Algorithmic power and scientific knowledge

This paper critically revisits traditional perspectives on technology within academic and scientific writing studies. It aims to comprehend the intricate, emerging, and dynamic sociotechnical configurations that underlie contemporary scientific practices. These practices increasingly involve language, text, and literacy practices, seen as products of the collaboration between humans and machines. The paper draws on empirical research on influential institutional metadiscourses in high-impact scientific writing produced and/or disseminated by public universities and a research institute in the State of São Paulo (Brazil), whose local policies of globalization are driven by international university rankings. I use a qualitative content analysis approach grounded in socio-anthropological, socio-semiotic, and pragmatic studies of linguistic ideologies to shed light on how ideological and socio-semiotic processes support the metapragmatics of scientific writing in university policy documents. This metapragmatics is utterly alien to the role of performative sociotechnical infrastructures in the production, distribution, and hierarchization of scientific texts. Additionally, these documents do not account for the diverse conditions and restrictions that shape the production and circulation of academic knowledge in geopolitically marginal and equally diverse regions within the country, including those within São Paulo.