Reto_Bachmann-Gmuer on Thu, 20 Dec 2007 21:48:26 +0100 (CET)


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

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


This is an OpenPGP/MIME signed message (RFC 2440 and 3156)

--------------enig675E20B39088A60C4F01A05D
Content-Type: text/plain; charset=ISO-8859-15
Content-Transfer-Encoding: quoted-printable

Dear Florian,

it's an interesting discussion as it seems to me that your critique of
the semantic web and my appreciation for it are based on similar
motivations. To me semantic web technologies are a tool to facilitate
the  decentralized construction of realities. A truly decentralized
semantic web isn't there yet but the technology has the potential to
obsolete centralized information selection. The vision is not an
Aritotelian universal classification system but the empowerment of
individuals by making the social ecosystem of Xenophanian web of guesses
more transparent.

I hope we can resolve the misunderstandings.

> My point actually was not that there would be a danger in the
> Semantic Web to confuse the two Nettimes (since avoiding such
> ambiguities as opposed conventional full text queries is its
> very design objective), but quite the opposite: That with its
> goal of unambiguous categorization, it reduces, or even fails
> to acknowledge, the cultural complexity of the phenomena it
> references.

On the semantic web names of resources are unambiguous: a resource
may have many names but there is no support for equivocal URI-names,
as equivocal labels can still be modeled (as property values) I don't
think this is a problem. The semantic web is not about unambiguous
categorizations but about means to talk about abstractions and
about linking abstractions to actual data. If the category "person"
comes in handy for usage in one of your current realities the Class
<http://xmlns.com/foaf/0.1/Person> might be useful in a graph
describing this world. Using the well-known foaf-ontology (rather than
creating your own class) increases chances that the merging of your
graph with another produces a result of greater value, if Santa Claus
is not part of your world, you wouldn't merge it with the graph at
http://www.jibbering.com/foaf/santa.rdf (a document comprised by the
"semantic web").

> "Nettime" is a good example because even in the case of the Nettime
> we're referring to here, the categorization "MailingList" is by no
> means uncontroversial. Nettime has also be called a "collaborative
> filter", and it has not only involved the mailing list, but also
> meetings, printed readers or, in the case of a work by Pit Schultz
> from 1997, an experimental slide projector screening in a Berlin
> techno club. Just recently, someone told me that she considered
> Nettime really a social network. In other words, there is no label,
> or no descriptive tags Nettime itself would agree upon, let alone an
> "ontological" categorization in relation to higher-order, synonymic
> or subordinate concepts or phenomena. 

I could hardly think of anybody understanding the
RDFS ontology and not agreeing that Nettime is a
<http://www.w3.org/2000/01/rdf-schema#Resource>. Personally I
consider Nettime as a ex:MailingList because I don't think the
Class ex:MailingList is disjoint with any other Classes you suggest
Nettime might be an instance of. But my goal is not to reach a global
consensus on this, but just to express my knowledge in a way that
people having similar pattern of interpretations could use it. The
beauty about the semantic web concepts (open world and graphs instead
of trees) is that it allows some use in merging graphs even if the
similarity between the pattern of interpretations is low.

> The Semantic Web, as your example shows, rests on the illusion
> that there can be common sense descriptions, while in real world
> semantics, you can't have meaning without viewpoint and cultural
> conflict. [A truism since Locke and Kant, and my reason to call the
> Semantic Web a scholastic project.] 

Centralized shared knowledge repositories offered by commercial
entities can streamline the interpretation models and foster
ontological convergence. This development is not specific to semantic
web technologies, it is true for wikipedia and for most so called web
2.0 applications. A truly decentralized Semantic Web will probably
not directly arise from these systems, but an environment of more
accessible data is not a hindrance to such development.

With the notion of anonymous resources the semantic has built-in
support for entities that have a contextual meaning by their extension
within the containing model. Especially commercial providers tend
not to support such anonymous (contextual) resources and treat
them as entities potentially spanning multiple models; I see this
as a danger as even if multiple names for the same resources are
possible it will give an advantage to the biggest player assigning
the most well understood names. In my opinion it is unfortunate that
Tim Berners Less promotes persons not being represented by such
anonymous resources[1] also his vision of a Giant Global Graph[2]
seems to reduce decentralization to the distribution of the data.
But the currently standardized semantic web technologies allow for a
construction of reality supported by the dynamic merging of multiple
non-coordinated RDF-Graphs. Currently these graphs are distributed
over the web, in future they might be over P2P systems.

> The argument that the Semantic Web can have multiple "ontologies"
> each of which could potentially tag and map Nettime in different
> ways seems to be moot if not an alibi;

This is how the semantic web looks like today.

> if you would map all the pluralities of meanings given to Nettime,
> you would end up just creating a one-to-one replica of the semantic
> mess that already exists in folksonomies and untagged texts.

Sure, the semantic web is not about simplifying the world, just about
making part of the complexity conceivable to software agents.

> And in comparison to concepts like, say, "god",
> "perversion", "culture", "madness" or "race", "Nettime" is even a quite=
> easy subject.


Even if such hard terms we can be more or less precise on what we are
talking about, in some case we refer to "the thing I call 'culture'
in this context" or we might refer to "the resource described by
<http://www.w3.org/2006/03/wn/wn20/instances/wordsense-culture-noun-2.
rdf= >".

Reto


1. http://dig.csail.mit.edu/breadcrumbs/node/71
2. http://dig.csail.mit.edu/breadcrumbs/node/215



--------------enig675E20B39088A60C4F01A05D
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.7 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHarMOD1pReGFYfq4RAqQcAJ4g5Qd3LNzoZ+wseQElZbS4aDvzJQCgz1iD
+gS5q+pUDaxWxeV8s3psAjE=
=Rk3G
-----END PGP SIGNATURE-----

--------------enig675E20B39088A60C4F01A05D--


#  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