Introduction - Social Semantics: The Search for Meaning on the Web - page 4

the topic of representation, this thesis is not an attempt to develop a philosophy
of computation (Smith 2002), or a philosophy of information (Floridi 2004), or
even a comprehensive “philosophy of the Web” (Halpin 2008b). These are much
larger projects than can fit within the scope of a single topic or even a single
individual's life-long calling. However, in combination with more fully-formed
work in philosophy, we hope that at least this topic provides a starting point for
future work in these areas. So we use notions from philosophy selectively, and then
define the terms in lieu of our goal of articulating the principles of Web architecture
and the Semantic Web, rather than attempting to articulate or define the terms of a
systematic philosophy or with reference to the many arguments over these terms in
analytic philosophy. Many of the terms in this thesis could be explored much further,
but by virtue of our scoping are not explored, as to constrain the topic to a reasonable
size. Unlike a philosophical work, in this topic counter-arguments and arguments are
generally not given for terminological definitions, but instead references are given
to the key works that explicate these terms further.
This thesis does not inspect every single possible answer to the question of What
is the meaning of a URI? , but only three distinct positions. An inspection of every
possible theory of meaning and reference is beyond the scope of the thesis, as
is an inspection of the tremendous secondary literature that has accrued over the
years. Instead, we will focus only on theories of meaning and representation that
have been brought up explicitly in the various arguments over this question by the
primary architects of the Web and the Semantic Web. Our proposed solution of
social semantics rests on a theory of meaning, a neo-Wittgensteinian theory, that is
one of the most infamously dense and infuriatingly obscure theories of meaning.
Finally, while the experimental component of this topic has done its best to be
realistic, it is in no way complete. Pains have been taken to ensure that experiments,
unlike much work in the Semantic Web, at least uses real data and users, and
are properly evaluated over a range of algorithms and parameters. Our work on
tagging systems takes its data from a real system, del.icio.us , as well. While various
parts of the experiments could no doubt be optimized and scaled up still further,
these experiments should be sufficient to motivate our movement towards social
semantics, although a full formalization of such a theory and testing of it would
require access to the data of a large-scale search engine such as Google, which for
the time being is outside of scope. For future work, we would like to pursue the
formalization and large-scale testing of social semantics.

1.2 Summary

The thesis of this topic must be stated in a twofold fashion: first to analyze the
problem, and then to propose a solution. To analyze the problem of representation
on the Web, one must ask the question: What is the meaning of a URI? First,
we will clarify the problem that the Web is a kind of new language that can be
defined by its engineering conformance to the principles of Web architecture, but