David Golumbia on Thu, 3 Jan 2008 13:52:27 +0100 (CET)


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: <nettime> Critique of the "Semantic Web"


Having written about these issues for quite a while, including in my
paper "Becoming-Encoded" at the 2006 MLA "Code" panel, I apologize for
having failed to check the list for a few weeks at the end of this
semester and missing this important discussion. I could not agree
more with Florian's perspective (indeed I think it is similar to the
perspective I offered in my paper and elsewhere), and that one can and
should go even further.

1) Where and how did the idea develop that something like the
"semantic web" COULD work? I argue that it stems exactly from
a powerful rationalist philosophy that walks hand-in-hand with
computers, and that it exists today for exactly the same reason Strong
AI once existed (and in some quarters still exists). This is most
apparent in the *goals* listed by Semantic Web advocates: they are
almost always examples of removing human decision-making from the
circuit of social decisions.

2) Along with the philosophical problems in building even relatively
general ontologies (as opposed to highly local ones, which certainly
do work, though often to the effect of *limiting* rather than
*increasing* the typical functions of the underlying data, there is a
profound question of language. In what language should tags--whether
on the surface or in the code--be written? So far they are largely
English, probably even moreso than most code is. Assuming that we can
cluster tags in different languages presumes the kind of universal
translation that will never happen (again, a subject on which I and
many others have written extensively). I would argue that there
is a profound desire on the part of many advocates--even if it is
unconscious--to keep the web "English-only," & that this ties to the
underlying computationalist philosophy. 3) What projects does the
Semantic Web displace? I think it precisely displaces the project of
making the web truly multilingual, a project that faces in almost
exactly the opposite conceptual direction from the SW project.

One more comment. I am very much in agreement with most of what Alan L
writes. But this comment requires response:


> criticism of the Semantic Web to end up confessing a utopianism
> of its own that is curiously similar to that being criticized in
> Berners-Lee et al. Today's cultural-critical dream is of a world
> in which there can be local cultures not enslaved by, indebted to,
> or otherwise under the shadow of the so-called "universal" culture
> of any dominating force, imperialist or Googlist. I don't see why
> the notion of an extensible ontology--where there is an overall
> framework for human interaction but not necessarily an overall
> meaning of humanity--is foreign to that dream. It is an extension of
> that

I would really resist equating these two "utopianisms," and Alan is
not the first person to equate them this way. One refers to what
may happen in a dream-like way (a typical utopia); the other tries
to characterize the way the world is now. Even if the latter is an
idealization, it is not utopianism. Utopianism refers to a world in
which significant problems have been resolved; critics of the Semantic
Web are simply describing today's world in its very problematic
nature.

I can give an example from linguistics, where I spend quite a lot of
time working. For quite a while the computationalists who are working
hard to computerize anything they can in linguistics wanted to create
something they called the "GOLD Ontology." The point of it was to
develop a series of universal, *flexible* terms that could accomodate
all language description. Linguists argued that even such apparently
central terms as NOUN and VERB have widely varying meanings across
languages. Project advocates responded that they weren't trying to
define these terms, but merely to come up with a shared vocabulary,
and that NOUN and VERB could be defined locally in whatever way the
linguist desired. But the linguists countered: then why call them
NOUN and VERB to begin with? It is already a problem in the "paper"
world that people assume NOUN and VERB are substantively identical
across languages. Computerizing these categories will make comparisons
even more likely despite their falsehood. If people want to develop
cross-linguistic typological comparisons for research purposes, as
many do, more power to them. But to install something like this at the
heart of linguistic research is directly to imply that there *is* a
universal categeorization scheme that applies to languages, which is
not merely conceptually tendentious: it is likely wrong. I think it
was the right thing for linguists to refuse the GOLD Ontology, not at
all from utopianism but from a careful examination of the world as it
is; and I am quite concerned that in other domains the utopian dream
of universalizing concepts and categories lives on quite powerfully,
though under the hood it is very much a dream of *changing*--not
*representing*--the world.

DG

--
David Golumbia
Assistant Professor
Media Studies, English, and Linguistics
University of Virginia

dgolumbia at virginia.edu
dgolumbia at gmail.com






#  distributed via <nettime>: no commercial use without permission
#  <nettime>  is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: http://mail.kein.org/mailman/listinfo/nettime-l
#  archive: http://www.nettime.org contact: nettime@kein.org