EdTech routinely imagines the learner as an autonomous, self-regulating agent - capable of rational decision-making, goal-setting, and continuous self-improvement. Whether embedded in learning dashboards, adaptive algorithms, or “personalised” pathways, this vision reflects what Charles Taylor calls a social imaginary: the background moral and cultural framework that shapes what feels natural, inevitable, or legitimate within a given domain (Taylor, 2004). In this imaginary, the learner is not situated within a community or formed through relation, but is instead treated as a sovereign individual responsible for their own outcomes.

To understand the philosophical roots of this vision, we turn to Taylor’s (2007) concept of the buffered self. In contrast to the porous self - open to others, to the world, and to shared meaning - the buffered self is imagined as sealed off from external demands. It is a disengaged self, protected from vulnerability, tradition, and transcendence, and concerned primarily with inward-directed processes of reasoning, choice, and identity construction. Taylor associates this development with the rise of modern secularity, in which belief and meaning become privatised projects rather than shared horizons (Taylor, 2007).

This model finds a powerful analogue in how learners are constructed within educational technologies. Learning Management Systems (LMSs), AI tutors, and performance dashboards frame users as isolated agents tasked with interpreting feedback, adjusting behaviours, and optimising productivity. These platforms construct learning as a process of measurable self-enhancement: a data-driven exercise in personal growth. Learners are represented not as participants in communities of inquiry but as accountable subjects in a regime of metrics and nudges. Such figures presuppose a pedagogical model in which formation is individual, and success is the result of self-discipline rather than mutual dependence or care.

The result is a digital ecology that encodes a particular version of autonomy: not as collective self-governance or dialogic agency, but as responsibilised self-optimisation. As Williamson, Bayne and Shay (2020) argue, the datafication of teaching reconfigures educational practices through a logic of surveillance and measurement - one that positions both teachers and students as accountable subjects within platform-driven systems. Within this framework, the learner is expected to constantly monitor and improve their own performance, treating education as a project of continuous enhancement. Even the ethical and political dimensions of learning are reduced to individualised metrics, flattened into dashboard goals or optimised behavioural choices.

Yet this model is not pedagogically neutral. As Taylor (1989) insists, our visions of the person are always moral visions - choices about what kind of life is worth aspiring to. To imagine the learner as buffered is to marginalise forms of education that depend on intersubjectivity, shared inquiry, and care. It disqualifies porous forms of learning - those that value dialogue, vulnerability, and community - as inefficient or irrational.

This post argues that the dominance of the buffered self in digital education is not simply a design decision - it is a philosophical orientation with deep pedagogical consequences. By examining how this vision is built into the architecture of EdTech, we can begin to ask: what might it mean to reimagine the learner not as a self-contained unit, but as a relational, dependent, and situated subject? What kind of digital pedagogies would honour that vision?

Charles Taylor’s A Secular Age (2007) offers a rich account of how modern Western societies have come to imagine the self. Central to his analysis is the contrast between two models of subjectivity: the porous self and the buffered self. These are not just psychological categories but reflections of deeper shifts in what Taylor calls the social imaginary - the background framework of moral and cultural assumptions that shape how we experience the world and ourselves within it.

The porous self is characteristic of pre-modern societies. It is open to spiritual forces, moral claims, and communal identities that transcend individual will. In this mode, the boundaries between self and world are fluid. Meaning is not generated privately but received through embeddedness in cosmic, social, or divine orders. Emotions, motivations, and transformations are seen as arising through relationship, ritual, and encounter.

The buffered self, by contrast, emerges gradually through key historical shifts, especially the Protestant Reformation, the Enlightenment, and the rise of scientific rationalism. In this configuration, the self is enclosed, protected, and self-sufficient. Its boundaries are clearly marked: it no longer relies on external sources of meaning but constructs its own identity through internal processes of reflection and reasoning. The buffered self is disembedded and disengaged - less vulnerable to external influence, but also less connected to shared moral horizons (Taylor, 2007).

For Taylor, this transformation represents a fundamental change in how the self is constituted. Where the porous self is formed through relation, the buffered self is imagined as autonomous, rational, and in control. It is not merely a descriptive model but a normative ideal that has come to underpin liberal individualism. In this view, human agency is defined by independence, choice, and mastery over external conditions - values that are also foundational to modern conceptions of education.

