The ModelSEN Framework: Current Insights and Future Directions

In the ever-shifting scholarly landscape of the history of knowledge, a fundamental question is and has been: how does human knowledge evolve? In the poster, we aim to present the framework of socio-epistemic networks (SEN), intended to explore the dynamics that govern the formation and evolution of knowledge systems. It combines network analysis and collaborative inquiry approaches with questions about knowledge formalisation, dissemination and acquisition of interest to historians and sociologists of knowledge and science. The framework is meant to be used with empirical data, but it can and has been extended with formal models, such as simulations and agent-based models.

The ModelSEN team at the Max Planck Institute for the History of Science in Berlin has been developing the SEN framework and the associated digital methods (ModelSEN Project 2023). Continuing the work begun by Jürgen Renn and others in the early 2010s, seeking to create a quantitative, network-inspired approach to the history of knowledge that integrates micro- with macro-history, accounting for global, transregional effects of knowledge processes (Renn et al. 2016; Lalli et al. 2020).

In the poster, we will present infographics explaining the underlying theoretical assumptions and the principal ideas behind SEN. ¹ Renn et al. define knowledge as a “codified experience and potential for problem-solving” (Renn et al. 2016: 16) Codification occurs through the formation of cognitive, material, and social structures. For example, the processing of experiences into mental models, but also through their external representations, such as the formation of sign systems both within a realm of communication between actors. Accordingly, cognitive, material and social dimensions of knowledge emerge. These dimensions can be modelled by multi-layered, time-evolving socio-epistemic networks: the social, the semantic and the semiotic/material. As a framework, socio-epistemic networks build on the conceptualisation that knowledge is culturally and physically embedded, aim to uncover structural changes in the constructed networks and reflect on these changes from both quantitative and qualitative historical perspectives.

Social networks are ones whose nodes represent individual or collective actors and whose edges represent interactions between these actors. In the context of knowledge systems, these nodes and edges revolve around knowledge-related interactions like collaborations, being colleagues, attending the same conference or being supervisor/supervisee. Analysis on this layer primarily uses tools drawn from Social Network Analysis (SNA). Semiotic networks are those whose nodes represent material objects or externalised representations and whose edges represent interactions or physical transformations between these objects, like publications that cite each other. These objects often serve as the starting point for constructing other levels of SENs simply because historical events are typically accessed through archival materials. Methods from Bibliometrics are commonly used to analyse this layer. Semantic networks comprise nodes representing cognitive elements, while edges represent cognitive operations or interactions among these elements. These can be concepts, mental models, or other cognitively linked entities, but as with the other two layers, the network is understood as the structural carrier of knowledge rather than the individual element. Methodologically, approaches from semantic network research or computer linguistics are mostly used in analysing this layer, meaning operationalisations like co-occurrence networks (Segev 2021), quantum semantics (Koponen / Södervik 2022) or approaches using word/sentence/document embeddings (Bizzoni et al. 2020).

These three interconnected layers of the SEN framework and their operationalisations will be illustrated and explained with examples in further detail on the poster. We will also make clear in the poster that there are constraints of knowledge-related interactions on each of these layers governing the formation of structures. We will show that the proposed framework has a strong hermeneutic function because representing knowledge at different network levels structures research hypotheses and interpretations, making them transparent, verifiable, and open to scrutiny.

Furthermore, we will call attention to the future directions we see for the SEN framework by showcasing its strengths, limitations and potential applications. We identified several compatible methods for extending research outputs produced through the application of the SEN framework, which we will highlight in the poster. ²