Autonomy and the Educated Subject: Rethinking Learning Beyond Instrumentality

A faceless figure stands before a cosmic backdrop with a glowing sun, geometric networks, and floating shapes, evoking a surreal vision of thought, space, and digital interconnectedness.

Introduction: What Kind of Learner?

What kind of learner do our technologies assume? This deceptively simple question opens onto a complex terrain of pedagogical values, institutional priorities, and political imaginaries. In much of technology-enhanced learning (TEL), the learner is increasingly framed through data: a subject to be measured, nudged, and optimised. Embedded within this framing is a behaviourist logic that reduces learning to observable outputs and education to a sequence of system-managed interventions.

Learning analytics platforms, intelligent tutoring systems, and gamified dashboards all enact a particular vision of the learner: one whose motivations can be inferred from clicks, whose progress is best tracked numerically, and whose needs can be met through automated feedback loops. This figure – the “data-driven learner” – is assumed to thrive in environments where efficiency, prediction, and compliance are paramount. But this vision is not neutral. It emerges from and reinforces a conception of education as instrumental: a means to measurable ends, governed by external metrics of success rather than intrinsic or shared purposes (Biesta, 2010).

By contrast, Cornelius Castoriadis offers a radically different account of the educated subject. For Castoriadis, autonomy is not merely about individual freedom within a given system – it is the capacity to question and create the very rules by which one lives. The autonomous subject is one who engages in self-institution, not simply self-regulation (Castoriadis, 1997). Such a vision foregrounds the political and creative dimensions of education: to be educated is to become capable of participating in the collective shaping of social meaning, institutions, and imaginaries.

These competing visions – the data-driven learner and the autonomous subject – reflect deeper tensions in the philosophy of education. On one side lies an instrumental rationality that treats education as a tool for achieving predefined outcomes; on the other, a commitment to education as a space for the emergence of new forms of subjectivity and sociality. As Feenberg (2017) argues, technical rationality tends to obscure human agency by embedding value-laden decisions into supposedly neutral systems. When applied to education, this tendency risks replacing the question of what education is for with the managerial concern of how best to deliver it.

To rethink the learner, then, is to rethink education itself – not as a pipeline to productivity, but as a site of democratic and imaginative possibility.

The Autonomous Subject – Castoriadis’ Vision

Autonomy, for Castoriadis, is not reducible to the liberal ideal of individual independence. It is not simply the freedom to choose within a predefined system, nor the capacity to self-manage according to external goals. Rather, autonomy is the collective and reflexive capacity of individuals and societies to question, reinterpret, and create the very rules by which they live. It is the possibility of self-institution – of recognising that all social forms are historically contingent and thus open to transformation (Castoriadis, 1997).

At the heart of this vision lies what Castoriadis calls the radical imaginary – the generative, creative force by which human beings institute meaning. It is through the radical imaginary that societies bring forth their own norms, values, and institutions, and through which they may also come to interrogate and transform them. Autonomy, in this sense, is not a given but an achievement: it involves the conscious creation of meaning through collective deliberation and reflection. As Adams (2005) explains, Castoriadis’s account of autonomy emerges from this capacity for creation ex nihilo – the imaginative act of bringing new social forms into being. Autonomy is thus inseparable from the political project of questioning inherited institutions and actively participating in the creation of new ones (Adams, 2005; Castoriadis, 1997).

This understanding of autonomy stands in stark contrast to its appropriation in contemporary educational discourse, particularly under neoliberalism. In many learning technologies, autonomy is framed behaviourally – as “self-regulated learning,” where the learner takes responsibility for goal-setting, time management, and performance monitoring (Zimmerman, 2002). While seemingly empowering, this framing often collapses autonomy into compliance: the learner is autonomous to the extent that they optimise their behaviours within a predetermined system. Autonomy is rendered as individual discipline, not collective creation.

Moreover, this notion aligns closely with the broader neoliberal subject: entrepreneurial, responsibilised, and governable through metrics. Learners are expected to track their own progress, respond to system prompts, and cultivate “grit” or “resilience” in the face of systemic constraints (Gillies, 2011). In this paradigm, the radical potential of autonomy is flattened. Education ceases to be a space for the questioning of social forms and becomes, instead, a site for their reproduction.

A Castoriadian approach reclaims autonomy as both a personal and political capacity. It invites educators to foster conditions where learners do not merely navigate a system, but participate in imagining and re-instituting the very frameworks through which learning is understood and enacted.

