Home   Hughes 1   Hughes 2   Hughes 3   Hughes 4

Shirley Hughes

Rules and Exceptions

One of the factors which have led many philosophers to adopt a more or less sceptical attitude in moral philosophy has been the recognition that most rules have exceptions. This has commonly been regarded as a threat to the entire moral enterprise. How can a philosopher even attempt to find an account of the moral relations that obtain among things which will weave them into the unity of a stable system if every principle, every rule, every judgment has to be qualified by who knows how many exceptions?

Plato was acutely aware of how devastating the admission of an exception might be. In the Republic Socrates completely invalidates Cephalus' thesis that justice is simply a matter of returning to others what is due to them by pointing out that if a friend deposited a weapon with us for safekeeping and then asked for it when he was not in his right mind, there would be justice in not returning it to him. Ordinarily we should return what does not belong to us, but this case would seem to be a legitimate exception. Socrates mentions another. It would be right in such circumstances he says to lie to a person who was out of his mind. On the other hand Plato also realised that by no means all alleged exceptions are justified. In the Euthyphro Socrates upon being informed that Euthyphro intends to prosecute his own father for murder suggests that perhaps it would be right to prosecute his father if he killed a relative but not if he murdered a stranger. Euthyphro rebukes Socrates for suggesting such an exception. Socrates offers no defence except to express amazement at the certainty with which Euthyphro claims to know what is right.

There are several ways to resolve the problem of uncertainty which the existence of exceptions seems to introduce. One way is that of the extreme sceptic; simply deny the validity of morality altogether. There can hardly be a problem of exceptions if there is no such thing as morality. To most philosophers this solution to the problem seems too drastic and unnecessarily defeatist.

Another method of disposing of the problem is to dispense with rules and principles in determining what is moral. We have seen this approach taken by Sartre and to a considerable extent by Fletcher. Certainly if the problem is reconciling in a consistent fashion the existence of exceptions with the existence of rules then one solution is to deny the latter (just as one classic solution of the problem of evil is to deny the existence of God).

The third approach is the converse of the one just mentioned. Instead of denying that there are rules one simply denies that genuine moral rules and principles ever have exceptions. Perhaps the only explicit philosophical exposition of this view that moral rules are exceptionless is Kant's discussion in his essay "On a Supposed Right to Tell Lies from Benevolent Motives". Although this essay may not provide us with a complete or altogether accurate statement of Kant's position on exceptions it is none the less the interpretation which is made of Kant's views by most ethical commentators.

Still a fourth approach to the problem is to assert that while all or at least most rules have exceptions, principles do not. We need only to appeal to them or to some ultimate principle among them in order to determine which exceptions to rules are admissable and which are not.

Miller attributes this view to classical utilitarians such as Bentham and Mill, but also and more generally to "any theory in which it is maintained that some me rule always takes precedence over all others". Miller finds reasons for rejecting all such attempts. A fifth theory is suggested by Sidgwick and developed hypothetically by Miller. This theory attempts to resolve the problem of exceptions by appending to any given rule, a list of all exempting conditions; that is to say all the cases to which the rule is related but to which it does not apply.

Miller finds that this approach too is objectionable. A final position regarding exceptions is what's known as the Aristotelian theory of exceptions. Briefly it is the view that rules do not, and because of their general nature, cannot apply to all the cases of a certain class, thereby leaving some cases (the exceptions) to be decided by an appeal to intuition rather than by an appeal to rules and or principles. According to the usual interpretation of Kant's views he is said to believe that "moral rules are absolutely inflexible and without exceptions". Certainly much in Kant's essay "On a Supposed Right to Tell Lies from Benevolent Motives" would justify such an interpretation. He speaks of the "unconditioned principle of veracity" as a "command of reason not to be limited by any expediency" and of rules "which in their nature do not admit exceptions'.

He offers at least two reasons for answering and asserting the existence of exceptionless rules, namely that rules or principles which have exceptions are self-contradictory and that "exceptions destroy the universality of which they are called principles".

Nonetheless in his Fundamental Principles and elsewhere, Kant makes an important distinction between what he variously calls 'perfect' or 'strict' duties on the one hand and 'imperfect' or 'broad' duties on the other. A 'perfect' duty is he says one that admits of no exceptions in favour of inclination. He is less explicit regarding 'imperfect' or 'broad' duties. The distinction between these two types of duties permits several possible interpretations of Kant's view.

Regarding 'imperfect' or 'broad' duties he does say the following about them in later work The Metaphysical Principles Of Virtue p. 48

For the law can command only the maxim of actions and not the actions themselves then this is a sign that the law leaves in its ... observance of latitude ... for free choice i.e. it cannot definitely assign in what way and to what extent something should be brought about by an action directed to an end which is at the same time a duty. But by a broad duty is not understood a permission to make exceptions to the maxim of actions but only the permission to limit one maxim of duty by another (e.g. the general love of one's neighbour by the love of one's parents).

