How, then, should assisted dying work? For many the model is Oregon’s Death with Dignity Act. It allows (but does not oblige) doctors to prescribe lethal drugs to patients with less than six months to live who ask for them, if a second doctor agrees. There is a cooling-off period of 15 days.

We would go further. Oregon insists that the lethal dose is self-administered, to avoid voluntary euthanasia. To the patient the moral distinction between taking a pill and asking for an injection is slight. But the practical consequence of this stricture is to prevent those who are incapacitated from being granted help to die. Not surprisingly, some of the fiercest campaigners for doctor-assisted dying suffer from ailments such as motor neurone disease, which causes progressive paralysis. They want to know that when they are incapacitated, they will be granted help to die, if that is their wish. Allowing doctors to administer the drugs would ensure this.

Oregon’s law covers only conditions that are terminal. Again, that is too rigid. The criterion for assisting dying should be a patient’s assessment of his suffering, not the nature of his illness. Some activists for the rights of the disabled regard the idea that death could be better than a chronic condition as tantamount to declaring disabled people to be of lesser worth. We regard it as an expression of their autonomy. So do many disabled people. Stephen Hawking has described keeping someone alive against his wishes as the “ultimate indignity”.

One exception to this distinction should be children.