Corinna Bath | Knowledge production for the Semantic Web: Dream of a universal language or user participation and openness?


The Semantic Web was first announced in 2001. As an approach to make Web content meaningful to computers, it invokes the vision of the intelligent, social machine. An essential prerequisite of making the meaning of text machine-understandable are classifications and ontologies, i.e. formal descriptions of “reality”, “the world” or certain domains.
In their analysis of similar projects science and technology studies (STS) researchers, particularly feminist scholars, have already pointed out that such formal approaches have politics. They produce inclusions and exclusions (e.g. Bowers 1992, Bowker/ Star 1999). Formal ontologies cannot represent embodied, skilled or tacit knowledge. They ignore minority views and quieter voices and allow the majority view to speak for everyone. Epistemologically speaking these approaches assume a consensus reality and a consistent representation of knowledge. They often presuppose that the knowing subject is male, white, middle-class etc. (e.g. Adam 1995, 1998, Sherron 2000). Others have shown that formal modelling approaches are lacking expressiveness for social processes, especially conflicts, change and fluidity (e.g. Crutzen 2000).
Thus, formal ontologies which seem to be the core infrastructure of semantic technologies are suspected of transforming Cartesian epistemology into IT. The knowledge represented in by new generation of the internet is therefore in danger of socio-materially reproducing existing structures of inequality, particularly the structural-symbolic gender order.
Contextualized in this way, the semantic web represents typical modelling approaches of computer science and software development. A closer look, however, reveals that this research & development has changed since its first proclamation in 2001. Today, concepts as well as technologies are not only based on traditional approaches to knowledge production, but also on social epistemologies (e.g. Web 2.0’s “wisdom of the croud”) or the promises of open, free knowledge adopted from the open source community.
In my contribution I will present some preliminary results of my ongoing research project on the semantic web as a technology-in-the-making in the tension of objectivist approaches and their counter-movements. I am focussing on questions such as: Which knowledge is coded in the semantic web? Which knowledge is rather excluded? Which meanings (semantics) are configured by formal ontologies? Who is assumed to be the knowing subject? What are the basic epistemological and ontological assumptions of the new infrastructures of the information society? And how do these questions relate to the existing gender order?
On the basis of theoretical considerations and empirical research within the field of semantic technologies research & development (e.g. expert interviews, participatory observation, conference participation) I will discuss the question whether the semantic web project reflects some of the criticism of the formalization approach expressed by STS researchers, philosophers and feminist scholars during the last decades. Does it, furthermore, take into account recent shifts towards the aim of user participation, openness and inclusion? Or is it rather an reinvention of the old dream of a universal language that in effect re-establishes power relations, exclusions and the existing gender order?