Why Does ‘Ought’ Play by Different Rules? A Journey into Deontic Logic

The Puzzle of Broken Promises

Imagine you promise your best friend you’ll help them build a model rocket on Saturday afternoon. Later that same day, without thinking, you promise another friend you’ll meet them at the park to play soccer at exactly the same time. When Saturday comes, you realize you can’t do both. You feel a heavy “must” toward each promise. But can both truly be obligatory if it’s impossible to fulfill them together?

This is not just a messy social moment. It’s a doorway into deontic logic, the branch of logic that studies the meaning and rules of words like obligatory, permissible, impermissible, and optional. Deontic logic asks: what can we correctly infer when we reason about duties and rules? And it turns out the answers are weirder—and more important—than you might guess.

Philosophers and logicians have been puzzling over deontic notions for centuries. In medieval Islamic thought and later in Europe, scholars noticed a striking similarity between the way we talk about duties and the way we talk about what must be true. That similarity launched a project to build a logic of obligation modeled on the logic of necessity and possibility. But as we’ll see, the fit is not perfect—and the cracks are where things get really interesting.

The Secret Twin: Necessity and Possibility

Before we can understand the logic of “ought,” we need to meet its older twin: alethic modal logic. Alethic modal logic is the logic of necessary truth and related notions. Consider these sentences:

• It is necessary that 2+2=4.
• It is possible that it will rain tomorrow.
• It is impossible for a square to be circular.
• It is contingent (neither necessary nor impossible) that you are wearing blue.

Logicians use a box (□) for “necessarily” and a diamond (◇) for “possibly,” and they define the others from these. For instance, “impossible” means “□ not,” and “contingent” means “◇ and ◇ not.” They also noticed a tidy pattern: the modal square of opposition shows how necessary, possible, impossible, and non-necessary statements oppose each other, much like the corners of a square.

This square reads like rules of a game. If “necessarily p” is true, then “possibly p” is true, and “impossible p” is false. If “possibly p” is false, then “necessarily not p” must be true. These relationships are so reliable that logicians treat them as axioms—starting points for deduction.

Around the tenth century, Islamic philosophers and, later, medieval European thinkers began to see that normative words like “obligatory” and “permissible” fit a nearly identical square. It was as if the logic of duties had been hiding in plain sight, mirroring the logic of truth.

The Deontic Square: Obligatory, Permissible, Forbidden

Swap “necessary” with obligatory (OB), and “possible” with permissible (PE), and you get the Traditional Definitional Scheme. Something is impermissible (IM) exactly if its negation is obligatory. Something is omissible exactly if it is not obligatory. And something is optional (OP) if neither it nor its negation is obligatory—so it’s up to you. This scheme appears in the work of William of Ockham (c. 1287–1347) and later, very clearly, in Gottfried Wilhelm Leibniz (1646–1716).

Here is the heart of the analogy:

• OB p → PE p  (if obliged, then permitted)
• IM p → OM p  (if forbidden, then omissible)
• ¬(OB p & IM p)  (nothing is both obligatory and forbidden)
• OB p ↔ ¬PE ¬p  (obligatory = not permitted not)
• PE p ↔ ¬OB ¬p  (permitted = not obliged not)

These make intuitive sense. If you must help your friend, you’re certainly allowed to. If it’s forbidden to run in the hallway, then it’s omissible—you don’t have to do it. And the rules can be packed into a deontic square that looks just like the modal square, with “obligatory” and “permissible” at opposite corners.

The Traditional Threefold Classification follows neatly: every proposition is either obligatory, optional, or impermissible—never more than one. This tripartite division mirrors the alethic split into necessary, contingent, and impossible truths. It’s elegant, and for a long time philosophers thought it captured all there was to say.

Standard Deontic Logic: Rules for Obligation

In 1951, the Finnish philosopher Georg Henrik von Wright (1916–2003) launched symbolic deontic logic as a full branch of modern logic. The system that became most influential is called Standard Deontic Logic (SDL). SDL takes classical propositional logic and adds a single primitive operator, OB (“it is obligatory that…”), along with a few special axioms.

The key deontic principles in SDL are:

1. OB-K: If a conditional “if p then q” is obligatory, and p is obligatory, then q is obligatory too.
2. NC (No Conflicts): Nothing can be both obligatory and forbidden: OB p → ¬OB ¬p.
3. OB-NEC: If something is a logical theorem (like “it is raining or it is not raining”), then it is obligatory.

