In this article, I argue that Michael Moore’s (1997), and other similar formulations of desert retributivism – viz., the theory that holds punishment to be justified because of the deserved suffering it imposes on guilty offenders – are epistemically problematic. The argument draws on John Dewey’s inchoate critique of retribution, and on Dewey’s more general contention that the justification of ethical judgments and principles proceeds ex post – viz., that it depends on the experiences elicited by acting on those judgments (...) and principles. Based on Dewey’s ex post take on justification, I more specifically argue that, given its commitment to moral naturalism and to coherentism, Moore’s version of desert retributivism is epistemically unwarranted. This is because we have evidentially grounded reasons to think that many of the retributive judgments that underlie the principle of retributive justice are not clearly supported ex post in the relevant, desert retributivist sense. (shrink)

Robotization is an increasingly pervasive feature of our lives. Robots with high degrees of autonomy may cause harm, yet in sufciently complex systems neither the robots nor the human developers may be candidates for moral blame. John Danaher has recently argued that this may lead to a retribution gap, where the human desire for retribution faces a lack of appropriate subjects for retributive blame. The potential social and moral implications of a retribution gap are considerable. I argue that the retributive (...) intuitions that feed into retribution gaps are best understood as deontological intuitions. I apply a debunking argument for deontological intuitions in order to show that retributive intuitions cannot be used to justify retributive punishment in cases of robot harm without clear candidates for blame. The fundamental moral question thus becomes what we ought to do with these retributive intuitions, given that they do not justify retribution. I draw a parallel from recent work on implicit biases to make a case for taking moral responsibility for retributive intuitions. In the same way that we can exert some form of control over our unwanted implicit biases, we can and should do so for unjustifed retributive intuitions in cases of robot harm. (shrink)