For a long time, the field of education has been grappling with a ‘crisis of representation’, i.e. the challenge to articulate and ‘teach’ non-declarative forms of knowledge (Eisner, 1997). These debates began by exploring the relationship between epistemology and experience (Dewey, 1909), but evolved over the years to consider the tensions between paradigmatic and narrative forms of knowledge (Bruner, 1985), and the often-tacit ‘nous’ of epistemic communities (Lave & Wenger, 1991). The advent of new forms of educational governance and delivery - technology-assisted, aggressively comparative,  data-based and increasingly granular - has compounded this crisis by exposing a widening rift between quantifiable epistemologies, grounded in an individualistic understanding of learning as psychological and rational-economic, and more humanistic, socio-historical and holistic perspectives on education (Biesta, 2015).

How is the praxis of education (pedagogy) changing in response to these trends? In this article, I aim to identify the socio-technical influences that are contributing to the emergence of what I term ‘panspectric pedagogy’. In proposing this term, the article brings together critical perspectives in contiguous areas of scholarship: critical studies in education, critical algorithm studies in education and research on data-based forms of educational governance.  The rationale underpinning this exercise is that whilst much has been written in recent years about the multiple, problematic intersections between data, digital platforms and education governance and policy (Decuypere et al., 2014; Hartong & Förschler, 2019; Landri, 2018; Williamson, 2017), the topic of pedagogy as a central concern in educational research has been somewhat neglected. The question of whether a genuinely new form of praxis is emerging because of algorithmic logics of surveillance and control is worth pondering in educational research, and I hope that this article will convince others of the urgency of such a discussion.

Among philosophers, education has always been an explicitly ethical and normative project concerned with the practical moulding of personal agency and civic participation, that is, a situated process of initiation into a liberal and just society. Traditional as well as critical understandings of pedagogy were profoundly shaped by these concerns. Dewey’s pedagogy, for example, revolves around a notion of individual self-realisation in a western-style democracy, while placing a great emphasis on the experiential dimension (Dewey, 1998). For Dewey, the ideal learner is a free socio-historical agent operating for the greater good.  Another relatively popular ‘theory’ of pedagogy is Paulo Freire’s Latin-American rendition of humanist Marxism. His ‘pedagogy of the oppressed’ is rooted in materialist relations of dominance and subjugation and a great emphasis is placed on emancipation through class conscience and grassroots agency (Freire, 1996). For Freire, learning is always political and all those involved - students, teachers and educational researchers - are members of local collectives motivated by solidarity and the pragmatic pursuit of justice.  The threads running across such diverse yet related understandings of pedagogy are an overt focus on the ethical and axiological aspects of education, coupled with only a mild interest in instructional procedures. These ideas have surely influenced scholarly debates but were largely drowned under a coordinated ideological tsunami which, as the world transitioned to the current neoliberal order, turned education into a project entirely beholden to the dictates of utilitarianism and economic rationality.  

Against this background, this article presents a theoretical examination of pedagogy in the context of datafication. Pedagogy is treated as a broad conceptual dimension that incorporates a focus on learning (the main theme of this Special Call) but from the perspective of practice, agency and labour, that is, the socio-historical experiences of teachers and educators more broadly.

The article is conceptual in nature. As such, it undertakes an examination of the literature to propose a categorisation of the sociotechnical dynamics that are shaping an emergent form of pedagogical practice at the intersection of algorithmic surveillance and automation. The three dynamics are as follows: i) The reconfiguration of the pedagogical space-time. Drawing on the always-productive metaphors of spatialized thought used by such authors as Lefebvre (Lefebvre & Nicholson-Smith, 1991) and Harvey (Harvey, 2006), it could be argued that digital platforms and algorithms act as ‘spaces of representation’ that bind and reproduce practice. In the spatial-temporal framing of algorithmic pedagogy, the twin processes of computational imagination and abstraction (i.e., the re-imagining of social processes as enumerable and quantifiable) create symbolic fields where the reach of words and actions is bounded, and their meanings channelled along predefined routes. ii) The reconfiguration of pedagogical participation: as learning takes place across the multiple spaces/times generated through algorithmic abstraction, the interactive and epistemic demands placed upon individuals increase exponentially, rapidly leading to pedagogy by proxy, as multiple interactional tasks (delivering content, communicating feedback, coordinating interactions, resolving conflicts, etc) are offloaded onto automated systems. These systems then regulate teaching and learning unbeknownst to all but the most technically minded. This is the educational equivalent of the milieu-based governance described by Andrejevic (Andrejevic, 2020) and informed by the Foucaldian notion of environmentality: the development of surveillance infrastructures where ‘action is brought to bear on the rules of the game rather than on the players.’ (Foucault et al., 2008, pp. 259-260). Unlike more traditional forms of disciplinary surveillance, milieu-based surveillance is based on a compenetrating and mostly invisible web of normative and discursive assumptions endowed with implicit political force and underpinned by a tacit form of rationality. Its goal is to surveil and control not through discipline (external or internalised), but by modulating the contexts in which subjects operate. iii) The reconfiguration of pedagogical infrastructure: digitisation and the meteoric yet unstoppable rise of platformisation changed how we understand infrastructure and its role in shaping pedagogy. Education has always been supported by infrastructural arrangements: physical schools, curricular standards and progression protocols. However, educational infrastructures have been altered by the logics of digital platforms which have aggressively colonised social life. By reconfigured pedagogical infrastructure, I therefore refer to the sprawling collection of technological properties, normative discourses, governance principles and ideological leanings that, simultaneously, bracket and extend pedagogical practice (Perrotta et al., 2020).

