Beyond Recognition: Towards Sonic Opacity as Algorithmic Resistance

The doctoral thesis Beyond Recognition: Towards Sonic Opacity as Algorithmic Resistance, presents an intervention at the intersection of critical theory, media studies, sound studies and artistic research that seeks to critically assess recognition as a regime traversing epistemological, ethical, aesthetic and technological domains. In particular, the thesis concentrates on listening as a sensory-specific mode of recognition; on human voice as the overdetermined object of auditory recognition; and on algorithms as contemporary technologies of recognition. Adopting a feminist, new materialist philosophical framework, Beyond Recognition understands recognition (in all of its modes) as a process that produces, rather than merely observes, its objects and subjects. Listening-as-recognition is thus decoupled from its purported neutrality, and exposed as contingent on ingrained racial, colonial, classed and gendered premises. The first Part of the thesis takes up the methodology of situating to offer a critical genealogy of recognition as an episteme sustaining its operations across philosophical, colonial, and computational apparatuses of capture. Throughout this Part, the thesis aims to show how recognition is uncritically inscribed in philosophical and aesthetic models, as well as in technological assemblages such as algorithms. In a contrapuntal theoretical movement, Beyond Recognition aims to propose opacity (Glissant) as a conceptual and performative tool able to actively resist recognition. The second Part of Beyond Recognition builds on the previous to substantiate it through a case study. This Part unfolds as a detailed critical audition of the “accent recognition” algorithm in use at the German Migration Office to process undocumented asylum seekers’ applications, doing so by engaging with artist and researcher Pedro Oliveira’s work. This critical assessment of the algorithm is grounded, on one hand, in media archeological methodologies and Actor Network Theory’s practice of script analysis; on the other, it proposes a second-degree interpretation of Oliveira’s own methodology, to highlight his radical and innovative approach to algorithmic audition, especially when confronted with the epistemological and practical limits of classic methods of analysis. In this case as well, opacity offers the conceptual and pragmatic not only to overcome this impasse, but actively resist to the voice recognition algorithm. The thesis concludes with an analysis of Oliveira’s live sound art work DESMONTE (2022), first, framing it in the broader context of contemporary artistic research practice on voice/recognition/algorithm, and lastly by adopting a participatory perspective as a privileged listener, to put to the test sonic opacity as a practice able to move beyond recognition.