The last one surprises many people: it means some purely logical truths are automatically obligatory, which seems odd if you think obligations are about real actions. But logicians often accept this as a technical convenience, because it makes SDL fit neatly into the well-studied family of normal modal logics.

What does “obligatory” mean in SDL? Here the semantics uses possible worlds. Imagine a huge set of possible ways the world could be. Among those worlds, some are acceptable from the standpoint of a given world—worlds where all your current duties are fulfilled. A proposition p is obligatory at a world i if and only if p is true in every i-acceptable world. It’s permissible if p is true in some acceptable world; impermissible if p is true in no acceptable world.

This picture relies on seriality: there must always be at least one acceptable world—otherwise every p would be trivially obligatory and forbidden at once, which NC prevents. You can visualize the acceptable worlds as a bubble in logical space. Obligation then just says: p holds everywhere inside that bubble.

SDL validates the Traditional Scheme and the deontic square beautifully. But life isn’t always so tidy.

When Duties Collide: Chisholm’s Puzzle

In 1963, the philosopher Roderick Chisholm (1916–1999) presented a now-famous puzzle. Consider these four sentences:

1. Jones ought to go help his neighbors.
2. It ought to be that if Jones goes, he tells them he’s coming.
3. If Jones doesn’t go, he ought not to tell them he’s coming.
4. Jones doesn’t go.

These four seem perfectly consistent and logically independent. Yet, if you translate them directly into SDL, you get a mess. With (1′) OB g, (2′) OB(g → t), (3′) ¬g → OB ¬t, and (4′) ¬g, you can derive both OB t and OB ¬t using OB-K and modus ponens—a flat contradiction. Alternative uniform translations lose either independence or consistency. SDL simply cannot represent the third sentence, a contrary-to-duty obligation (what you should do if you violate a primary duty), without breaking.

This puzzle exposed a deep rift. Two main camps emerged. The factual detachment camp says: from a conditional obligation “if p then it ought to be that q” and the fact that p, we can detach “it ought to be that q.” But this leads to the problem that if you violate a duty, you still get a new actual obligation—like killing your mother gently if you kill her. The deontic detachment camp insists that conditional obligations are a primitive, unbreakable unit (written OB(q|p)), where you can only detach when the antecedent itself is obligatory. This preserves the idea that even when you go wrong, the original “ideal” obligations don’t vanish, but it raises the question: when are conditional obligations ever useful for guiding action?

Neither side won outright. Instead, deontic logicians have developed a zoo of enriched systems: dyadic obligation operators, preference-based orderings of worlds, and non-classical logics that abandon some SDL axioms. Other paradoxes, like Ross’s Paradox (from “you must mail the letter” SDL infers “you must mail or burn the letter”) and the Good Samaritan Paradox (from “you must help the robbed man” SDL infers “the man must be robbed”), show that the simple inheritance principle OB-RM—if p entails q, then OB p entails OB q—is also problematic.

Why It Still Matters: Reasoning Through Real-Life Rules

You might think this is all just a game for logicians with too much time. But deontic logic is at work whenever you think about rules, promises, or laws. When a parent says, “If you finish your homework, you may watch TV,” and you haven’t finished, does it follow that you may not watch TV? When a school code says, “No running in the halls,” but you run to help a fallen classmate, is there a clash of duties? How should a self-driving car’s software weigh conflicting rules in an emergency?

These questions need a logic of obligation that can handle exceptions, conflicts, and conditional duties. Deontic logic today has expanded to include agency (who does what), time (duties change), and epistemic states (what you know). It connects with law, ethics, artificial intelligence, and game theory. The early dream of a simple perfect square gave way to a rich field full of live arguments—and that’s exactly what makes philosophy tick.

The next time you make a promise you can’t keep, or weigh two “musts,” you’ll be standing in the middle of deontic logic. And you’ll know that even the simplest “ought” has a hidden skeleton of rules—rules that philosophers are still mapping.

Think about it

1. If you make two promises and can only keep one, do you still have an obligation to do the thing you can’t do, or does that obligation vanish? What would a friend say?
2. A teacher says, “If you’re late, you must sign the late book.” Suppose you’re never late. Does that rule create any real duty for you? Why or why not?
3. Should a good legal system allow genuine conflicts of obligation, or should it always be designed so that duties never clash? Give an example.