This buffered model is strongly echoed in the architecture of contemporary educational technologies. Adaptive learning systems, gamified dashboards, behaviour-tracking apps, and AI tutors all invite learners to see themselves as isolated agents in charge of monitoring and improving their own performance. These platforms privilege a vision of autonomy aligned with self-regulation, efficiency, and individual goal-setting. Learning becomes a private project, measurable through data and optimisable through feedback loops.

The implications are far-reaching. As Taylor (1989; 2007) notes, the way we imagine the self determines the kinds of practices we design and the forms of life we legitimise. The buffered self is not simply a psychological assumption - it is a cultural orientation embedded in the moral fabric of our institutions. In digital education, this orientation encourages a turn away from relationality, interdependence, and shared formation. Technologies that presuppose the buffered learner marginalise the porous dimensions of education: dialogue, vulnerability, spiritual inquiry, and community.

Taylor does not suggest a simple return to the porous self. Rather, he invites us to see both models as partial and historically situated. In recognising the buffered assumptions behind EdTech design, we create space for pedagogical alternatives - ones that honour the constitutive role of relation, openness, and shared meaning in the formation of the learner.

The buffered self, as described by Taylor (2007), finds vivid expression in the logic and design of contemporary educational technologies. While such platforms often claim to promote learner autonomy, the form of autonomy they operationalise is grounded in self-containment, individual responsibility, and continuous optimisation. Technologies like Learning Management Systems (LMSs), AI-driven tutors, and personalised dashboards do not merely support learning; they configure a particular kind of learner - disembedded, self-governing, and accountable primarily to the system’s metrics.

LMSs such as Canvas, Moodle, and Blackboard exemplify this model. Although they support some collaborative tools, their core architecture centres on individual access, assessment, and achievement. Each learner logs in to their personal space, tracks their own progress, submits their own work, and receives individualised feedback. Even seemingly mundane features - submission time stamps, gradebooks, progress bars - subtly position the learner as a self-monitoring unit of performance. These interfaces invite the student to view learning as a private project of compliance and output management.

This logic is further intensified by AI-based learning systems. These tools generate adaptive pathways by aggregating data on individual learners, tailoring tasks and recommendations based on predicted need or likely success. The learner is cast as an optimisation problem - one whose performance can be fine-tuned through granular feedback, behavioural nudges, and real-time data visualisation. As Knox (2020) argues in his study of AI in China, such technologies exemplify a broader global trend in which education is framed as a system to be governed through automation and data. These tools presume not only that learning is measurable but that learners should continuously regulate themselves in response to algorithmic outputs - reproducing a model of the self as rational, responsibilised, and individually accountable (Williamson, 2017; Knox, 2020).

The metaphors that structure these platforms reinforce this imaginary. The dashboard positions the learner as a driver navigating a system through instrumental control. The analytics profile visualises progress via engagement metrics and behavioural data, turning reflection into self-surveillance. The productivity tracker imagines value in terms of speed, regularity, and task completion. These metaphors do not simply describe learner activity; they encode a design logic that assumes and encourages a buffered model of the self: autonomous, rational, and accountable.

This logic aligns with what Williamson, Bayne and Shay (2020) identify as the datafication of teaching and learning. Educational processes are reframed as datasets to be captured, visualised, and acted upon. Dialogue, ambiguity, vulnerability, and collective inquiry - dimensions that resist quantification - become invisible. What is visible - and therefore valuable - are actions that produce data. Clicks, completions, quiz scores, and engagement durations stand in for learning itself.

Crucially, this infrastructure enacts a process of responsibilisation. Learners are encouraged to see themselves as solely responsible for interpreting their data traces and acting accordingly. This responsibilisation masquerades as empowerment, but in practice, it individualises responsibility while masking the broader systems and relationships that support meaningful learning. As Eynon (2015) argues, the quantified self in education encourages students to view self-tracking as a route to personal improvement, yet rarely questions how this shapes their sense of identity, agency, or what counts as valuable learning. The result is a learner positioned as both subject and object of continuous monitoring - bounded, analytic, and optimising - rather than as a porous being formed through dialogue, care, and community (Selwyn, 2022; Eynon, 2015).

