Intra-linking the Research Corpus: Using Semantic MediaWiki as a lightweight Virtual Research Environment

In recent years, virtual research environments (VREs) emerged as a topic referring to the established research field in the digital humanities: enabling research practices with digital tools. First projects in this area are realized and discussed by the community (Carusi 2010; Dunn 2009; Neuroth et al. 2009). In the humanities, researchers point out that the so-called ‘data deluge’ (Hey et al. 2003), which has influenced several national and supranational information policy agendas in the sciences, does not cover the full range of aspects of research practices in the humanities. While digital libraries and archives offer a new plurality of research resources in the humanities, the ‘complexity deluge’ (Dunn 2009) formulates an opposite agenda addressing the sometimes fuzzy, interfering and dispersed practices of humanities research.¹In this paper we want to address this tension between research data, metadata and collaborative action in the design of research corpora carried out within a VRE. Therefore we will focus the ongoing corpus re-arrangement and the potentials of Social Semantic Media technologies to expand metadata creation and use in qualitative research. Tools for qualitative research target the flexible coding system, i.e. allowing researchers to annotate research resources according to a classification system that may evolve over time. However, the tools have been criticized for the limited metadata interoperability of resources and research findings (Corti et al. 2011). In our paper, we outline the aspects of interoperability in qualitative corpora research and focus on researchers’ capabilities to intra-link the corpus. We use the term intra-linking²to address a main aspect in qualitative research: to create and to describe entities while allowing for the ongoing re-arrangement of entities and their properties in the research process. These capabilities will be discussed and exemplified within the scope of the project Semantic MediaWiki for Collaborative Corpora Analysis (SMW-CorA) which aims to reconfigure Semantic MediaWiki as a lightweight virtual research environment.³

The field of corpora centered research in the digital humanities offers interesting insights into the design of VREs. In the early 1990s, Biber pointed out main aspects of corpus design by problematizing a priori determinations of its boundaries and formal specifications. He recommends the selection of relevant objects and the formal description to be realized as a cyclic or iterative process of corpus work (Biber 1993: 256). While a linguistic approach mainly aims at a statistical ‘representation’ in relation to a target population, qualitative corpus research, which is focused here, pursues a so called qualitative selection, i.e. a typification of yet unknown properties in research (Bauer 2000: 20). We argue that this indeterminacy of entities and properties in qualitative research emphasizes the affordance of a VRE enabling researchers to intra-link the corpus – it means to give them the ongoing capabilities to create, modify and re-arrange entities and properties while doing research. This topic of qualitative corpus research addresses the research and design desideratum of qualitative annotations (Juola 2008) and a demanded shift to further capabilities for the researcher to control the data (Smith 2008: 178).

While the SMW-CorA project targets the re-use of its VRE infrastructure in different research contexts in the mid-term, its initial design is subject to a co-operation with a research project in the history of education, involving a major library in this field. The educational research project encompasses the analysis of a corpus of 25 educational lexica dating from 1774 to 1942, reconstructing the development of educational science. Discourse, field and content analysis are supported and applied to grasp the networked relationships in this scholarly field. The Library for the History of Education (BBF) hosts a large part of the lexica at the digital library Scripta Paedagogica Online (SPO).⁴ The collection amounts to nearly 22,000 articles. Each lexicon is bibliographically described and accessible online as image files. Participatory and agile design approaches are used to offer an adequate shared space for the different stakeholders to articulate possible potentials and boundaries of the VRE.

The SMW-CorA project builds on a Social Semantic Media technology which in turn is based on the Web 2.0 software MediaWiki (MW),⁵ which is used at the well-known online encyclopedia Wikipedia, and the extension Semantic Media Wiki (SMW).⁶ The latter enables the use of semantic annotations and integration within the Semantic Web through import and export of semantic data and linking to external entities. While other VREs⁷ have been realized using this wiki technology, the SMW-CorA project promotes research on a corpus by configuring the facilities of MW and SMW and by developing further extensions to offer a configurable and lightweight VRE.

As such, the basic framework MediaWiki offers some facilities for creating and typifying entities in a corpus for research (in our case: lemmas, authors, institutions etc.). The online encyclopedia Wikipedia demonstrates, besides the collaborative creation of texts, the capability to store documents and digital objects. Therein a wiki page can be related to other entities by hyperlinks and arranged within a hierarchical category system. SMW extends this by offering an increased granularity for describing and linking the corpus by adding metadata descriptions used in the Semantic Web. The unspecific hyperlinks between entities can be typified by metadata and thereby entities can be enriched with attributes or semantic relations to other entities. Furthermore, it is possible to import and export metadata in the Semantic Web standard RDF.

Within the scope of the project a set of use-cases is explored comprising the scholarly work in the research life cycle from importing research resources, coding, classifying and analyzing these to the export for re-use. The supported and envisaged use cases have in common that entities can be integrated, created, modified and interlinked (i.e. intra-linking). This functionality is focused within the project and supported by tools to enable researchers to carry out an ongoing re-arrangement of their corpora. While the aspects of importing lexica from a digital library and of exporting the bibliographic data in RDF are discussed in (Schindler et al. 2011), further use cases exemplify these capabilities. Besides this import functionality, the enrichment of entities such as editors, authors, and related affiliations with properties by using Semantic Web Browsing technologies is a further example.⁸ Thereby, semantic properties from digital archives, libraries or further collections can be integrated and adapted to the locally used metadata schema in the VRE. This enables researchers for example to add biographical data of authors or editors from authority files (e.g. German National Library GND). Furthermore the import and collaborative development of taxonomies (e.g. classification or coding schema) are supported by interlinking entities within the VRE and linking entities into the Web of Data. These links to the world external to the corpus, together with the data export facilities of the VRE, enable the reuse of the content created within the VRE. Additionally, the VRE provides functionalities for creating and re-arranging metadata schema as well as a bottom-up task management to allow supervision of the research process.

