Daniel C Dennett and Gregg D Caruso debate free will in Aeon:
[Dennett] You weren’t responsible for becoming an autonomous agent, but since you are one, it is entirely appropriate for the rest of us to hold you responsible for your deeds under all but the most dire circumstances. As [the American country singer] Ricky Skaggs once put it: ‘I can’t control the wind, but I can adjust the sails.’ To suppose that some further condition should be met in order for you or anyone else to be ‘truly deserving’ is to ignore or deny the manifest difference in abilities for self-control that we can observe and measure readily. In other words, the rationale or justification for excusing someone, holding them not deserving of criticism or punishment, is their deficit in this competence. We don’t try to reason with bears or babies or lunatics because they aren’t able to respond appropriately. Why do we reason with people? Why do we try to convince them of conclusions about free will or science or causation or anything else? Because we think – for good reason – that in general people are reasonable, are moved by reasons, can adjust their behaviour and goals in the light of reasons presented to them. There is something indirectly self-refuting in arguing that people are not moved by reasons! And that is the key to the kind of self-control which we are justified in treating as our threshold for true desert.
Caruso: I don’t disagree with you that there are important differences between agents who have the kind of rational control you highlight and those who lack it. Such a distinction is undeniable. A normal adult who is responsive to reasons differs in significant ways from one who is suffering from psychopathy, Alzheimer’s or severe mental illness. I have no issue, then, with acknowledging various degrees of ‘control’ or ‘autonomy’ – in fact, I think you and other compatibilists have done a great job highlighting these differences. My disagreement has more to do with the conditions required for what I call ‘basic desert’ moral responsibility.
More here.