Lists Home |
Date Index |
An idle (but perhaps relevant) question:
Doesn't the ability to establish ontologies require that all ontological
transforms act upon data that has identical atomicity?
In case I didn't state that properly, let's take for example the
portmanteau word 'chortle', which is a combination of 'snort' and 'chuckle'.
In one sentence they may be separate ontological constructs: "He snorted as
he chuckled." One could map the sentence fragment 'snorted as he chuckled'
to 'chortled'. However, there are myriad similiar fragments to map:
"He snortingly chuckled."
"He had outbreaks of snorting between his chuckles."
"He chuckled in a snorting fashion."
All of these of course mean, "He chortled."
It's a frivolous example, but the point is a serious one: domains often
normalize data according to their needs or custom. For languages without a
Lewis Caroll, snorting a chuckling may remain forever separate. In some
domains, an address may be three fields, others five.
I guess what I'm asking: Is there really a way to come up with a general
ontological scheme that doesn't resort to rules that involve sophisticated
and arbitrary syntax analysis or pre-ontological normalization of data
across target domains?
(And I know Len's going send me links to incredibly dense philosophical
treatises, but I guess that's the price I pay for being curious.)