RUTH Barcan MARCUS AND MODALITY (2): Morals

Marcus turns from talking about possibility and necessity in general to talking about permissions and duties, that is, to deontic logic. Deontic logic is a kind of formal logic that deals with words like “ought”, “must”, “may”, “forbidden”, and “permitted”. Philosophers use a symbol, often “O”, for “it ought to be that” or “it is obligatory that”, and another symbol, often “P”, for “it is permitted that”.

The special twist  is that Marcus is interested in “iterated” deontic modalities. “Iterated” here just means “one on top of another”. So instead of just “O A”, which you can read as “you ought to do A”, she looks at things like “O O A”, which you can read roughly as “you ought to see to it that you do what you ought to do”, or “it is forbidden that something forbidden happens”, and so on.

To get into this, she first wants you to feel the difference between simple rules and rules about rules. A simple rule is like “You must eat your greens before eating cake”. A rule about rules is like “You must always obey your bosses’ rules”. The second one does not tell you directly what to do in each case. It tells you what to do with other rules. That is the idea of an iterated obligation. It does not just say “Do X”. It says “Do what you ought to do” or “Make sure the rules are followed”. In formal deontic logic people often make systems with neat properties. One way to make things neat is to keep the language “extensional”, like Quine advises. There you try to have very simple substitution rules, so that if two sentences are equivalent, you can swap them wherever you like without changing the truth. Some logicians tried to do something similar for duties. They wanted to treat “ought” like a kind of box that behaves in very tidy ways. But as with modality in general, this can make the logic too simple and it begins to misdescribe real life. Marcus says something like this is happening with many early systems of deontic logic.

She calls those systems “extensional deontic logics”. In such a system you take O as just another operator that should obey very strong rules. For example, if A and B are logically equivalent, then O A and O B should always be equivalent. Also, people often make a rule that says: if it is obligatory that A, then A is not forbidden. This sounds fine at first, but when you start nesting the “O” operator, you run into trouble. She points out a striking problem. In some extensional systems, as soon as you allow sentences like “O O A”, the rules force very odd results. At one point she notes that, given those rules, something as harmless as wiggling your eyebrows would end up being forbidden. She uses this absurd example on purpose. Eyebrow wiggling is a funny, rather neutral action. There is nothing obviously wrong with it. If a fancy logical system says that such a silly act must be forbidden, then something has gone wrong with the system.

To see the basic idea, think of a world in which there are only two possible actions you could do at some moment, call them A and B. Maybe A is “help your neighbour” and B is “ignore your neighbour”. Suppose the moral rule is that you ought to do A, and it is wrong not to do A. In ordinary moral thinking, that does not tell you anything about, say, secretly wiggling your eyebrows, which might be a third, entirely separate thing. It is not automatically forbidden. But in some extensional deontic logics, once you pack all the rules together, everything that is not required can look as if it is forbidden. It is as if the logic insists that the world must be neatly divided into what is required and what is forbidden, with no extra space for silly neutral acts. Marcus thinks that does not match how we actually think about ethics. Everyday moral life has lots of “free space” for harmless behaviour.

She pushes this further with the idea of a Manichean universe. Manicheans imagined the world as a battle between absolute Good and absolute Evil. Marcus asks you to imagine a world where every single action is either morally good or morally bad, with no middle ground, and where the rules about what is good are fixed in a very rigid way.  In such a world, an extensional, very strict deontic logic might be more at home. If every action is either commanded by good or commanded by evil, with no neutral actions, then strong formal rules that divide everything into “ought” and “must not” might match the picture. But our actual world is not like that. There are heroic actions that go far beyond duty, such as risking your life to save a stranger. There are also tiny, indifferent actions, like tapping your fingers, wiggling your eyebrows, or humming a tune. They are not morally required, and they are not morally forbidden. Marcus is interested in building a logic that can talk about duties and permissions in a way that leaves room for these facts. If she can then she would be building a model that is more fruitful than the Quinean alternative which seems to have problems doing that.

