Ground Projects
Ground projects are the commitments constitutive of an agent's identity — the things without which her life would cease to be recognizably hers. Williams introduced the concept to distinguish them from ordinary preferences, which can be swapped without existential cost. The scientist whose life is organized around inquiry, the engineer whose identity is inseparable from building, the craftsman whose self is constituted by mastery — these are persons whose ground projects are not external to their identities but internal to them. Asking such a person to abandon her project is not asking her to change her preferences but asking her to become someone else. The AI transition threatens ground projects by transforming the practices around which they were formed, and threatens their replacement by denying the temporal space in which new ones could constitute themselves.
In the AI Story
Williams introduced ground projects in his 1976 essay 'Persons, Character and Morality' to defend the claim that impartial moral theories misdescribe what agents owe themselves. A morality that requires its agents to treat all reasons as impartially weighable cannot make sense of the fact that agents have reasons to live at all — reasons grounded in the specific commitments that give their lives forward momentum. Ground projects are the reason-giving structures without which moral deliberation itself loses its footing.
The concept illuminates a dimension of the AI transition the productivity discourse cannot see. When a senior software architect confronts the commoditization of his expertise, the loss is not primarily economic. The economic dimension is real but manageable. The deeper loss is the destabilization of the ground project: the commitment to craft that organized his decades of effort and constituted the self that accumulated them. 'Reskill and adapt' misdescribes this situation because it treats the craft as a skill swappable for another rather than as the constitutive commitment it actually was.
Ground projects form slowly. They are not chosen from a menu but emerge through sustained engagement with practices that prove meaningful. This temporal fact is consequential. When the AI transition demands that practitioners demonstrate value in the new landscape before they have had time to discover what the new landscape makes possible, it demands the formation of commitments under conditions hostile to their formation. Constitutive commitments require exploratory uncertainty and the slow recognition of what matters — both incompatible with quarterly metrics.
Williams distinguished ground projects from conscious life-plans. A life-plan can be revised; a ground project announces itself through the discovery that revision would be self-destruction. The engineer who moves from backend systems to frontend work in response to AI augmentation may be exercising reasonable career judgment. The engineer who is told that her entire identity as a builder of careful systems is obsolete and should be replaced by something called 'creative direction' is being told that her ground project should be abandoned, which is a different kind of demand and produces a different kind of response.
Origin
The concept first appeared in Williams's 1976 essay 'Persons, Character and Morality' (collected in Moral Luck, 1981). It was the central pivot of his critique of utilitarianism's demand for impartial maximization, and it became one of the most frequently cited concepts in late-twentieth-century moral philosophy, especially in debates about the relationship between morality and the good life.
Key Ideas
Constitutive, not elective. Ground projects are not things the agent happens to want but things without which the agent would not be the agent she is.
Reason-generating. They provide the motivational structure that makes deliberation her deliberation rather than an impersonal computation conducted in her skull.
Slow to form. Ground projects emerge through sustained engagement with practices that prove meaningful; they cannot be chosen from a menu or assembled on a deadline.
Transformable, not swappable. An existing ground project can be transformed by new circumstances and new commitments, but the transformation is an existential event, not a preference adjustment.
Threatened in two ways by AI. The transition destabilizes existing ground projects by transforming their practices, and threatens their replacement by denying the temporal space new ones require.
Debates & Critiques
Some critics, including Susan Wolf and Peter Railton, have argued that Williams's defense of ground projects licenses a form of moral parochialism — a willingness to prioritize one's own attachments in ways that impartial morality rightly condemns. Williams's response was that this is not a bug but a feature: a morality requiring its agents to treat their own lives as items among others has already made a philosophical error about what agents are.
Appears in the Orange Pill Cycle
Further reading
- Bernard Williams, 'Persons, Character and Morality' in Moral Luck (Cambridge, 1981)
- Bernard Williams, 'A Critique of Utilitarianism' in Smart & Williams, Utilitarianism: For and Against (Cambridge, 1973)
- Susan Wolf, Meaning in Life and Why It Matters (Princeton, 2010)
- Samuel Scheffler, The Rejection of Consequentialism (Oxford, 1982)
- Cheshire Calhoun, Doing Valuable Time (Oxford, 2018)