Recognising these technologies as instantiations of the buffered self allows us to interrogate their assumptions. Rather than treating platform defaults as neutral, we can begin to ask what kinds of learners they invite us to become - and what alternative pedagogical futures they may be foreclosing.

Educational technologies often presume that learners are rational, self-regulating individuals - capable of setting goals, tracking progress, and adjusting behaviour independently. This vision aligns closely with Taylor’s (2007) concept of the buffered self: inwardly governed, autonomous, and detached from formative social relations. It reflects a moral ideal of the learner as self-sufficient and responsible, an ideal encoded into the design of EdTech platforms that prioritise personalisation, optimisation, and continuous self-monitoring.

Personalised learning systems, AI tutors, and data dashboards are often promoted as tools for learner empowerment. But the control they offer is narrowly structured. Rather than supporting open-ended inquiry, they channel learners into pre-configured pathways based on predictive modelling. As Knox (2020) shows in his analysis of AI in China, such technologies promote an automated governance of education - where students are expected to respond to system-generated nudges, track their own metrics, and adapt behaviour in line with algorithmic expectations. Control is reframed as responsiveness to data; agency becomes synonymous with self-regulation.

These systems reward behaviours that conform to a managerial ideal of the learner: efficient, consistent, and measurable. Engagement is tracked through logins, clicks, and task completions. Success is defined through dashboards and analytics scores. Slowness, confusion, hesitation, or collaborative struggle are hard to visualise and therefore undervalued. As Selwyn (2022) notes, what is not captured by the platform often does not count. Affective, relational, and embodied dimensions of learning are marginalised - not because they are unimportant, but because they resist quantification.

This narrowing of what is visible in learning reflects broader neoliberal logics. As Brown (2015) argues, neoliberalism extends economic rationality into all domains of life, transforming individuals into entrepreneurs of the self. In education, this manifests as responsibilisation: learners are expected to treat themselves as projects in perpetual development - constantly upgrading, correcting, and performing. As Williamson (2017) observes, the learner becomes a data subject: someone to be measured, visualised, and governed through metrics. These systems rarely ask what the learner wants to become; they ask only how the learner can be made more efficient.

The consequences are profound. When autonomy is defined purely in terms of behavioural self-management, it loses its ethical and political dimensions. There is little space for autonomy as critical agency, as the capacity to question the structure of tasks, to reshape goals, or to resist platform logics. What becomes of learning as a shared, affective experience - when the system cannot register hesitation, care, or surprise? For educators, this raises urgent design questions: how can digital environments support autonomy understood not as control, but as co-agency and mutual formation?

To challenge these assumptions is to reclaim pedagogy as a relational and imaginative act. EdTech platforms do not simply support learning; they make particular forms of learning thinkable, and others difficult to realise. If we are to resist the moral narrowing of autonomy into optimisation, we must begin by asking what kinds of selves our systems invite - and what kinds they exclude.

The buffered self, as articulated by Taylor (2007), is not only a model of inwardness but also one of insulation. It describes a form of personhood in which individuals are imagined as self-contained, protected from external claims, and capable of generating meaning without reference to others. In the context of digital education, this model is operationalised in systems that valorise independence, self-regulation, and personal optimisation. But in privileging these attributes, educational technologies systematically marginalise a foundational aspect of learning: relationality.

Learning is not solely a cognitive or individual process. It is shaped through encounters - with other people, with ideas, with difference. As Biesta (2013) argues, education is a process of subject-formation: we become who we are not in isolation, but in relation to others. This process is dialogic, risky, and often uncomfortable. It involves moments of dissonance, vulnerability, and recognition. These dimensions are essential to deep learning, yet they are rarely captured - or even acknowledged - by platforms that measure educational engagement through clicks, completions, or time-on-task.

When learners are imagined as isolated agents - dashboard-driven, data-managed, and perpetually self-correcting - the social, affective, and ethical dimensions of education are diminished. Platforms that prioritise individualised pathways often do so at the expense of shared inquiry. A group discussion may be recorded merely as a participation checkbox, its richness invisible to the system. Collaborative learning is reduced to task coordination rather than mutual exploration. Reflection becomes a private log, detached from the dialogic process that gives it meaning. As hooks (1994) reminds us, education at its most transformative happens in community. It requires spaces where learners can speak, listen, and be recognised.

