Unlocking Historical Media Dynamics — impresso’s groundbreaking corpus offers unprecedented opportunities for advancing data-driven historical research. By providing an unparalleled view of content, formats, and ideas across media in a transnational context, scholars can now reconstruct historical media ecosystems across modalities, time, languages, and national borders, relying on data derived from published content.
Media must be read as an “active construction of collective representations” (Kaenel & Vallotton 2021), where selection, prioritisation and form of content are the result of processes of choices and negotiations. They are active constructions in the sense that they do not strictly reflect reality but produce representations of it which are affected by journalists’ points of views, editorial lines and external influence. To understand these processes, we will look at media as an organised ecosystem that interacts with its environment. This term, which we borrow from ecology, neatly describes two aspects that we want to study through the lens of published content: the internal organisation and functioning of the media, and their interactions with their environments.
Several conceptually and methodologically complementary case studies help guide the project. They are united under the overarching theme of “influences”. We define influence as the presence and circulation of traces left by individuals, groups, values and mechanisms for the development and dissemination of news. Each case study will investigate influences from four perspectives: external to the media, within the media ecosystem, within a single media institution, and between content formats.
Our approach to studying influences from a transnational and transmedia perspective contributes a comparative and content-based view of the classic question of media autonomy, advance research on the functioning of media ecosystems by analysing their content and by revealing the mechanisms of production and circulation of information, and emphasise interactions and transfers between radio and newspapers, usually studied in light of their competition.
Historians have long mastered the art of making meaningful comparisons based on incomplete archives and in consideration of different source types. Comparisons, however, do not merely serve to enumerate differences and similarities. Rather, they make visible the idiosyncrasies of what is being compared and help make sense of the observations. But how do we translate this practice to a transmedia corpus of heterogeneous source types connected by semantic enrichments? How can we operationalise our goal to trace influences in historical media in transnational perspective? How can we conduct comparisons at scale?
We raise these questions against the background of a critical perspective on algorithms, data, tools and interfaces. Our sources are diverse but similarly enriched, and thus provide a basis for comparative approaches: named entities, topics or detected content reuse offer unparalleled versatility to shift between large and small scales, to compare over time and across media, national boundaries and languages, and to shift between intra-, inter- and trans-media perspectives. But comparing does not mean flattening. To assess whether a pattern is actually meaningful, we need to establish a dialectic relationship between historical heuristics and interpretation, and the computed measures of relatedness and similarity.
To this end, we develop a comparative framework that allows us to operationalise the detection of influence traces and how they are distributed: we do so from a critical digital hermeneutics perspective which exploits the opportunities and addresses the challenges inherent in our data, tools and interfaces, and we will evaluate our findings to ensure their validity.
Collaborative data-driven research
impresso’s corpus is the first of its kind and offers many hitherto unexplored opportunities to advance data-driven historical research. Aligned with digital hermeneutics, we pioneer collaborative, data-driven historical research. Leveraging impresso data lab’s API and Jupyter notebooks, we implement transparent, reproducible research practices. Publishing adaptable building blocks and templates aims to democratize accessibility. For both novice and experienced researchers, these tools facilitate collaborative work, ensuring transparent, reproducible, and versatile research standards. Our vision is to build a wider community of researchers around the impresso system.