On this page you can download the discussion paper that was submitted for publication in the Journal for Media Linguistics. The blogstract summarises the submission in a comprehensible manner. You can comment on the discussion paper and the blogstract below this post. Please use your real name for this purpose. For detailed comments on the discussion paper please refer to the line numbering of the PDF.
Blogstract of
From Dyadic Reciprocity to Giant Interaction Engines. Posthumanist Perspectives on Interaction ‘Under Construction’
by Britta Schneider & Rita Tamara Vallentin
Machine-learning technologies are transforming not only communicative practices but also the very terms through which we analyze them. With the rise of large language models (LLMs) such as ChatGPT, the question emerges of how we should conceptualize terms like “interaction”. In our article, we draw on posthumanist theories to examine how central linguistic concepts may shift in the context of changing media ecologies.
Our starting point is the observation that language does not exist as a fixed structure but emerges within social and material practices. While cultures of literacy shaped linguistic theorizing through the linear and sentence-based logics of writing, digital infrastructures and machine-learning culture create new forms of understanding sign- and meaning-making. Language assemblages are changing: not only written and printed text but also algorithms, datasets, and infrastructures shape how language is understood and used.
At the center of our discussion lie the conceptual reconfigurations of the term “interaction” when machines appear to “speak”. Building on interactional approaches such as those of Goffman and Luhmann, interaction has traditionally been understood as entailing co-presence, reciprocity, and double contingency. In the context of human-machine encounters, however, these categories undergo transformation. Attributions of agency to machines can no longer be dismissed as mere projections by human participants; rather, they signal emergent forms of contingency that align with Esposito’s notion of “virtual double contingency”.
To explore these questions in practice, we turn to illustrative examples drawn from Reddit threads in which users describe their experiences with ChatGPT. Some construct the system as a rational collaborator that surprises them or prevents missteps. Others describe it as an empathetic partner that provides belonging, yet simultaneously as a kind of echo chamber where one’s own thoughts are mirrored. Still others note how ChatGPT’s ways of structuring arguments reshape their own communication practices. These examples reveal that interaction can no longer be described as a dyadic exchange between autonomous human actors transferring meaning, but as a process distributed across networks of data, algorithms, users, and infrastructures.
We therefore propose that interaction in machine-learning culture should be understood as part of a giant interaction engine: a vast, historically layered, and materially distributed assemblage of human practices, algorithmic operations, and economic interests. Within this assemblage, attributions of agency and interactivity are constantly negotiated. Our analysis shows that core linguistic concepts are themselves ‘under construction’. They do not exist independently but emerge in relation to the material, technological, and political conditions that shape language. Posthumanist linguistics provides a framework for recognizing these shifts: language and meaning are not stable entities but processes in motion, shaped by the assemblages in which they are realized.