Papers

I show that attempts to avoid the paradoxes of infinite value through non-standard choice theories come at a cost: they entail extreme and sometimes implausible behavior in the pursuit of evidence. 

I argue that the standard refinement hierarchy in game theory no longer holds in strategic scenarios where one or more players have lexicographic preferences.

We examine a formal account of norm change which leverages the fact that individuals can change both their (i) behavior and (ii) conceptual schemata. 

I argue that the well-known paradox concerning the value of social learning (Rogers’s paradox) can be resolved when agents are allowed to use mixed strategies and interactions are positively assorted. 

It is well established that positive assortment promotes cooperation in social dilemmas. I demonstrate, however, that positive assortment can also undermine the prospects for cooperation. 

Fair division remains a contested problem, but I show that one approach achieves a kind of stability that others fail to match. 

Mixed strategies are usually dismissed in the one-shot prisoner’s dilemma, but I show that randomizing can both destabilize defection and sustain partial cooperation under conditions once thought hopeless for cooperation. 

Social tags reshape when and why spite evolves, revealing new forms of conditional harm. 

I show that the same forces that enable cooperation also make the dominance of ethnocentric strategies highly likely. 

When choices determine who exists, experiments show that moral motivation weakens. 

I show that majority rule isn’t just procedurally fair—it’s often the most egalitarian in outcome, equalizing satisfaction better than its rivals. 

I use game theory to reinterpret Locke’s state of nature as a coordination problem, not a war of all against all. On this view, Nozick’s path to the minimal state stalls on a free-rider problem, revealing limits to his invisible-hand account of political order. 

Experiments reveal no agreement on what impartial reasoning requires, challenging the idea that a single model—Rawlsian, Scanlonian, or otherwise—captures the impartial point of view.