One view is simply that whereas 'strict' or 'perfect' duties allow no exceptions whatever 'broad' or 'imperfect' duties do.

However this appears to be an oversimplification of Kant's view and excluded by his statements that broad duties do not grant "permission to make exceptions to the maxim of actions.

Kant's position on exceptions would seem to be, no moral rules whether of 'perfect' or 'imperfect' duty can incorporate in their statement any mention of an exception.

Myself I take issue with Kant's claim that certain moral views are totally without exceptions by appealing to the almost universal belief that such rules have exceptions or possibly to my own insight that this is so. Assuming that moral rules can and do have exceptions the problem becomes one of understanding the nature of the logical relations between exceptions and rules.

One way to conceive this relation is to think of exceptions as cases which fall outside and restrict the scope of a rule. We may speak of exceptions conceived in this way as exceptions 'toll a rule.

Thus if killing in self defence is an exception to the rule "killing is wrong" we think of the rule as applying only to cases of non defensive killing.

Another way of conceiving the relation is to think of exceptions as built in negative specifications of the rule.

In this sense an exception is part 'of' the rule not a case that limits it so to speak. from the outside.

Thus if killing in self defence is an exception we no longer think of the rule as being "killing is wrong" but as "killing in self defence" is wrong. Singer appears to adopt the first way of construing exceptions. Baier appears to adopt the second. According to Singer there is a considerable difference between moral principles and moral rules. Moral principles he says hold in all circumstances, allow no exceptions and are always relevant, whereas rules hold only generally, do allow exceptions and are not always relevant.

Furthermore principles are apt to be more abstract than moral rules, though he says, they are not necessarily less definite. It is by appealing to principles that moral rules as well as exceptions to them are justified and established.

Ultimately such a justification procedure must involve an application of what Singer calls the generalisation argument which he says may be thought of indifferently as either an argument or as a moral principle. It may be expressed as follows; "if everyone were to do that the consequences would be disastrous therefore no-one ought to do that". Actually he finds that this principle includes two subordinate principles 1) the generalisation principle and 2) the principle of consequences (if the consequences of A's doing X would be undesirable then A ought not to do X).

In many respects Singer's generalisation argument and Kant's categorical imperative are similar. Both involve generalisation and both provide a criterion for the application of the principle. Both are supreme principles. The differences include the fact that Kant's principle involves a reference to 'willing' and to the maxim of an action. whereas Singer's does not. On the other hand Singer's incorporates an appeal to undesirable consequences and this is completely absent from the categorical imperative.

Singer's basic task is to show that 1) that not every application of the generalisation argument is a valid one and so determine the conditions under which it is valid and 2) determine the conditions to show that the generalisation argument, while capable of establishing and justifying certain moral views, does not establish them as holding always or in all possible circumstances and so explain the conditions under which exceptions to rules may be justified. One of the conditions which Singer mentions in connection with his first task, that of establishing moral rules, has to do with what he calls the 'invertibility' of the generalisation argument. He means that in order for a given application of the generalisation argument to be valid it must not be 'invertible' in the sense that the consequences of everyone's acting in a certain way would be undesirable and the consequences of no-ones acting in that way would also be undesirable. Thus the argument that no-one should grow food because if everyone did the consequences would be undesirable (no-one would provide the other basic necessities) is an invalid argument because it also follows that if no-one grew feed the consequences would also be undesirable (we would all starve).

Another condition of the validity of an application of the generalisation argument is that it is not 'reiterable' A generalisation argument fails to meet this condition if it is made to apply to some arbitrarily specified aspect of an action. Thus the argument that no-one has a right to eat at 6 pm because if everyone did then there would be no-one to perform certain essential functions, is reiterable and invalid, because the same reasoning could just as well be applied to other arbitrarily selected times e.g. eating at 5 pm or 7 pm. It is this last condition of validity, namely non-reiterability, which is particularly involved in Singer's second task, that of accounting for exceptions. The problem of exceptions in terms of Singer's theory may be stated as fellows; "By what procedure can one justify acting in a way in which it would be undesirable for everyone to act?" Singer's answer is that exceptions to moral rules are justified if one can show either I) that the individual involved is a member of a class of persons such that if every member of that class were to act in that way the consequences would not be undesirable or 2) that the circumstances of one's action are such that the consequences of everyone's acting in that way in those circumstances would net be undesirable.

For example, it can be shown that it would not be undesirable if everyone having below a certain minimum income did not pay taxes. The only restriction Singer places on this application of the generalisation argument is that it not be 'reiterable' with the respect to the class of persons or circumstances selected. That is to say the class cannot be so defined that everyone may regard himself as a member of the class or as being in those circumstances. On the other hand if a. person reasons that he should not pay his share of the tax burden because society will not miss his small contribution he will find if he tries to apply the generalisation argument to this ease that everyone else can equally argue the same way i.e. claim to belong to the same class. The argument is therefore reiterable, making everyone an exception. Singer regards this as self-contradictory.

