Joel Feinberg concludes that “the state has a right to prevent self-regarding harmful conduct only when it is substantially non voluntary or when temporary intervention is necessary to establish whether it is voluntary or not.” (quoted in Arneson, 1980: pg 470). In the example of the bridge referenced earlier, an ill informed member of the public stepping on a bridge which would collapse under their weight could not be said to be acting voluntarily; the state would be well within its rights in such a scenario under Mill’s logic to station a guard patrolling the area, to leap in and tackle such ill-informed people, stopping them from involuntarily taking the fatal step. If, however, they are running towards the bridge and shouting about how they know it will kill them, his logic would forbid the guard from taking any direct action. Whatever we feel about this compulsion to allow people to harm themselves if they so choose, it is a basic tenet of liberalism, that people know what the best for themselves is, and that interfering in their desire to pursue their own good in their own way is intrinsically wrong.This can, however, be harder to test than it would originally appear; what if the person is, for example, mentally ill? Indeed, if a person wishes to take an act which can obviously further their good but contains in it some degree of harm, or pain, we can easily understand both why a person, or state, would want to interfere and why they should be prevented from doing so. Yet if a person states that they wish to follow a course of action that will bring them no obvious benefit, and yet will clearly generate a great deal of harm, at what point should the state step in to take action? Is a mental assessment to determine their sanity unjust interference? After all, if a person is mentally unstable, surely they cannot truly be said to be acting of their own free will. Informed consent means more than merely knowing what the risks are, it means understanding what the consequences mean.