The Behaviourist Legacy of Learning Technologies

Contemporary learning technologies - ranging from learning analytics systems to gamified apps - often embody a behaviourist logic that reduces learning to quantifiable inputs and outputs. Stemming from early behaviourist psychology, this logic treats learners as passive recipients whose responses to stimuli (e.g. clicks, quiz results, badges) can be tracked, analysed, and shaped (Wilson et al., 2017).

Learning analytics frequently frame learner activity as discrete events to log and optimise. As Wilson, Thompson, Drew, and Doyle (2017) note, such analytics “relies on [digital] traces as proxies for learning,” underscoring the behaviourist tendency to equate data with knowledge. The emphasis on what is measurable privileges easily captured performance over deeper, contextualised understanding - effectively narrowing the notion of learning to what can be counted.

Similarly, gamification strategies - such as points, badges, and leaderboards - leverage extrinsic reinforcement to shape user actions. While some meta-analyses show small effects on engagement and perceived autonomy, these effects are uneven and can undermine intrinsic motivation if the design lacks depth (Liu et al., 2024; Sailer et al., 2017). Worse still, poorly designed gamification can reduce learning to compliance with system prompts - an interaction driven by the overjustification effect, whereby external rewards diminish a learner’s internal motivation (Deci, Koestner & Ryan, 1999).

This entrenched behaviourism reinforces a narrow framing of autonomy: learners are ’autonomous’ only when they self-regulate within predetermined parameter settings - completing tasks, meeting targets, and responding to algorithmic nudges. Autonomy becomes behavioural optimisation rather than political self-determination. As Kaliisa et al. (2023) demonstrate, learning analytics dashboards often fail to improve achievement or foster meaningful motivation, instead perpetuating a loop of system-informed compliance.

By reducing complex, creative processes to clicks and metrics, such technologies enact a kind of closed-loop compliance. Learning becomes about producing the right data signals, and autonomy is reframed as successful alignment with system exigencies - not the capacity to shape, question, or withdraw from them.

Heteronomy by Design – When Platforms Define Possibility

The term heteronomy refers to being governed by rules or forces external to oneself, in contrast to autonomy. In Kantian ethics, heteronomous actions are those driven by impulses or external authorities rather than rational self-legislation (Kant, 1996). Castoriadis extends this critique to social and digital institutions: learning platforms often function heteronomously by shaping behaviour and perceptions without acknowledging their own constructed nature (Castoriadis, 1997).

Heteronomous structures in learning platforms

Most Learning Management Systems (LMS) and digital course environments are heteronomous by design. These systems rely on predefined content hierarchies, sequential modules, and algorithmic triggers. Learners are positioned as system components: to click the next lesson, submit assignments by a fixed deadline, or respond to automated nudges. Such interfaces leave little room for learner-led exploration or reinterpretation of the learning path. As Rosário and Dias (2022) note, despite their pedagogical potential, LMSs are frequently configured to support monitoring, compliance, and delivery at scale rather than co-construction or critical engagement.

In this context, platforms often reflect what Mengay (2020) terms the “digitalisation of heteronomy” - where work and learning environments increasingly rely on algorithmic structures that standardise behaviour while concealing their own contingency. Learning becomes entangled with unseen systems that operate beyond the learner’s control or awareness.

Fixed pathways and default nudges

Design conventions in many LMS platforms include linear progression (“modules must be completed in sequence”) and “nudges” - notifications or reminders timed to drive completion rates. While these afford accountability, they also constrain learning to institutional timelines and expectations. Fawns (2022) argues that such pedagogical technologies often operate within a broader entangled infrastructure, where design choices reflect managerial rather than educational priorities. Nudges and constraints, though technically flexible, tend to reproduce default institutional assumptions.

Kerssens and van Dijck (2022) highlight how edtech governance can undermine pedagogical autonomy, particularly when platform features are determined more by commercial and administrative logic than by educational values. Even when intended to support learners, these default settings can reduce the space for critical judgement, reflection, and learner-led experimentation.

Automated feedback and learner disempowerment

Automated feedback systems - such as quizzes that reveal right/wrong answers or dashboards that visualise progress - serve as gatekeepers of knowledge. They standardise evaluation and reinforce system-defined metrics. As a result, learners orient towards satisfying algorithmic expectations instead of pursuing curiosity or developing critical judgement. Reflection becomes informational, not transformative.

In these ways, digital learning becomes a heteronomous encounter: learners comply with external structures that appear natural, while their role in co-constructing the learning environment is hidden. Castoriadis warns that when institutions appear objective and immutable, they mask their own self-created origins - keeping individuals trapped in heteronomy. A Castoriadian pedagogy would expose these hidden structures and invite learners to critique and redesign them, moving from compliance to self-institution.

