Universal prescriptivism (often simply called prescriptivism) is the meta-ethical view which claims that:
Ethical sentences do not express propositions.
Instead, ethical sentences function similarly to imperatives which are universalizable — whoever makes a moral judgment is committed to the same judgment in any situation where the same relevant facts obtain.
This makes prescriptivism a universalist form of non-cognitivism or expressivism. Prescriptivism stands in opposition to other forms of non-cognitivism (such as emotivism and quasi-realism), as well as to all forms of cognitivism (including both moral realism and ethical subjectivism).
Since the concept was introduced by philosopher R. M. Hare in his 1952 book The Language of Morals, it has been compared to emotivism and to the categorical imperative of Immanuel Kant.
For an illustrative example of the prescriptivist stance, consider the moral sentence "Murder is wrong". According to moral realism, such a sentence claims there to be some objective property of 'wrongness' associated with the act of murder. According to moral relativism, such a sentence simply claims that murder is disapproved of by society. According to emotivism, such a sentence merely expresses an attitude of the speaker; it only means something like "Boo on murder!" But according to prescriptivism, the statement "Murder is wrong" means something more like "Do not murder" — what it expresses is not primarily a description or an emotion, it is an imperative. A value-judgment might also have descriptive and emotive meanings, but these are not its primary meaning on a prescriptivist account.
Hare would allow utilitarian considerations to enter into such a formulation, but he would not base the formula or his ethical theory solely on a principle of utility. Hare believed that all of our ethical propositions ought to conform with logic.
Peter Singer has expressed sympathy with Hare's position, though he is more strictly representative of the preference utilitarian school.