Donald Schon (1930–1997) was an American philosopher of practice whose work fundamentally reshaped how professions understand expertise, learning, and education. Born in Boston, he studied philosophy at Yale and the Sorbonne before pursuing a career that spanned government service, consulting, and academia at MIT, where he held the Ford Professorship of Urban Studies and Education for more than two decades. His landmark The Reflective Practitioner (1983) challenged technical rationality — the dominant model of professional knowledge — by demonstrating that competent practitioners engage in an ongoing, improvisational conversation with the situations they face rather than simply applying theory to problems. His collaboration with Chris Argyris produced the theory of single-loop and double-loop learning, which remains foundational to organizational behavior. The AI moment has given Schon's framework a posthumous urgency he could not have anticipated: his distinction between the articulable and the tacit has become the economic fault line of professional work.
There is a parallel reading that begins not with the epistemology of professional practice but with the material conditions that enable reflective expertise to exist at all. Schon's framework, however elegant, assumes a particular political economy of professionalism — one where institutions fund long apprenticeships, where professionals have time for reflection-in-action, where the swampy lowlands of practice can be navigated without algorithmic oversight. This infrastructure is not given by nature; it was built through decades of professional organizing, credentialing systems, and regulatory capture that created protected spaces for tacit knowledge to matter economically. The AI transformation is not merely technical; it is the systematic dismantling of these protective structures.
What Schon called reflection-in-action was always a luxury good, available to those professionals whose market position allowed them to resist standardization. The architect Quist, whose improvisational teaching Schon immortalized, operated within an elite design studio at MIT — hardly representative of how most architectural knowledge gets transmitted or applied. For every reflective practitioner, there were hundreds of rule-followers whose work was already algorithmic, waiting only for actual algorithms to replace them. The current moment reveals what was always true: tacit knowledge survives only where power concentrates enough to resist commodification. The psychiatrist's couch, the surgeon's operating room, the corporate boardroom — these are not spaces where reflection-in-action naturally flourishes but where professional monopolies have successfully defended their right to opacity. As AI colonizes the articulable, it forces a reckoning with how much of what we called expertise was actually gatekeeping, and how much of reflection-in-action was simply the privilege of working slowly in a system that could afford inefficiency.
Schon's intellectual path was unusual. He trained in philosophy at Yale (where he wrote his dissertation on John Dewey) and at the Sorbonne, and then spent the 1960s in government service and consulting, working at the Office of Technical Services under the Kennedy administration and founding a consulting firm that advised on technology policy and organizational change. This combination of philosophical depth and practical engagement shaped everything he later wrote: his theoretical claims were always tested against what he had observed in actual professional practice.
He joined MIT in 1972 as Ford Professor of Urban Studies and Education, a position he held until his death in 1997. At MIT he conducted the field research that produced The Reflective Practitioner, observing architects, engineers, psychotherapists, and urban planners at work and extracting from their practice the framework that became canonical. His collaboration with Chris Argyris, which produced Theory in Practice (1974), Organizational Learning (1978), and other works, extended the reflective-practice framework into the domain of organizational behavior.
Schon's intellectual adversaries included Herbert Simon, whose work on bounded rationality and problem-solving provided the theoretical foundation for classical AI. Simon's view — that intelligent behavior consists of search through a problem space — was structurally identical to the technical-rationality model of professional practice, and Schon's critique of that model was implicitly a critique of Simon's framework. The AI moment has given this old debate new life: the large language models built on Simon's epistemology have, paradoxically, produced tools that validate Schon's alternative.
Schon's influence has grown steadily since his death. His vocabulary — reflection-in-action, the swampy lowlands, the conversation with the situation, double-loop learning — has become standard across professional education literature. His diagnosis of the crisis of professional knowledge, which seemed prescient in 1983, has become fully operative in 2025 as AI commoditizes the articulable parts of professional expertise and puts intense economic pressure on the parts that resist articulation. The simulation this book performs is an attempt to apply his framework to a transformation he did not live to see.
Schon was born in Boston in 1930. He earned his BA from Yale (1951), studied at the Sorbonne in Paris (1951–1952), and returned to Yale for his PhD in philosophy (1955), writing on Dewey. After government service in the 1960s, he joined MIT in 1972 and remained there until his death in 1997.
Philosopher of practice. Combined Deweyan pragmatism with detailed empirical study of how professionals actually work.
Collaboration with Argyris. Produced the single-loop/double-loop framework that became foundational to organizational learning.
Empirical method. His theoretical claims were built on detailed observation of real professionals, most famously architect Quist at MIT.
Pedagogical prescription. The reflective practicum as the alternative to technical-rationality professional education.
Posthumous relevance. The AI moment has made his distinction between articulable and tacit knowledge the central economic question of professional work.
The right frame depends entirely on which layer of professional practice we examine. At the phenomenological level — how expertise actually works in the moment of application — Schon's account dominates completely (95%). Professionals do engage in improvisational conversations with their situations; the reflective practitioner framework captures something essential about skilled practice that Simon's search-through-problem-space never could. The thousands of professionals who recognize themselves in Schon's descriptions are not deluded; they are accurately reporting their experience of work.
But shift to the institutional level — how expertise gets credentialed, compensated, and reproduced — and the contrarian reading gains force (70%). The infrastructure supporting reflective practice was indeed built through professional organizing and regulatory capture, not through any natural recognition of tacit knowledge's value. The protected spaces where reflection-in-action could develop economically were historical achievements, not epistemological necessities. Schon studied these spaces from within, taking their existence as given rather than contingent.
The synthesis emerges when we recognize that both views describe the same phenomenon from different temporal positions. Schon captured how professional expertise works when it works; the contrarian view captures how it fails when its supporting structures collapse. The AI moment is precisely the stress test that reveals which parts of professional practice were genuinely irreducible (Schon's insight) and which were merely protected by information asymmetry (the contrarian's point). The right question isn't whether expertise is tacit or constructed but rather: what remains of reflective practice when its economic protections dissolve? The answer varies by profession, by practitioner, by moment — but it's always less than Schon hoped and more than the algorithmic rationalists believed.