Drawing on the critical discussion and the resulting categorisation – both summarised in the previous sections - this article introduces the notion of ‘panspectric pedagogy’. The term is not entirely original, as it borrows from Manuel De Landa who coined ‘panspectron’ in 1991 to describe the environmental and ambient surveillance enabled by distributed networks of sensors and AI systems (De Landa, 1991). Compared to the relatively more popular panopticon, where the logics of control rely on assumptions of totalising visibility (the Foucauldian ‘gaze’), the panspectron is based on multiple detections which are captured by a ‘ubiquitous quantitative tracking system built into the very infrastructure.’ (Bolin & Andersson Schwarz, 2015, p. 3). In this sense, panspectric logics and their attendant algorithms are no longer in the realm of the visible but operate instead at a sub-observational scale. In conclusion, I argue that a panspectric pedagogy creates a new time/space for agency rather than eliding it altogether. In this new context of milieu-based surveillance, we are witnessing the co-existence of face recognition, algorithmic behaviour control (e.g., analytics for suicide and bullying prevention), comprehensive device management (e.g., the ability to monitor usage and establish ‘healthy media habits’ by proxy). In this novel infrastructural arrangement, AI-assisted processes of operational offloading and environmental surveillance nudge educational agency and learning towards algorithmic alignment and, cognitively, they undermine the ability to nurture forms of knowledge not amenable to enumeration and standardisation. The capturing of teachers’ labour and students’ learning within an algorithmic web of expectations, standards, routines and other implicit or explicit regularities is not however the full story. Within this web, there are openings for resistance and misalignment, but even these are regulated and ultimately made possible by the algorithmic reframing of the pedagogical experience.

Andrejevic, M. (2020). Automated media. Routledge. Biesta, G. J. (2015). Good education in an age of measurement: Ethics, politics, democracy. Routledge. Bolin, G., & Andersson Schwarz, J. (2015). Heuristics of the algorithm: Big Data, user interpretation and institutional translation. Big Data & Society, 2(2), 2053951715608406. https://doi.org/10.1177/2053951715608406 Bruner, J. (1985). Narrative and paradigmatic modes of thought. Teachers College Record, 86(6), 97-115. De Landa, M. (1991). War in the age of intelligent machines (Swerve eds. ed.). New York : Zone Books. Decuypere, M., Ceulemans, C., & Simons, M. (2014). Schools in the making: mapping digital spaces of evidence. Journal of Education Policy, 29(5), 617-639. https://doi.org/10.1080/02680939.2013.865081 Dewey, J. (1909). Moral Principles in Education. Project Gutenberg. http://www.gutenberg.org/ebooks/25172 Dewey, J. (1998). The essential Dewey: Pragmatism, education, democracy (Vol. 1). Indiana University Press. Eisner, E. W. (1997). The Promise and Perils of Alternative Forms of Data Representation. Educational Researcher, 26(6), 4-10. https://doi.org/10.3102/0013189X026006004 Foucault, M., Burchell, G., & Davidson, A. (2008). The birth of biopolitics: Lectures at the Collège de France, 1978–1979. New York, NY: Springer. Freire, P. (1996). Pedagogy of the oppressed (revised). New York: Continuum. Hartong, S., & Förschler, A. (2019). Opening the black box of data-based school monitoring: Data infrastructures, flows and practices in state education agencies. Big Data & Society, 6(1), 2053951719853311. https://doi.org/10.1177/2053951719853311 Harvey, D. (2006). Space as a keyword. In G. D. Castree N (Ed.), David Harvey: A Critical Reader (pp. 270-294). Blackwell. Landri, P. (2018). Digital governance of education: Technology, standards and Europeanization of education. Bloomsbury Publishing. Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge university press. Lefebvre, H., & Nicholson-Smith, D. (1991). The production of space (Vol. 142). Oxford Blackwell. Perrotta, C., Gulson, K. N., Williamson, B., & Witzenberger, K. (2020). Automation, APIs and the distributed labour of platform pedagogies in Google Classroom. Critical Studies in Education, 1-17. Williamson, B. (2017). Big data in education: The digital future of learning, policy and practice. Sage.