This is not merely a pedagogical concern. As Shotter (1993) points out, modernity has privileged modes of knowing that are abstract, objective, and decontextualised. In contrast, relational knowing is situated, embodied, and constructed through dialogue. Digital education, particularly in its datafied forms, risks reinforcing this epistemic narrowing. It renders visible only what can be measured and engineered - while devaluing the tacit, the co-present, the emergent. These conditions of isolation are not inevitable. They are designed, built into systems that mistake visibility for value.

This logic also undermines the ethics of care and mutual dependence. As Tronto (2013) argues, care is not an optional supplement to education - it is foundational to justice and responsibility. A platform that treats learners as self-contained data points leaves little room for attentiveness, responsiveness, or trust. It discourages educators from engaging in the subtle, relational labour that makes formation possible. It discourages learners from acknowledging their interdependence - not as weakness, but as the basis of ethical educational life.

To reassert relationality is to challenge the buffered model at its core. It is to affirm that education is not merely the transmission of content but the co-construction of meaning in relation to others. Digital platforms need not exclude this vision - but they must be redesigned. If they continue to treat learners as isolated nodes in a system of optimisation, they will foreclose the very conditions under which learning becomes meaningful.

If the buffered self has become the dominant imaginary of the learner in contemporary EdTech - self-contained, autonomous, and optimising - then recovering the porous self offers a way to reimagine pedagogy on fundamentally different terms. As Taylor (2007) describes, the porous self is shaped not through enclosure but through openness. It is formed in relation to others, to meaning beyond the self, and to traditions, histories, and social worlds that exceed individual control. Where the buffered self seeks autonomy through control and separation, the porous self is constituted through dependence, encounter, and transformation.

Reclaiming this vision in education means recognising that learning is not simply internal processing triggered by external inputs. It is emergent, social, and dialogic. It arises through trust, vulnerability, and the willingness to be affected by others. As hooks (1994) insists, meaningful education is not just about content - it is about community. Dialogue is not merely a pedagogical technique but an ethic: it affirms the learner’s subjectivity while also calling them beyond themselves. Similarly, Biesta (2006) argues that education involves interruption - moments when others disturb our settled understandings and invite us to become something more.

Pedagogically, a turn toward the porous self means designing learning environments that sustain interdependence and mutual formation. This includes practices such as dialogic teaching (Alexander, 2017), peer feedback, collaborative authorship, and shared reflection. These are not simply active learning techniques but invitations into co-presence and co-construction. Educators committed to this vision build structures that allow learners to speak and be heard, but also to listen, respond, and grow together. They create space for moments of friction - disagreement, discomfort, silence - recognising these as essential to learning.

Digital tools can support porous pedagogies. But this is only possible when their design is guided by a different imaginary - one that values connection over control. Platforms can be built to scaffold dialogue, to support collective authorship, and to highlight relational patterns rather than isolate metrics. As Greenhow and Lewin (2016) argue, social media affordances - when used critically - can enable forms of networked participation that challenge the individualism embedded in many learning platforms. But doing so requires resisting the dominant tendencies of educational technology: surveillance, standardisation, and individualisation.

This shift is not just technical - it is moral. A pedagogy of the porous self resists the drive to make learning frictionless and efficient. It embraces the slow, relational, and difficult dimensions of formation: the uncertainty of collaborative interpretation, the vulnerability of speaking into silence, the discomfort of having one’s assumptions undone. It sees interdependence not as a problem to be managed, but as the very ground of ethical education.

In this way, we are invited to imagine digital education not as a closed system to be optimised, but as a space of encounter - where selves are shaped, challenged, and sustained through the presence of others.

This post has explored how educational technologies translate a particular vision of the learner into platform design - what Charles Taylor (2007) calls the buffered self. In this vision, learners are imagined as rational, self-contained agents capable of self-management, optimisation, and meaning-making in isolation. LMSs, dashboards, and AI-driven tutors enact this model by treating learning as an individualised process, managed through performance tracking, feedback loops, and goal-setting interfaces. These systems reframe education as a matter of personal adjustment, increasingly stripped of its relational, social, and affective dimensions.

