Abstract
This thesis begins by arguing that hard indeterminism, a type of free-will skepticism (FWS), is the most defensible position within the free-will debate. Its competitors, libertarianism (both types: agent-causal and event-causal) and compatibilism fail for the following reasons. Agent-causal libertarianism succumbs to the problem of wild coincidences, and event-causal libertarianism falls foul of the disappearing agent objection. I distinguish between “leeway” and “source” compatibilism whereby the former asserts that a necessary condition for an agent to be morally responsible is to have alternative possibilities (“the agent could have done otherwise”), whilst the latter rejects this. “Leeway” compatibilism is refuted by the consequence argument, and “source” compatibilism is susceptible to the manipulation argument, on the proviso that a free-will skeptic can refute the “no generalisation” objection to the manipulation argument (MA), which it currently fails to do. I present a novel response to the “no generalisation” objection by arguing that to overcome it, the FWS needs to stretch the four-case MA to five cases, including a fifth case termed “the heritability case.” I then investigate the leading criminal punishment model on behalf of the FWS, the quarantine model of criminal punishment model, and argue that its implementation is likely to offer less than adequate deterrence. Due to this, the quarantine model is unlikely to reach democratic legitimacy and would thereby be detested by “the people”. Lastly, from these insights, I develop an original criminal punishment model on behalf of the free-will skeptic that I term “an integrative punishment model.” The integrative punishment model attempts to achieve democratic legitimacy by accommodating the “the people's” desire for vengeance, whilst “capping” the amount of punishment received by criminals, and lessening criminal sentences at the same time since I believe that no-one is ever morally responsible in a desert-based sense.