Now she returns to the idea of iterated deontic modalities. Why have rules about rules at all? She points out that we often talk this way quite naturally. A parent might say “You must always follow the posted safety rules at the swimming pool”. That is an obligation about other obligations, since the pool rules themselves tell you what you must do. A teacher might say “You are required to obey the school rules”. Some moral theories say things like “You ought always to do what morality requires”, which is roughly O(O A). These are ordinary sounding sentences, not strange inventions. So a logical system that refuses to let us write any sentences like “O O A” would already be too weak. It would be rejecting common patterns in moral talk.

On the other hand, if you allow “O O A” but treat it with very strict extensional rules, you risk the eyebrow problem and similar odd results. So Marcus thinks we need a more subtle account of how “O” works when you put it inside itself. She looks at two ways of understanding deontic operators that philosophers discuss: a “prescriptive” reading and an “evaluative” reading. On the prescriptive reading, “O A” is like a command. It says “Do A”. Then “O O A” would be like “Do what I just told you to do”. That is a rule that tells you to follow another rule. It is natural when you think about authorities like laws, parents, teachers, or religious commands. For example, a law might say “Citizens must obey all traffic laws”. That is a rule about rules.

On the evaluative reading, “O A” is more like “A is the right thing to do” or “A is obligatory from the point of view of morality”. Then “O O A” is something like “It is right that it be right that A”. That sounds repetitive, but it can matter when we think about clashes between different sets of rules, for example between legal rules and moral rules, or between the rules of a game and the rules of fairness. You can have “Legally, it is required that you follow the company policy, but morally, it is required that you do not follow it in this case”. Her point is that, depending on which reading you choose, the logic of nested “O” can behave differently. A good system of deontic logic should be flexible enough to represent that.

She also wants to keep the link with what she says about intensional languages. She argues that contexts involving “necessarily” and “possibly” are intensional, so you cannot simply swap equivalent sentences everywhere. Here, she argues that deontic contexts are also intensional. “It ought to be that A” is not just a neat box around A. The meaning of “ought” cares about how we think of the action, about the background rules, and about what counts as a violation. So she suggests that trying to force deontic logic into an extensional mould is a mistake of the same kind Quine made about modal logic. It ignores the real complexity of the concepts.Think about two ways of describing the same act. Suppose A is “telling the truth”, and B is “telling Zurko that I broke the vase”. In this situation, A and B pick out the same action, because the true thing you must tell Zurko is that you broke the vase. But the way the rule is written can matter. Suppose the rule at home is “You must always tell the truth”. Then “You ought to tell the truth” looks like a basic duty. “You ought to tell Zurko that you broke the vase” is a specific case of it. In simple logic, you might want these to be fully interchangeable. In real life, you might think “Yes, I must tell the truth in general, but maybe this one case is different”. So an “ought” that works on the general description may not feel the same as an “ought” that works on the specific description, even though they cover the same act. This is another sign that deontic contexts behave like intensional contexts.

Her main message is fairly simple. First, iterated deontic modalities, that is, rules about rules, are natural and needed to describe our moral thinking. Second, some neat extensional deontic logics produce absurd results when you allow such iterated “oughts”, such as making eyebrow wiggling forbidden or turning every non required action into a forbidden one, which does not match our sense of neutral or optional behaviour. Third, a better approach is to treat deontic operators as intensional and to be more careful about how we handle nested obligations, permissions, and prohibitions, so that there is room for neutral acts and for richer talk about rules and meta rules. 

So, after repairing our logical language so that it can handle talk of possibility and necessity without collapsing everything into a flat extensional picture, Marcus does something similar for moral concepts. She says that “ought” and “may”, especially when you repeat them and make rules about rules, belong in a more sensitive, intensional kind of language, one that respects the difference between required, forbidden, optional, and supererogatory, that is, above and beyond duty, actions.