The framework of socio-epistemic networks defines knowledge networks as being composed of three different layers: the social network, the semiotic network, and the semantic network. The first is defined as the collection of relations involving individuals and institutions, which encompass any kind of structured organization of individuals. The semiotic network is defined as the collection of the material or formal representations of knowledge, which include a broad variety of entities, such as books, articles, journals, instruments, reports, but also particular institutions that might be considered, under particular circumstances, embodiment of specific knowledge elements. Finally, the semantic network is a collection of knowledge elements and its relations, where for knowledge element we identify the more abstract level of knowledge including, concepts, topics, research agendas and methods.
Current state of research
In the last years the research field of Historical Network Research (HNR) has emerged, which applies formal methods of Social Network Analysis (SNA) to corpora of historical data. The website "Historical Network Research" gives a comprehensive and up-do-date account of the development of this field. As an interdisciplinary approach, HNR has the potential to combine different research traditions: classical quantitative social research, in-depth historical studies as well as sub-areas of mathematics and computer science, such as graph and network theory. Previous publications have shown that HNR can lead to new results or allow to substantiate former qualitative results through quantitative analyses.
There is another branch of research in computer science and in the humanities where graphs and networks are of central importance: the semantic modelling of knowledge systems, which developed from early approaches of the semantic web. A growing body of data on cultural heritage is formally described on the basis of ontologies. In recent years, integrated systems that enable both input and data retrieval on this basis have been developed, notably WissKI and Research Space. Both support in particular the Conceptual Reference Model proposed by the International Council of Museums. These ontology-based descriptions of cultural heritage data are also being advanced in other areas, e.g. the Europeana Data Model for cultural objects data or the Encoded Archival Description for archival finding aids.
Various historical studies have shown that bringing together the two graph-related approaches, semantic modelling and network analysis, leads to new results in historical research. However, a major obstacle to the implementation of these methods in the humanities is their lack of transparency as well as poor performance and flexibility in some cases. Both are critical aspects for the typical research cycle in network modelling: New data leads to new assumptions, which, in turn, lead to new models and thus to new results. With current methods this cycle can only be implemented for historical case studies with great expenditure of time, which severely impairs the investigation of well-founded theoretical modelling approaches for historical processes. The problem of transparency is closely linked to the lack of a historical-critical apparatus for classifying the results of a formal network analysis and the resulting problem of acceptance of these results in the specialist community of historians.
A second fundamental issue is that rarely individual case studies can be related to other similar case studies, since the specific concepts and data formats cannot be mapped directly to one another. Typical examples of this are individual projects resulting in ego networks of people who are particularly relevant in order to reconstruct a specific research area. A lack of standards prevents projects that are in principle similar, from being merged, although this would be desirable from a research perspective, as this merging would enable comparative perspectives or the analysis over longer periods of time.
It is therefore necessary to develop at the same time historical-theoretical approaches, new mathematical methods and a digital workspace in which these can jointly be tackled, analogous to the established cooperation between experimental and theoretical research in the natural sciences.