Social Semantics

This book is an inquiry into representation. Given the almost impossibly wide scope of possible kinds of questions pertaining to representations, we will deploy an analysis that is simultaneously both historical and scientific by restricting our inquiry to an investigation of representations on the World Wide Web.

In this chapter we give a brief synopsis of the history of the Web, starting with Licklider through Engelbart to Berners-Lee. We then procede to dissect the various components of the World Wide Web in order to get an overview of Web architecture. The Web is distinguished from the Internet as a “universal information space” where all items of interest are named by URIs, while the Internet is a transport protocol for the transfer of bytes. Beginning with protocols to communicate between clients and servers, we explore twin notions behind information, namely encoding and content. Then, moving to the Web of URIs and web-pages, we arrive at a tentative theory of representations. Finally, we deduce from our exploration of Web architecture a number of informal principles.

In this chapter we review the Semantic Web, a project to extend the use of URIs on the Web to identify non-Web accessible things like real-world objects and imaginary concepts. We trace the history of the Semantic Web via the trajectory of classical artificial intelligence. Features of the Semantic Web’s most basic language, RDF, are explained as extensions of the principles of Web architecture. Finally, Web architecture is itself formalized as a Semantic Web ontology. Lastly, a comparison between artificial intelligence and the Semantic Web shows that it is the ‘Web’ in the Semantic Web, primarily the use of URIs, that is the main difference between the two projects.

In this chapter, we explore theories of semantics for the Web. In particular, we first inspect the source of the fission, what is known as the ‘Identity Crisis’ of the Semantic Web, which first became visible as an argument between Hayes and Berners-Lee. This problem has as its root the issue of how to determine what a URI means or refers to. We inspect the original theory of meaning, the theory of sense and reference by Frege. Then we move to theories of semantics that attempted to get rid of the notion of sense, namely logical atomism, and notice its similarity with certain interpretations of the formal semantics of the Semantic Web, even though the formal semantics of the Semantic Web makes no appeal to ‘sense-data.’ Then the famous counter-argument, the causal theory of reference of Kripke and its extension to natural kinds by Putnam is introduced. An extended parallel is shown between this Kripkean theory of the Web and Berners-Lee’s position is elaborated. Finally, we return to Frege’s notion of sense in search of a third tenable position.

This chapter puts forward the hypothesis that the Fregean sense of a URI can be constructed out of user-defined tags, or natural language terms applied to a web-page accessible via a URI using a ‘collaborative’ tagging site. We use the data from the social bookmarking site http://​del.​icio.​us to empirically examine the dynamics of collaborative tagging systems and to study how coherent categorization schemes emerge from unsupervised tagging by individual users, which can then be considered a ‘stable’ Fregean sense manufactured by a social consensus. First, we study the formation of stable distributions in tagging systems, seen as an implicit form of consensus reached by the users of the system around the tags that best describe a resource. We show that final tag frequencies for most resources converge to power law distributions and we propose an empirical method to examine the dynamics of the convergence process, based on the Kullback-Leibler divergence measure. The convergence analysis is performed both for the most utilized tags at the top of tag distributions and the long tail. It is assumed such a ‘power-law’ comes from the support the user in the tag selection process by providing tag suggestions, or recommendations, based on a popularity measurement of tags other users provided when tagging the same resource. So we investigate the influence of tag suggestions on the emergence of power-law distributions as a result of collaborative tag behavior. Although previous research has already shown that power-laws emerge in tagging systems, the cause of why power-law distributions emerge is not understood empirically. The majority of theories and mathematical models of tagging found in the literature assume that the emergence of power-laws in tagging systems is mainly driven by the imitation behavior of users when observing tag suggestions provided by the user interface of the tagging system. We present experimental results that show that the power-law distribution forms when tag suggestions are not presented to the users, and the power-law distribution does not hold when there are tag suggestions presented to the user. Looking to see if we can move beyond tagging as in our search for sense, we study the information structures that emerge from collaborative tagging, namely tag correlation (or folksonomy) graphs. We show how community-based network techniques can be used to extract simple tag vocabularies from the tag correlation graphs by partitioning them into subsets of related tags. Furthermore, we also show, for a specialized domain, that shared vocabularies produced by collaborative tagging are richer than the vocabularies which can be extracted from large-scale query logs provided by a major search engine. Although the empirical analysis presented in this paper is based on a set of tagging data obtained from del.icio.us , the methods developed are general, and the conclusions should be applicable across all websites that employ tagging.

This chapter explores the hypothesis that the terms from entire representations can be used as the sense of a URI. Furthermore, we show that this is not limited to Semantic Web documents, but extends to the hypertext Web. Based on this intuition, we investigate the possibility of using Semantic Web data to improve hypertext Web search. Using queries from a real-world search engine, we show that relevance feedback creates a ‘virtuous cycle’ between data gathered from the Semantic Web of Linked Data and web-pages gathered from the hypertext Web, and vice versa. Previous approaches have generally considered the searching over the Semantic Web and hypertext Web to be entirely disparate, indexing and searching over different domains. Semantic Web data can boost the performance of hypertext search, and vice-versa. From this empirical finding, we conclude that the sense of resource does can be composed of an entire representational nexus, from tags to Semantic Web data to hypertext.

This chapter presents the argument for the position of social semantics, where the meaning of representations on the Web is posited as both socially constructed and objective. The philosophical roots of social semantics are traced to Wittgenstein, in particular via the development of statistical methods of language understanding from Masterman and Spärk-Jones. Wittgenstein’s objections to both the causal theory and descriptivist theory of reference are explained. How an objective notion of sense leads to an understanding of representation as part of ‘extended’ things over time and space is given. The manifestation of an organically developing statistical Semantic Web is then given as the evidence of social semantics. Finally, the thesis of social semantics is extended in terms of determining the meaning of URIs on the Semantic Web. A tentative theory of collective intelligence then becomes our future work.