March taught Don Quixote at Stanford for years, not as a literary curiosity but as a model for organizational leadership. The figure is disciplinary scandal: a rigorous formal modeler insisting that the most important things about leadership cannot be captured by models, that the knight who charges windmills illustrates a structural feature of decision-making under ambiguity that rational-choice theory cannot see. When the situation is genuinely uncertain, when the correct interpretation is unknowable in advance, the quality of the decision cannot be evaluated by its outcome. A good decision that produces a bad outcome was still a good decision. The quality resides in the process, not the result — in the range of alternatives considered, the tolerance for ambiguity maintained throughout, the willingness to act with total commitment to a vision the actor does not fully understand.
March used Quixote to illustrate what rational-choice theory cannot accommodate: action whose justification is not available in advance and whose outcome cannot serve as verdict on the decision. The knight does not know whether the windmills are windmills or giants. The knight who charges has committed before knowing; the knight who does not charge has already surrendered. The charge is an act of integrity precisely because it cannot be justified by rational calculation. It is the only form of integrity available to creatures that must act under genuine ambiguity — which is to say, to creatures facing the most consequential decisions of organizational life.
The leader navigating the AI transition is Quixote. The windmills are the organizational challenges AI presents: the ratchet, the competency trap, the myopia, the organizational forgetting, the garbage can dynamics that produce unmanaged scope expansion. Some of these are windmills — manageable challenges that will yield to competent leadership. Some may be giants — structural threats that will destroy organizations not prepared for them. The leader does not know which is which. The environment is too new, the precedents too few, the data too sparse. What the leader can do is maintain the organizational capacity for both responses: the capacity to exploit the windmill (capture AI productivity gains) and the capacity to explore the giant (investigate whether the organizational model itself needs rethinking).
March's Quixote is not the figure of comic delusion familiar from popular culture. He is a figure of principled commitment to values that rational calculation cannot validate. The knight's madness is not belief that windmills are giants — it is the refusal to accept that the difference between windmills and giants is the only question worth asking. The deeper questions — what kind of life is worth living, what kind of organization is worth building, what kind of vision is worth charging for — are questions rational choice cannot answer because they are prior to rational choice. They are the questions that constitute the self that does the choosing.
The framework refuses resolution. March would not have ended his framework on a note of optimism about finding the right balance. He would have ended — and in his writing consistently did end — with the observation that the balance cannot be resolved, only managed; that the management is perpetual and imperfect; that the organizational sin that is unforgivable is the pretense that the tension has been resolved, the right strategy has been found, the organization can stop adjusting. The pretense is the trap. The tension is the work. And the work, like all genuinely important work, never ends.
March's engagement with Quixote spanned decades of Stanford teaching and writing. His lectures on leadership, literature, and decision-making appeared in various forms, culminating in the documentary Passion and Discipline: Don Quixote's Lessons for Leadership (2003) and the edited volume On Leadership (with Thierry Weil, 2005). The choice of Quixote was deliberately scandalous to the management-science orthodoxy March had helped build. His point was precisely that the formal modeling he had mastered did not exhaust what could be said about organizations and leadership — that the humanities preserved insights about judgment, commitment, and action under ambiguity that the social sciences had largely abandoned.
The Quixote framework also tied together the themes of March's late career: ambiguity, foolishness, the limits of rational calculation, the moral dimensions of organizational life. Each of these themes found in Quixote a figure who embodied them more fully than any formal model could. The knight was not an example; he was an argument.
Integrity through commitment. Under genuine ambiguity, the quality of a decision resides in the process — range of alternatives, tolerance for not-knowing — not in the outcome.
Windmills or giants. AI presents challenges that may be manageable or catastrophic; the leader cannot know in advance which, and must maintain capacity for both responses.
Beyond rational choice. Questions about what is worth doing are prior to rational calculation; they constitute the self that chooses.
Refusal of resolution. The exploration-exploitation balance cannot be solved, only managed; the tension is permanent, the adjustment never ends.
Foolishness as discipline. The knight's charge is disciplined foolishness — action committed to a vision that cannot be justified but cannot be abandoned without loss of integrity.
Whether Quixote is the right figure for organizational leadership in the AI age is contested. Critics argue that the knight's commitments — chivalric ideals that his society had already abandoned — make him a figure of nostalgia rather than forward-looking adaptation. Defenders respond that the content of Quixote's commitments is not the point; the point is the structural posture of acting under ambiguity with integrity, and that posture is more necessary in the AI age than in any previous period. March's own position was characteristically subtle: he used Quixote not as a model to imitate but as a lens through which features of organizational leadership became visible that had been obscured by the dominance of rational-choice frameworks.