Reclaiming Autonomy – Towards Self‑Institution in Pedagogy

To reclaim autonomy, pedagogy must move beyond compliance-based learning and instead position learners as active authors of their educational journeys. This requires practices that cultivate the capacity for self-institution - the creative and political act of defining and shaping one’s own modes of learning and engagement.

Open‑ended tasks

Open-ended tasks resist behavioural scripting by inviting learners to frame questions, explore multiple pathways, and produce original artifacts. Papastephanou (2021) argues that reclaiming learning requires acknowledging learner diversity, fostering epistemic justice, and supporting transformative engagement - moving beyond standardised metrics to embrace creative, co-constructed educational experiences.

Collective co‑design of learning environments

Inviting learners to co-design their learning environments foregrounds political agency and turns education into a collective practice. Brown et al. (2021) document a graduate-level design-based research project in which students, instructors, librarians, and coordinators collaboratively developed an open educational resource via Pressbooks. They outline how a structured co‑design framework - clarifying process, making thinking visible, building relationships, and sustaining beyond the course - transformed learners from passive receivers into empowered knowledge creators. This participatory approach exemplifies self‑institution: students configure meaningful digital artifacts and co-legislate what counts as learning.

Critical media literacy

Critical media literacy repositions learners as interpreters and producers of media, unpacking power, representation, and ideology. In their foundational essay, Alvermann and Hagood (2000) describe critical media literacy as a pedagogy that enables students to enjoy and critique popular culture simultaneously, revealing the often-invisible mechanisms of social control embedded in media texts. This dual focus - pleasure and critique - positions learners not merely as recipients but as agents in re-authoring media narratives.

Building on this, Funk, Kellner and Share (2016) argue that critical media literacy in the digital age must go beyond decoding media content. It should enable learners to create and circulate counter-narratives, challenge dominant ideologies, and participate meaningfully in democratic discourse. This kind of literacy is central to a pedagogy of autonomy, enabling learners to critically intervene in the cultural and technological systems they inhabit.

Participatory and democratic assessment

Assessment practices themselves should invite learners to reflect and participate. Democratic models allow students to propose assessment formats (e.g. portfolios, peer review, open rubrics), engaging them as co-assessors. This shifts assessment from surveillance to shared responsibility, as advocated by emancipatory pedagogy theorists (hooks, 1994).

These pedagogical practices collectively reposition learning as a creative-political process. They value learners as subjects who can imagine, critique, and reconfigure learning and living environments. In doing so, they enact Castoriadis’s vision: education not as training in compliance, but as praxis in self-institution.

Beyond Instrumentality – Education as a Site of Political Becoming

Education, to be truly emancipatory, must transcend the narrow aim of crafting “functional individuals” and instead nurture autonomous subjects - those capable not only of navigating systems, but of imagining, questioning, and reshaping them. Only then does learning become a political practice of becoming rather than a technical means to predefined ends.

1. Nurturing autonomous subjects, not compliant performers

Freire (1970) emphatically critiques the “banking” model of education, where students are passive vessels and teachers deposit knowledge into them. Instead, he proposes a problem-posing pedagogy, in which learners and teachers engage in dialogic exchange - co-investigating reality to foster critical consciousness (conscientização) and collective liberation. Education becomes a practice of freedom, not conformity.

2. Learning as collective meaning-making and institutional creation

Following Freire, hooks (1994) portrays education as inherently political: the classroom becomes a space of transgression, where dominant structures of race, gender, and class can be critically dismantled. Through engaged pedagogy, both teachers and students cross boundaries - intellectually, socially, and institutionally - co-creating meaning and reconfiguring institutional norms.

Education, then, gestures toward what Castoriadis would call self-institution: the collective development of new norms, values, and imaginaries through conscious praxis rather than passive reception.

3. Reconnection with critical pedagogy traditions

Paulo Freire remains the foundational voice in this tradition: liberation education requires learners to become subjects of their own learning, capable of transforming themselves and the world. bell hooks extends this vision into feminist and multicultural contexts, asserting that teaching must be a political act of love, hope, and resistance, where everyone participates as fully human.

4. Implications for digital and TEL environments

If education is to be a site of political becoming, then digital learning environments must be reimagined not merely as platforms for content delivery, but as spaces for democratic engagement, institutional imagination, and co-creation. This requires a decisive break from instrumental design principles rooted in automation, efficiency, and behavioural control.

