My larger aim in this chapter and the next is to demonstrate the relative inadequacy of deontological models of blame-responsibility compared to consequentialistic utilitarian models of task-responsibility. In opening that attack, let me first focus on the inadequacy of models of blame-responsibility where they might be expected to be on firmest ground.

Dictating how to distribute credit and blame among agents for what they have done or not done is, on the face of it, something that models of blame-responsibility should be able to do well. In truth, those models perform that task quite poorly, at least in an arguably common class of cases. Where the credit or blame is being distributed among people whose actions jointly either underdetermine or overdetermine the outcome, models of blame-responsibility are radically incomplete guides to the distribution of credit or blame. At most, they fix only broad parameters for those distributions.

My analysis on this point begins with some commonplace observations. Outcomes are characteristically the products of actions and omissions of many people. Furthermore, different people make different, and differentially important, contributions to the end product. Both these facts are well mapped in our standard notions of “responsibility.” Whatever particular account we want to give of that notion, it is clear that responsibility can be shared, and shared unequally, between various different agents.

What is less familiar, but nonetheless true, is that these fractional responsibilities do not necessarily sum to one.