These assumptions are not technical necessities - they are the result of design decisions shaped by specific ideological commitments. Rooted in liberal individualism and extended through neoliberal responsibilisation, the buffered learner model encodes a moral and political imaginary that aligns closely with logics of measurement, efficiency, and behavioural nudging (Williamson, 2017; Brown, 2015). Platform features such as personalised dashboards, self-paced modules, and progress analytics give concrete form to this imaginary, positioning learners as entrepreneurs of the self - constantly regulating, tracking, and refining their own outputs.

Against this backdrop, we have argued for the importance of reimagining learning through the lens of the porous self. Drawing again from Taylor, this conception understands the self not as sealed and sovereign but as formed through relation - with others, with tradition, and within shared ethical worlds. It foregrounds the dialogic, situated, and affective character of learning as subject-formation. As Biesta (2006) and hooks (1994) remind us, education at its most transformative involves encounter, interruption, and co-presence. It cannot be reduced to data flows or individual metrics.

To move toward a pedagogy of the porous self is not to abandon digital tools but to interrogate the assumptions they carry. It means asking how we might design learning environments that sustain interdependence, support ethical formation, and make space for dialogue, slowness, and uncertainty. It is a call to reclaim education as a humanising practice - one grounded in care, vulnerability, and the difficult work of mutual recognition. Platforms do not merely support learning; they shape what learning is allowed to become.

Yet platform imaginaries shape more than just our view of the learner - they also embed particular visions of what is fair, valuable, and worth achieving. From performance dashboards to ranking systems and behavioural analytics, EdTech platforms do not merely measure learning - they construct moral narratives about merit, effort, and success. In the next post, we will examine the moral order behind these systems, drawing on Taylor’s account of market society. We will ask how platforms encode assumptions about fairness, reinforce meritocratic ideals, and naturalise inequality - while exploring how educators might reclaim plural, relational sources of value in educational practice.

• Alexander, R.J. (2017) Towards dialogic teaching : rethinking classroom talk, University of Cambridge. Fifth edition. Thirsk: Dialogos.
• Biesta, G. (2006) Beyond Learning: Democratic Education for a Human Future. Boulder, CO: Paradigm Publishers.
• Biesta, G. (2013) The Beautiful Risk of Education. Boulder, CO: Paradigm Publishers.
• Brown, W. (2015) Undoing the Demos: Neoliberalism’s Stealth Revolution. New York: Zone Books.
• Eynon, R. (2015) ‘The quantified self for learning: critical questions for education’, Learning, Media and Technology, 40(4), pp. 407–411. https://doi.org/10.1080/17439884.2015.1100797
• Greenhow, C. and Lewin, C. (2016) ‘Social media and education: reconceptualizing the boundaries of formal and informal learning’, Learning, Media and Technology, 41(1), pp. 6–30. Available at https://doi.org/10.1080/17439884.2015.1064954
• hooks, b. (1994) Teaching to Transgress: Education as the Practice of Freedom. New York: Routledge.
• Knox, J. (2020) ‘Artificial intelligence and education in China’, Learning, Media and Technology, 45(3), pp. 298–311. Available at https://doi.org/10.1080/17439884.2020.1754236
• Selwyn, N. (2022) Education and Technology: Key Issues and Debates. London: Bloomsbury Academic. Available at http://dx.doi.org/10.5040/9781350145573
• Shotter, J. (1993) Cultural Politics of Everyday Life: Social Constructionism, Rhetoric and Knowing of the Third Kind. Buckingham: Open University Press.
• Taylor, C. (1989) Sources of the Self: The Making of the Modern Identity. Cambridge, MA: Harvard University Press.
• Taylor, C. (2004) Modern Social Imaginaries. Durham, NC: Duke University Press.
• Taylor, C. (2007) A Secular Age. Cambridge, MA: Belknap Press of Harvard University Press.
• Tronto, J. C. (2013) Caring Democracy: Markets, Equality, and Justice. New York: NYU Press.
• Williamson, B. (2017) Big Data in Education: The Digital Future of Learning, Policy and Practice. London: SAGE Publications.
• Williamson, B., Bayne, S. and Shay, S. (2020). ’The datafication of teaching in Higher Education: critical issues and perspectives’. Teaching in higher education, 25(4), pp.351–365. https://doi.org/10.1080/13562517.2020.1748811