared human language, but many of them. (There isnt even a
>shared human language within California.) And if you check out all the
>various attempts to create a universal human language (such as Esperanto)
>you will find that they have all been dismal failures.
This seems disingenuous. There are differences in colloquialisms and
differences in context, but when two native English speakers get together
and talk, even if one is from Wales and one is from Massachusetts for
example, that they have *huge* amounts of shared understanding. There's
value in sharing a language even if we only share 99.9999%.
Adam
>Pat
>
>---------------------------------------------------------------------
>(650)859 6569 w
>(650)494 3973 h (until September)
>phayes@ai.uwf.edu http://www.coginst.uwf.edu/~phayes
Adam Pease
Teknowledge
(650) 424-0500 x571