Dialogic design

Too often, TEL platforms favour unidirectional flows: lectures streamed, quizzes marked, feedback automated. In contrast, dialogic design foregrounds mutual inquiry, sustained conversation, and relational accountability. Inspired by Freirean pedagogy, discussion boards, annotation tools, or synchronous meetings should be repurposed not to reinforce correct answers but to facilitate collective sense-making. Tools like social annotation platforms (e.g. Hypothes.is) can be used not simply for commenting on text, but for enabling civic discourse and public pedagogy - as Kalir and Garcia (2019) demonstrate through their analysis of open-web annotation in civic writing spaces, revealing how learners develop political voice and collaborative meaning-making through digital annotation practices.
Learner-generated curricula and participatory governance

Inspired by Freire’s use of generative themes, digital learners can co-define the focal points of a course, using surveys, collective tagging, or forum discussions to surface what matters to them. These learner-identified themes can inform weekly modules, guest speaker invitations, or assessment priorities. Beyond content, students might participate in course governance, helping shape the timing of submissions, rubrics for evaluation, or the weighting of collaborative versus individual work. Veletsianos and Houlden (2020) argue that such practices exemplify radical flexibility - a pedagogical stance that centres the needs, contexts, and agency of learners during times of uncertainty. Rather than enforcing standardised structures, radical flexibility invites responsiveness, co-authorship, and care into the design of digital learning environments.
Critical platform literacy

Following hooks’ insistence on intersectional critique, learners must be equipped to question how platforms themselves shape what can be known and said. This includes understanding the affordances and limitations of algorithms, the economic logics of edtech providers, and the ways digital architectures embed assumptions about who the learner is. Courses in digital pedagogy and media studies should include critical analysis of institutional platforms like Canvas, Moodle, or Blackboard - not only how to use them, but what they presuppose.
Ethical infrastructuring

Finally, institutions should consider open, learner‑owned platforms (e.g. blogs, Fediverse tools, self‑hosted portfolios) that allow students to retain their data, publish on their own terms, and build networks outside the institutional LMS. Fawns (2019) argues that postdigital educational design must break binary distinctions between digital and non‑digital, and instead embrace entangled, ethical infrastructures that foreground learner agency and relational accountability. These infrastructures decentralise authority and enable learners to see themselves as creators of educational spaces - not just users of predefined systems.

In redefining education not just as skill acquisition but as political becoming, we reclaim its capacity to shape democratic subjects - capable of self-governing and institution-building. Digital education must echo this transformation: from instrumental delivery systems to spaces of collective freedom, critique, and emergence.

Conclusion: A Call to Educators

If education is to reclaim its democratic, creative, and humanising potential, educators must learn to see beyond the interfaces they are given. The default settings of educational technologies - structured pathways, automated nudges, and data-driven feedback - may appear neutral or efficient. Yet they encode assumptions about what learning is, who learners are, and what they are for. These defaults are not merely technical choices; they are pedagogical commitments made in advance, without the learner’s voice.

To resist these defaults is not to reject technology, but to insist on the primacy of pedagogy. It is to ask: What kinds of learners do our platforms imagine? What forms of thinking, collaboration, and becoming do they afford or foreclose?

Educators must therefore work not only within systems but on them - reshaping the structures of digital learning so that they serve dialogic, imaginative, and participatory ends. This means:

Designing spaces for learner agency and co-creation, not just content delivery.
Making institutional platforms open to questioning, remixing, and critique.
Supporting collective authorship of meaning, not just individualised performance.

In the spirit of Castoriadis, Freire, and hooks, education must be reframed as an institutional imaginary in flux - a place where learners do not merely inherit norms, but actively participate in creating and transforming them (Castoriadis, 1997; Freire, 1970; hooks, 1994). Such a vision demands courage: to refuse the instrumentalisation of teaching, to critique the architectures of platform governance, and to reassert the value of learning as a form of political and ethical becoming.

In a time of automation, standardisation, and surveillance, the most radical act may be to teach as if the world could be otherwise.