To summarize, we identified the need of researchers to create, manage and intra-link entities and metadata objects in a research corpus. This functionality is relevant for multiple use cases where researchers perform a qualitative analysis on a corpus of resources such as digital/digitized documents and images. Our main contribution is the development of a lightweight collaborative and adaptive VRE that enables researchers to perform these tasks as well as to enable export of the created data and the content’s sharing and reuse. Since the VRE is based on a flexible Open Source platform it can be tailored by the researchers towards their specific needs. Therefore this lightweight environment may serve as a starting point for further re-uses and re-configurations in unforeseen research settings and functionalities required in the future.

Funding

This work was supported by the German Research Foundation (DFG) in the domain of ‘Scientific Library Services and Information Systems’ (LIS). The funded project is entitled: ‘Entwicklung einer Virtuellen Forschungsumgebung für die Historische Bildungsforschung mit Semantischer Wiki-Technologie – Semantic MediaWiki for Collaborative Corpora Analysis’ [INST 367/5-1, INST 5580/1-1]

References

Barad, K. (2003). Posthumanist Performativity: Toward an Understanding of How Matter Comes to Matter. Signs: Journal of Women in Culture and Society 28(3): 801-831.

Bauer, M. W., and B. Aarts (2000). Constructing a research corpus. In M.W. Bauer and  G. Gaskell (eds), Qualitative researching with text, image and sound: a practical handbook. Thousand Oaks: Sage, pp. 19-37.

Biber, D. (1993). Representativeness in corpus design. Literary and linguistic computing 8(4): 243-257.

Carusi, A., and T. Reimer (2010). Virtual Research Environment. Collaborative Landscape Study. A JISC funded project. http://www.jisc.ac.uk/media/documents/publications /vrelandscapereport.pdf (accessed 30 October 2011).

Corti, L., and A. Gregory (2011). CAQDAS Comparability. What about CAQDAS Data Exchange? Forum: Qualitative Social Research 12. http://www.qualitative-research.net/index.php/fqs/article/viewArticle/1634. (accessed 30 October 2011).

Dunn, S. (2009). Dealing with the complexity Deluge – VREs in the arts and humanities. Library Hi Tech (27)2: 205-216.

Edwards, P. N., M. S. Mayernik, A. L. Batcheller, et al. (2011). Science friction: Data, metadata, and collaboration. Social Studies of Science 415): 667-690.

Hey, T., and A. Trefethen (2003). The Data Deluge: An e-Science Perspective. In F. Berman, G. Fox, and T. Hey (eds.), Grid Computing: Making the Global Infrastructure a Reality. Chichester: John Wiley & Sons, pp. 809-824.

Huvila, I. (2008). Participatory archive: towards decentralised curation, radical user orientation, and broader contextualisation of records management. Archival Science 8(1): 15-36.

Juola, P. (2008). Killer Applications in Digital Humanities. Literary and Linguist Computing 23: 73-83.

Neuroth, H., F. Jannidis, A. Rapp, and F. Lohmeier (2009). Virtuelle Forschungsumgebungen für e-Humanities. Maßnahmen zur optimalen Unterstützung von Forschungsprozessen in den Geisteswissenschaften. Bibliothek. Forschung und Praxis 2: 161-169.

Schindler, C., C. Veja, M. Rittberger, and D. Vrandečić (2011). How to teach digital library data to swim into research. Proceedings of I-SEMANTICS 2011: 7th International Conference on Semantic Systems, Sept. 7-9, 2011, Graz, Austria New York: ACM International Conference Proceedings Series (2011), 142-149

Smith, N., S. Hoffmann, and P. Rayson (2008). Corpus Tools and Methods, Today and Tomorrow: Incorporating Linguists’ Manual Annotations. Literary and Linguist Computing 23: 163-180.

Notes

1.It should be mentioned that a similar field of tension is articulated for the sciences as ‘science friction’ (Edwards et al. 2011) by addressing the problems of different disciplines working on the same phenomena and trying to interoperate.

2.This term refers to the concept of ‘intra-action’, which describes the interrelations and re-configurations of research apparatuses in respect of ‘locally stabilized phenomena’ (Barad 2003: 817).

3.This project is realized in co-operation between the German Institute for International Educational Research (DIPF), the Karlsruhe Institute of Technology (KIT), the Library for Research on Educational History (BBF) and educational researchers mainly of the Georg-August-University Göttingen. See http://www.dipf.de/en/projects/virtual-research-environment-for-research-in-the-history-of-education-smw-cora.

4.The lexica are available at http://bbf.dipf.de/digitale-bbf/scripta-paedagogica-online/digitalisierte-nachschlagewerke.

5.See http://www.mediawiki.org

6.The extension Semantic MediaWiki is described at http://www.semantic-mediawiki.org and offers further extensions at http://semantic-mediawiki.org/wiki/Help:Extensions.

7.Some interesting cases using MW or SMW as a VRE are http://www.docupedia.de/, a reference work in the area of historical research, http://wiki.digitalclassicist.org and in the context of archives (Isto, 2008). Some further examples of using SMW are listed on the webpage http://smw.referata.com/wiki/Special:BrowseData/Sites?Data_type=Science.

8.A prototype of this SMW-based extension Semantic Web Browser is developed in collaboration with KIT – Benedikt Kämpgen, Anna Kantorovitch, and Denny Vrandečić – and is accessible at http://www.mediawiki.org/wiki/Extension:Semantic_Web_Browser.