The Space Between

Web 2.0

Web 2.0 is a primarily commercial term used to describe the user-friendly Web that has emerged since the beginning of the twenty-first century. This is the Web which comprises blogs and social networking sites that allow their users to have an impact on the layout of the site. A blog reader, for example, can comment on the author’s posts. A social networking site, like MySpace or Facebook, is built from user contributions; the pages designed and maintained by readers make up the whole of the site. But one of the most important developments for Web 2.0 had yet to be fully implemented: the Semantic Web.

The Semantic Web will affect how users, those ubiquitous “readers” of the Web, interact with the online world, giving them far more control over their reading experience and even endow them with some of the powers of textual organization traditionally given to authors. One of the biggest improvements will come in the world of search engines. Modern search engines compare hypertexts based on the words present in them, notably just the text and not multimedia elements like photographs or videos. Ideally, the more words the two hypertexts have in common, the more the documents themselves will have in common (Green), but this sort of comparison does not account for incidences of polysemy and synonomy. Polysemy is a single word that has several meanings, and synonomy denotes two words with similar meanings. Basically these are the sort of things educated readers would catch through contextual knowledge of their respective languages, but it is the kind of thing non-intelligent computers have an extremely hard time doing. Both of these phenomena “disrupt the simple identity relationship” that information retrieval programs use to determine document likeness (Green 1).

The designers of the Semantic Web hope to build “lexical chains” of words that have similar meanings in the context of the document. For example: vim, vigor, zip, élan, pep, and vitality might be classified by semantic software as synonyms for life. With this lexical chain, a search engine could find documents that are more closely related to one another. Green describes the task of building semantic links between documents as “trying to discover the inter-textual cohesive relations” (3). Hence the entire lexical chains belonging to the documents will be similar rather than just one or two similar words. The hope is that the meanings of the two hypertext pages will be similar too.

This works thanks to tags that computer networks can read (and search) which describe the content of pages or images or videos. For semantics to classify and organize the entire Web, a huge number of tags will need to be written, and the most readily available source of labor is readers. Though this sounds like a lot of work and responsibility, it is already happening on sites like YouTube, Technorati, and Del.icio.us. YouTube users classify their streaming videos by entering keywords when they are uploaded. The same is true for blog content on Technorati and shared Web bookmarks on Del.icio.us. In all cases, the method for implementing the Semantic Web is not to release it officially like HTML, which you may remember is governed through a central lawmaking body, but to allow popular Web sites to build up databases of tags slowly. As readers read, comment, and upload content to Web 2.0, they are also helping to organize that giant Technicolor Dreamcoat of a hypertext system.

Doubtless, people will disagree over appropriate classifications, and there is also the danger that lexical chains of tags could be arbitrarily created by simply linking a hypertext to an online thesaurus. That would provide an efficient source of synonyms but a source lacking in the depth an empowered reader could bring to the hypertext. Could, for example a lexical chain created by software explain the word pharmakos as Derrida uses it in Disseminations? Automated procedures might provide some context, but it is doubtful the Semantic Web’s computer elements could make the same sort of links human readers can.