Material-discursive performativity in software design: a sociopolitical approach

As research on the co-construction of society and technology and on the social shaping of technological artefacts and processes has highlighted social dimensions such as hidden assumptions on use contexts implicitly inform development practices. Moreover, software development takes place within organisations and therefore systems specifications and their implementation are co-determined by the organisational setting in which they are developed. Therefore, design decisions - although mediated by methods and tools of software engineering - represent the outcome of processes of negotiation and meaning construction; in this sense, societal discourses become operative in the development process as implicit assumptions and beliefs. Based in science and technology studies (STS), critical technical practice and research in software design, the aim of the project is to elaborate a socio-political approach to software design based on a theory of material-discursive performativity. The interventionist turn in STS increasingly involves researchers with practices of technology development and thus entails the need for appropriate methodologies. Thus, I will moreover conceptualize a deconstructivist design approach generating insights on the "intra-actions" between hegemonic social discourses and material practices in development processes. The projects` methodological approach integrates a deconstructivist technique and is therefore called "deconstructive design". Deconstructive design provides a tool to support cooperative work practices and provides a means to software development teams to disclose the implications of their implicitly agreed upon "ways-of-doing". It allows a team to make visible how structures such as in/formal hierarchies and discursive hegemonies affect their development processes and emerge with material phenomena and relations.

Computer programs are not neutral; they often reproduce social discrimination. In recent years, numerous cases have shown how algorithms, machine learning, and artificial intelligence objectify structural discrimination and reinforce injustice, sexism, racism, classism, and ableism. This project focused on how machines learn from language and everyday knowledge. Systems development incorporates social assumptions, values, and norms, reflecting inequalities and differences in society. Asking how computing practices embed social power relations and thus also co-emerge with them, I conducted ethnographic case studies in three computer science fields: 1) the development of Semantic Networks that help computers understand the meaning of everyday language; 2) research on Fairness in Machine Learning, which is concerned with removing bias and discrimination from systems; 3) the field of Critical Requirements Engineering, which is a pivotal point where social perspectives get translated into technical specifications. This research provides empirical insights into the intertwining of knowledge practices and social power relations in the approaches, concepts, and methods in these fields. Trying to teach computers human-like understanding, the first two work on building a 'semantic infrastructure': a sort of 'meaning-centered' foundation that relates concepts or entities, such as data from text, speech, and images. This infrastructure is already the result of machine learning techniques and again feeds resources into further (semi-)automated processes. Based on the empirical results, I developed the concept of 'infrastructural power'. It describes the entanglement of historically grown, structural power relations and their emergent materializations in sociotechnical systems. Furthermore, the project contributed to the computer science fields under study by developing a methodological approach that makes visible their implicit values and scientific norm-setting: the deconstructive method 'Mind Scripting' supports computer scientists and developers to reveal the social antagonisms that obscure the political relevance of their work practices, methods, and concepts. Thus, it develops collective, interdisciplinary agency for the emancipatory transformation of society-technology relations. The project linked research on the co-emergence of society and technology with critical computing and approaches to power from political science. Theoretically, it tied in with the concept of 'material-discursive performativity' suggested by Karen Barad. This concept explains how power relations reproduce in everyday practices through the intra-acting agency of social discourses and material (e.g., technical) phenomena and was brought into conversation with approaches of ideology, hegemony, and affect.