There are several points to notice with respect to Singer's theory. First, if moral rules do not always hold i.e. if the generalisation argument with respect to them is not always valid then moral rules are always of restricted universality. Second, the notion of a rule without exceptions is self-contradictory because the reasons which would be sufficient to establish the rules can be the same ones which in certain circumstances would override it. For instance a rule against killing might be established by showing that if everyone acted in accordance with it the consequences would not be undesirable but killing in certain circumstances (such as in self defence) could also be established by means of the same reason. Third, Singer denies that an exception to a rule can ever be singular in nature and asserts that in justifying a class of exceptions one is in effect modifying the understanding of the original rule by restricting its scope.

From this it would seem to follow that the exception conceived of as a limitation on the scope of the original rule is not a part of the expected rule but external to it. No doubt many philosophers including Kant would challenge the claim that the notion of an exceptionless rule is self-contradictory, even those who are perhaps not convinced that any of the. usual moral rules are in fact exceptionless.

If this thesis is challenged then certainly one could also challenge Singer's claim that moral rules are never truly universal. The Aristotelian theory of exceptions stands opposed to Singer's denial that exceptions can ever be singular. The problem of conceiving of an exception as a wholly external limitation on a rule consists in understanding how it can be intelligibly regarded as an exception to rather than a violation of the rule. For if the case A not governed by the rule then it is either totally unrelated to the rule or it is a violation of the rule. But normally we would not say that an exception to the rule against lying is a violation of it. Singer attempts to avoid the problem by saying that no-one has a right to violate a moral rule 'without a reason' thereby implying. that exceptions are indeed violation of' rules but net necessarily wrongful violations.

This may not be a satisfactory way of putting it. Baier's view makes an interesting contrast with Singer's. It is Baier's contention that in making exceptions to laws and legal regulations, magistrates must decide what are the deserving cases i.e. the justified exceptions, by appealing to principles and specifically to moral principles, but this is not so with respect to the justification of exceptions in the case of moral views. There is simply no higher set or kind of rules or principles to which to make such an appeal. It is his view therefore that every genuinely moral view already contains all of the deserving cases as parts of the rule. Thus for Baier a rule against homicide, if it is a genuinely a moral view, then the rule includes as part of itself the legitimate case of killing in self defence. As he says "when we say therefore that a person who has killed a burglar in self defence has not done anything wrong", we are not making an exception in the houseowners' favour.

It is much nearer the truth to say that in our morality the rule "thou shalt not kill" has several recognised exceptions among them "in self defence". We can say that a man does not know fully our moral view "thou shalt not kill" if he does not know it has among others, this exception. It is clear that Baier cannot be charged with treating an exception as a "violation" of a rule in any sense of that term. If the exception is literally a pert of the rule and is governed by it and if the rule says "no killing except In self defence" the exception to killing (self defence) is perfectly consistent with and is in no way opposed to the rule.

The problem here though is whether in conceiving of an exception as a part of the rule, it makes any more sense to call the excepted cases e.g. killing in self defence, exceptions to a rule against killing than to call the non excepted cases e.g refraining from killing someone, exceptions to a rule permitting self defence. The point is that though we can distinguish types of cases i.e. non killing in certain circumstances from killing, in certain other circumstances, are we really making a logical distinction or merely a verbal one in calling me set of cases exceptions and not the others?

Furthermore since one set of cases is as related to and unopposed to the rule as the other why call either of them exceptions? Perhaps we have here another but more covert way of expressing the view that genuine moral rules are exceptionless.

The two views of Singer and Baier are similar in many respects to a contemporary development of rule-utilitarianism. It is the view that the rightness of an action is determined by applying the principle of utility to the rule under which it falls, not directly to the act.

The final theory to be considered is in a sense a modified version of rule scepticism. It does not seek to do away with the rules altogether but to restrict them to the cases to which they properly may be applied. It insists that rules are for the most part absolutely essential guides to conduct.

It questions only the value of rules to determine which cases and situations are exceptions.

According to Aristotle in every class of human actions are some cases to which rules cannot apply. Rules must of necessity be formulated as universal statements, but about some things, particularly human actions, it is not always possible to speak both universally and correctly. That is to say rules inevitably fail to take account of all the relevant peculiarities of some cases and situations but in saying that rules are deficient by virtue of their universality.

On the contrary many are perfectly good rules. It's just that rules are rules. Aristotle does not mean to suggest that all rules are bad or useless. Therefore unless we are to become rule worshippers and so fail to do the right thing in individual cases just to preserve the integrity and universal application of rules we must appeal in certain cases to another method in order to decide those cases. Fortunately there is such a method: the appeal to intuition.


Immanuel Kant Fundamental Principles and The Metaphysical Principles Of Virtue

Kurt Baier Ethics and Society (1966)

Kurt Baier The Moral Point Of View (1965)