Re: Fwd: SUO Quo Vadis
On Saturday 17 December 2005 03:34, John F. Sowa wrote:
> Rob,
>
> I agree:
> > Anyway, I think there is a lot in common between what
> > you are saying, what I am saying, and now what I understand
> > Mark to be saying: that we need to focus on the human character
> > of ontology, that we need to think of ontology as the result
> > of a process, and "different perspectives for different purposes".
> >
> > It would be nice if we could capitalize on that commonality.
>
> But I would avoid the term "human character" because it is
> much too confusing and much too likely to lead to the kinds
> of topics that I *don't* want to address.
Fair enough. I was just trying to be a little inclusive.
Mark is characterizing this thing as human. I don't think that is wrong. I'd
like him to see we are talking about similar things.
If we're going to endlessly philosophize about the meaning of words we are
going to get nowhere, everybody. The debate which I find interesting on this
list that between a world which can be described in a single and complete
way, and one which reflects some necessarily multiplicity of description
(note: multiplicity of description not multiplicity of the world.)
I think first we need to identify that division, and then argue about the
details.
Mark initially presented somewhat cryptically, but I think he seeing the same
thing.
> As an example of the kind of topic we must address, consider
> the definition of the word _business_. People have gone round
> and round in circles trying to define a business as a set or
> a mereological sum or other kinds of nonsense. But they omit
> the fundamental principle:
>
> The goal of a business is to make money.
>
> There is, of course, the intermediate principle:
>
> In order to make money, a business must have the goal of
> producing a product or delivering a service that earns money.
>
> And there are many other intermediate goals implied by these two.
>
> As I said before, you can talk about the goals and strategies
> used in a chess program and implement them in a purely formal way.
> You can also talk about the goals, intentions, strategies, and
> deterrents used in business, medicine, engineering, law, etc.,
> in purely formal ways. If an ontology is to have any use whatever
> in any of those areas, it must address goals and intentions.
I don't disagree. The question is how we address this.
As you know I think the direction to seek solution is by treating knowledge
representation as a search problem. I guess I would address this by saying
the user's goals and intentions are encapsulated in the question he/she asks
of the system.
Alternatively that if the user submits a query for which the goals and
purposes of a business are relevant, then information in the system relating
to businesses will tend to organize itself in that way: the structure of the
search results reflecting the structure of the query.
A third stab, if goals and intentions are really important to business, which
I don't doubt they are, then goals and intentions will tend to force a
pattern on information relating to businesses. Search on businesses, and you
will tend to find those patterns (perhaps the word "money" will come up a
lot.)
That is how I would tend to address the problem in terms of ad-hoc structure.
Naturally--you can address "different perspectives for different purposes"
simply by allowing them to exist, _if_ your model is fundamentally one of
finding structure, rather than imposing it.
In the interests of generating broader consensus let me ask how you would
address this?
-Rob