University of Oxford
Utilitarianism is a normative ethical theory which states that the morally right action is the one that maximises the balance of happiness over suffering. However, ‘action’ can be interpreted in two different ways: (i) a unique action in a given circumstance; (ii) a type of action, such as ‘lying.’ (i) produces act utilitarianism and (ii) produces rule utilitarianism.
Much confusion could be generated if we did not also distinguish between two types of act utilitarianism and two corresponding types of rule utilitarianism. Normative act utilitarianism (hereinafter denoted by AU-N) is the doctrine that the right action is the one that produces the most happiness in that particular situation, while normative rule utilitarianism (hereinafter denoted by RU-N) holds that the right action is the one which conforms to a rule which, if followed generally, would maximise happiness. Practical act utilitarianism (hereinafter denoted by AU-P), on the other hand, is a decision procedure whereby the agent is required to weigh up the predicted consequences of each and every moral act that he performs and to perform the action that seems to produce the most happiness overall in that given situation. Practical rule utilitarianism (hereinafter denoted by RU-P) is a decision procedure whereby the agent is required to judge each possible act by virtue of the consequences that that type of act tends to produce; thus, the act of murder is always the wrong action by the standards of RU-P because murder usually produces more unhappiness than happiness. RU-P states that we should follow those rules (such as ‘do not murder’) that contribute the most to happiness in the long run – the consequence of the rule being in place should be more happiness than if the rule wasn’t in place. These rules are learnt through experience and established and developed by society throughout history and they must be followed even on an occasion where good consequences would best be promoted by breaking the rule. Blackburn (2008) compares RU-P to a referee enforcing rules and not worrying about the particular consequences in that case, because he knows that generally the rules are good.
But which is preferable, act utilitarianism or rule utilitarianism? Unsurprisingly, I shall argue that neither single-level act utilitarianism (AU-N plus AU-P) nor single-level rule utilitarianism (RU-N plus RU-P) is preferable, but that elements of both need to be combined in order to produce the best outcome. Moreover, a split-level approach (AU-N plus RU-P, or RU-N plus AU-P is not enough – a more complex, multi-level approach is needed.
RU-N seems to have arisen in response to criticisms against AU-N, such as that AU-N fails to recognise the intrinsic value of enforcing justice, protecting the innocent and minorities and keeping promises; the moral force of these actions, it is argued, is not wholly reducible to the happiness/unhappiness balance. Is RU-N, then, a sensible modification or an ad hoc defence mechanism? It seems the latter, since there appears to be no rational grounding for the modification other than our basic initial intuitions in particular circumstances and this leads to superstitious ‘rule-worship’. As Smart says (as referenced in Rachels 1995), so what if RU-N better maps our intuitions? It is also widely-known that the human brain is incapable of comprehending large numbers, and so we naturally underestimate the total positive effects of breaking a rule, when this is the sum of many small positive effects. For example, we intuitively think that an innocent man should not be hung in order to make the public feel safer, no matter how many people are made happier as a result, but this is because our minds cannot properly grasp the scale of the happiness produced in total. RU-N should not really even be called ‘utilitarianism’, since the fundamental criterion of maximisation subordinated to our intuitions.
So AU-N seems the correct position. However, AU-P is not an ideal decision procedure, since the calculations are themselves an action, and lengthy calculation can often waste precious time and thus decrease the utility of an agent’s response to a situation. RU-P, then, appears superior.
But we can go one step better. Hare suggests a ‘two-level’ system in reference to decision procedures, whereby we follow RU-P unless the situation is complicated and it is not clear which rule to abide by because, for example, there is more than one rule that is considered to maximise utility. In these situations, we should resort to AU-P. In reference to Williams’ ‘Jim and the Indians’ thought-experiment, Hare claims that this two-level system explains both why Jim should feel repugnance at shooting the Indian and why he ought, nevertheless, to do it.
Sidgwick proposes a similar system with regard to decision procedures, which he calls ‘government-house utilitarianism.’ Under this system, a collection of intelligent and educated people follow AU-P and they rule over the rest of the people, who follow RU-P.
Both Hare’s and Sidgwick’s systems are multi-level systems with the following format: AU-N plus RU-P, but sometimes AU-P; in other words, act utilitarianism is the correct normative description of ethics and in practice we should follow utilitarian ‘rules of thumb’ unless it seems worth calculating the consequences of our actions. Both systems seem plausible and either way, it is clear that neither rule utilitarianism nor act utilitarianism is preferable in itself, but that a combination of both is necessary.