In his 1980 Contract as Promise, Charles Fried famously argued that in the event one side fails to perform on the main subject of a contract, we should vindicate the moral bindingness of promises by giving the nonbreacher expectation damages. Over the years, other liberal contract theorists have argued just as strenuously for specific performance on the same grounds.

I argue here that both sides in this dispute are wrong. The right position for a will theorist or any other stripe of liberal contract theorist to take with respect to acceptable alternatives to performance is, in the first instance, no position at all – or more precisely, that each side should have whatever alternatives the parties agreed she should have. If liberal contract theorists ought to have no first-order moral preferences among possible alternatives to performance, what accounts for the almost universal view to the contrary? The short answer, I suggest, is blame it on Holmes.