Last modified: 2011-12-16
Abstract
This paper, based on doctoral research, will consider the relative merits of two different visions of the purpose and potential of semantic technologies. The first part of the paper will summarize a number of trends in the development of Semantic Web, including those in cultural heritage and archaeology, and demonstrate that two divergent approaches have been adopted. The first, here dubbed Mixed-Source Knowledge Representation (MSKR), bears many similarities to the Artificial Intelligence and Knowledge Representation research of the 1990s, but takes particular account of Web technologies. The second, Linked Open Data (LOD), is also related to these movements but takes particular account of the philosophy of the Web, as espoused by Tim Berners-Lee. Linked Open Data is concerned with creating a relatively open and strongly interlinked network of resources, whereas MSKR more commonly attempts to locally integrate heterogeneous datasets. A 2010 survey of practitioners applying semantic technologies to Cultural Heritage suggests that both approaches have been influential in this domain but that there is a particularly high correspondence with whether a project is fixed-term (MSKR) or open-ended (LOD).
The second half of this paper will be a more in-depth discussion of the MSKR approach and in particular whether semantic technologies can provide beneficial economies of scale within a large but closed consortium. In order to do so, it is necessary to provide suitable and functional infrastructure for both the production and consumption of semantically-formatted data across a community with highly variable technical literacy. Furthermore, the results must arguably be a marked improvement over those that can be obtained from more traditional methods if they are to justify the additional effort from archaeologists. This paper argues that achieving such improvements is likely to be extremely difficult within a closed environment. In particular, it notes that many of the benefits of the Semantic Web rely on network effects which are necessarily curtailed by access restrictions and that the examples of implementation which are so crucial to an emerging technology are also reduced. In contrast, more lightweight approaches are demonstrably easier to implement and build communities around but offer fewer possibilities for inferencing. The paper will conclude by returning to the issues raised at the beginning. What is it that archaeologists (individually and collectively) intend to achieve when they employ semantic technologies and to what extent are they willing to make data their available in order to do so? Until greater clarity has been reached on these topics it may be difficult to evaluate the future potential of digital semantics to the archaeological community.