I hope, my message will now be posted at last. Why is that so hard for me?
Are my posts moderated? I first tried it four hours ago... Now I'm posting
via the Listserv-Web-Interface.
I think this discussion about FRBR and how to relate individual items and
create concepts and whether there should exist some kind of rule-framework
for doing this poses essential questions for needed developments in the
library and information field. I would like to consider it from a more basic
view. It's an epistemological question and it's all about identities: What
identities and which resulting classification with how deep a resolution
fits the respective purpose? This is a question that can't be answered from
a global point of view. Each person answers it in respect of her own
purposes. And that's as it always has been: A scientist might have a more
refined scheme for classifying things while a layman looking at the "same"
subject makes only view distinctions. A fashion designer or artist should
have a more refined colour-scheme and describes a jacket as ultramarine blue
while a person pointing at another person in a mass only denotes the same
jacket with 'blue'. And in the way different purposes arise and groups
emerge which deal with different subjects there are new classification
refinements made which coexist with the "common" ones.
N.B., that doesn't break with the postulate of tertium non datur (i.e. that
something is either A or not-A, either a work or not a work). Within a
classification scheme used by one person or a group one entity should be
either a work or not a work, either a fish or not a fish. Otherwise this
classification scheme is wrong and might be useless. There are good and bad
classification schemes and the quality of a scheme lies within itself not in
comparison with a beforehand structured world.
Contradictions to a binary logic emerge not until we compare classification
schemes of different people or of the same people but different contexts.
Let's take a look at biologists and kids: For the first a whale is not a
fish, for the second it might be none of the predicaments is wrong. (Even a
biologist talking to his kids about the whale being a fish doesn't get
caught in cotnradictions.)
To come to the subject of semantic web and relation-building: I think there
shouldn't be restrictions to the user like pre-constructed classification
schemes or relation-types and complex rules how to treat different cases. It
is useful to give recommendations but that's all because there is not one
right relation to pick, as well as there are always plenty of new relations
to be constructed and the user should know best what fits his purpose. And
I'm sure that the outcome would be rich material to induce global trends
from. And these trends would naturally underlie change like our language
changes which is our foremost classification tool.
I haven't really used LT though I got an account but it sounds as if LT
would use this bottom-up approach, like Tim said:
> Instead of local items being pointers to the "real" global items, local
items are the real things, with the global level derivative of them.
I think that is the right approach. At best users can easily relate
different items, like Gerry suggested (in: [NGC4LIB] WebQubed: Web 2.0 Mixed
and Mashed): Relating score and sound, sound and text, text and picture,
picture and picture/variation, picture and movie etc. while commenting the
established relations and those of others.
Adrian
P.S.: Thanks to Nelson Goodman.
Received on Wed Feb 18 2009 - 08:07:54 EST