Examples of ecosystems of knowledge
Modeling digital archive interpretation
The research conducted in the field of digital humanities produces new archive sources that are challenging the division traditionally used by historians and the literati to distinguish between “primary” sources, those produced by the object of study, and “secondary” sources, those produced by research activity. The use of digital technologies leads to the creation of “secondary” archives in the form of databases that, if they are accessible and interoperable, automatically become new “primary” sources for a reflexive analysis of research activities or for other researchers studying the same field. The creation of these digital archives and, more specifically, the durable dimension of their use, conditions the researcher’s task by putting an emphasis on the formalization of the task in such a way that it becomes open, interoperable and lasting. This scientific imperative is imposed upon researchers more and more by the simple fact that they work on projects where the digital dimension is central, as it guarantees financing. The question then arises, how can this data be produced and made visible without being an expert in computer science or knowledge engineering?
Figure 1.3. Recursive cycle of sources
Muriel Louapre and Samuel Szoniecky aim to tackle this question by analyzing the task performed in the framework of the ANR Biolographes project. This very concrete terrain allows for examination of the nature of digital archives produced by research to extract the special features particular to the field of human science. After a presentation of the digital practices implemented in this type of research, the specific case of visualization methods is dealt with by a review of the primary tools available on the Web in order to critique the epistemological and practical limits. Using the same body of data, the authors show the utility of these tools for quickly testing the coherence of data, for visualizing networks, or for multiplying the approaches and defining new research perspectives. Finally, they reflect on a generic method for modeling influence networks using a prototype developed specifically to help researchers describe their interpretations so that they are interoperable with other perspectives. The goal of this process is to provide cognitive cartographies serving as an aid for the elaboration of a scientific consensus.
Figure 1.4. Mapping the influence networks. For a color version of the figure, see www.iste.co.uk/szoniecky/collective.zip
From these reflections emerges the result of a sometimes-difficult dialogue between researchers coming from different fields of expertise. Faced with the digital “black box”, digital models can be imposed upon researchers whose needs in terms of information processing are too often not explained concretely. Even if the lure of a button that can simply be pushed to obtain the relevant information starts to disappear after disappointments and frustrations during the dialogue with the machine, the lack of knowledge engineering training remains flagrant at times. Beyond knowing what the machine can do, it is important for humanities researchers who use digital technology to understand in what way they also bring reorganization to the collective task and research practices.