Bibliography

Adams, S. (2005). ’Interpreting Creation: Castoriadis and the Birth of Autonomy#. Thesis Eleven, 83(1), 25-41. https://doi.org/10.1177/0725513605057135
Alvermann, D.E. and Hagood, M.C. (2000) ’Critical media literacy: Research, theory, and practice in “New Times”’, The Journal of Educational Research, 93(3), pp. 193–205. https://doi.org/10.1080/00220670009598707
Biesta, G.J. (2015) Good education in an age of measurement: Ethics, politics, democracy. London: Routledge.
Brown, B., Hurrell, C., Roberts, V., Jacobsen, M., Neutzling, N. & Travers‑Hayward, M. (2021) ’Open Education Co‑Design as a Participatory Pedagogy in an Online Graduate Program’, Oteṣṣa Conference Proceedings, 1(1), pp. 1–8. https://doi.org/10.18357/otessac.2021.1.1.41
Castoriadis, C. (1997) The imaginary institution of society. Translated by K. Blamey. Cambridge: Polity Press.
Deci, E.L., Koestner, R. and Ryan, R.M. (1999) ’A meta‑analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation’, Psychological Bulletin, 125(6), pp. 627–668. https://doi.org/10.1037/0033%E2%80%912909.125.6.627.
Fawns, T. (2019) ’Postdigital education in design and practice’, Postdigital Science and Education, 1(1), pp. 132–145. https://doi.org/10.1007/s42438-018-0021-8
Fawns, T. (2022) ’An entangled pedagogy: Looking beyond the pedagogy–technology dichotomy’, Postdigital Science and Education, 4, pp. 711–728. https://doi.org/10.1007/s42438-022-00302-7
Feenberg, A. (2017) Technosystem: The social life of reason. Cambridge, MA: Harvard University Press.
Freire, P. (1970) Pedagogy of the Oppressed. New York: Continuum.
Funk, S., Kellner, D. and Share, J. (2016) ’Critical media literacy as transformative pedagogy’, in Yildiz, M.N. and Keengwe, J. (eds) Handbook of Research on Media Literacy in the Digital Age. IGI Global, pp. 1–30. https://doi.org/10.4018/978-1-4666-9667-9.ch001
Gillies, D. (2011). ’Agile bodies: a new imperative in neoliberal governance’. Journal of education policy, 26(2), pp.207–223. https://doi.org/10.1080/02680939.2010.508177
hooks, b. (1994) Teaching to Transgress: Education as the Practice of Freedom. New York: Routledge.
Kaliisa, R., Misiejuk, K., López‑Pernas, S. and Khalil, M. (2023) ’Have Learning Analytics Dashboards Lived Up to the Hype? A Systematic Review of Impact on Students’ Achievement, Motivation, Participation and Attitude. arXiv:2312.15042v1 [cs.HC]
Kalir, J. H., & Garcia, A. (2019). ’Civic Writing on Digital Walls’. Journal of Literacy Research, 51(4), 420-443. https://doi.org/10.1177/1086296X19877208
Kant, I. (1996) Groundwork of the metaphysics of morals. Translated and edited by M.J. Gregor. Cambridge: Cambridge University Press.
Kerssens, N. and van Dijck, J. (2022) ’Governed by Edtech? Valuing pedagogical autonomy in a platform society’, Harvard Educational Review, 92(2), pp. 284–303. https://doi.org/10.17763/1943-5045-92.2.284
Li, L., Hew, K.F. & Du, J. (2024) ’Gamification enhances student intrinsic motivation, perceptions of autonomy and relatedness, but minimal impact on competency: a meta-analysis and systematic review’. Educational Technology Research and Development 72, 765–796. https://doi.org/10.1007/s11423-023-10337-7
Mengay, A. (2020). ’Digitalization of work and heteronomy’. Capital & Class, 44(2), 273-285. https://doi.org/10.1177/0309816820904032
Papastephanou, M. (2021). ’Reclaiming learning’. Policy Futures in Education, 19(1), pp.13-27. https://doi.org/10.1177/1478210320940141
Rosário, A.T. and Dias, J.C. (2022) ’Learning Management Systems in Education: Research and Challenges’, in Geada, N. and Jamil, G. (eds) Digital Active Methodologies for Educative Learning Management. Hershey, PA: IGI Global, pp. 47–77. https://doi.org/10.4018/978-1-6684-4706-2.ch003
Sailer, M., Hense, J.U., Mayr, S.K. and Mandl, H. (2017) ’How gamification motivates: An experimental study of the effects of specific game design elements on psychological need satisfaction’, Computers in Human Behavior, 69, pp. 371–380. https://doi.org/10.1016/j.chb.2016.12.033
Veletsianos, G. and Houlden, S. (2020) ’Radical flexibility and relationality as responses to education in times of crisis’, Postdigital Science and Education, 2(3), pp. 849–862. https://doi.org/10.1007/s42438-020-00196-3
Wilson, A., Thompson, T.L., Drew, V. and Doyle, S. (2017) ’Learning analytics: challenges and limitations’ Teaching in Higher Education, 22(8), pp.991–1007. https://doi.org/10.1080/13562517.2017.1332026
Zimmerman, B. J. (2002) ’Becoming a Self-Regulated Learner: An Overview’, Theory Into Practice, 41(2), pp. 64–70. https://doi.org/10.1207/s15430421tip4102_2.