Of Norms

Jaap Hage

Universities of Maastricht and Hasselt





This contribution aims to clarify the notion of a norm by elaborating the idea that norms are rules that lead to deontic consequences. The elaboration focuses both on the nature of rules and on the nature of deontic facts.

Rules, it will be argued, are a kind of constraints on possible worlds. They determine which facts necessarily go together or cannot go together. Three kinds of rules are distinguished: dynamic rules which attach consequences to the occurrence of events, fact-to-fact rules which attach one fact to the presence of some other fact, and counts-as rules, which make that some things (often events) also counts as something else.

The very existence of a rule makes that some fact obtains: the factual counterpart of the rule. The descriptive sentence that expresses this fact – the descriptive counterpart of the rule - is an open generalization and this generalization often  has the same formulation as the rule from which it derives its truth. Many things that have been defended in connection with the logic of norms are better defendable when interpreted as dealing with these descriptive counterparts.

By distinguishing between objective facts, brute social facts, and rule-based facts and thereby introducing facts that are mind-dependent, an attempt is made to overcome resistance against the idea that facts might be normative, that there might be deontic facts. Deontic facts are facts that tend to induce a motivation to comply with them in agents to which they apply. They are most often the result of the application of fact-to-fact rules (duties) or dynamic rules (obligations). However, they can also be the result of mere acceptance, or perhaps even exist because that is rational. A distinction is made between two kinds of basic deontic facts - the existence of duties and of obligations – and two kinds of supervening deontic facts: being obligated and owing to do something.


Keywords: being obligated, competence, constraint, direction of fit, duty, fact, norm, normativity, obligation, ought to be, ought to do, permission, possible world, power, rule

1.      Terminology and Overview

Norms play a central role in practical reasoning, in law as well as in morality. An understanding of the nature of norms is therefore desirable for anyone who is theoretically engaged in practical reasoning, but such an understanding is not easy to achieve. The first challenge one encounters when trying to give an account of norms and their nature is that the terminology around norms is rich, to state it mildly. Von Wright starts his classic Norm and Action  with the remark

The word ‘norm’ in English, and the corresponding word in other languages, is used in many senses and often with an unclear meaning (Von Wright 1963, p. 1).

Ullmann-Margarit  describes a social norm as ‘a prescribed guide for conduct or action which is generally complied with by the members of a society’ (Ullmann-Margalit 1977, p. 12). However, she adds in a footnote that the term ‘norm’ tends to be used by authors with a Continental background, where authors with an Anglo-Saxon background prefer the terms ‘law’ and ‘rule’ (ibidem).

Kelsen writes at the very beginning of his posthumous study Allgemeine Theorie der Normen that the word ‘norm’ denotes in the first place a command (Gebot), a prescription (Vorschrift), or an order (Befehl). He hastens to add, however, that ordering is not the only function of norms, but that empowering, allowing, and derogating are also functions of norms (Kelsen 1979, p. 1).

Although norms both play a role in law and in morality, the term ‘norm’ has gained more popularity in the former field than in the latter. The term is also used more frequently in research from countries with a Roman law or Scandinavian background than in research from countries in the common law tradition. The reader may notice that these biases are reflected to some extent in the present contribution.

Norms are both related to normativity and to rules. However, the notions of normativity and rules are as hard to pin down as the notion of a norm is. For example, the words ‘norm’ and ‘rule’ are sometimes used interchangeably for something that is normative. To clear out matters, this contribution will sharply distinguish between two oppositions: normative – non-normative and rule – description, and it will propose to use the term ‘norm’ for normative rules only. In doing so it must inevitably deviate to some extent from standard word use, if there is something such as standard word use on these issues. The two distinctions will in the following be used as a framework for discussing not only norms as they are defined here, but also related phenomena that historically also have been called ‘norms’.

The main distinctions will be explained in the sections 2 and 3 of this contribution: section 2 deals with the nature of normativity, while section 3 focuses on the nature of rules. Norms are often opposed to facts, because the former would be normative while the latter would not.  It will be argued that the distinction normative – non-normative is not the proper basis to distinguish between norms and facts and to that purpose the sections 4 and 5 respectively discuss different kinds of facts and more in particular deontic facts such as the existence of duties and obligations. Section 6 returns to rules and distinguishes three different kinds of rules. The distinction is then used to identify norms in the strict sense defined here and to discuss the related phenomenon of rules that confer competence and other forms of legal status. The contribution will be summarized in section 7.

2.      Normativity

Norms[1] are basically used for two purposes. The one purpose is to evaluate states of affairs and acts, the other is to guide human behavior.[2] In this contribution the emphasis will be on the second function, but in order to avoid possible confusions it is useful to say a little here about the evaluative function of norms.

2.1      Ought-to-be and ought-to-do

It has become customary to distinguish between norms of the ought-to-do type and the ought-to-be type.[3] A norm of the ought-to-do type tells us what to do, while a norm of the ought-to-be type informs us what should ideally be the case, without specifying that somebody should do something. An example of an ought-to-do norm is the norm that house owners should clear away the snow from the pavement before their houses. This norm specifies that something should be done, and also indicates who should do it. An example of an ought-to-be norm would be that letters ought to be stamped. This norm does not specify that some action ought to be undertaken, let alone who is responsible for undertaking this action.[4]

While ought to be norms indicate what should ideally be the case, there is no logical connection between what ought to be done and what is ideally the case. If there is to be a connection between what ought to be done and what is ideal, this connection must be created by some perfectionist theory of practical reasoning, such as utilitarianism.

Since ought-to-be norms do not specify what ought to be done, they have no use in guiding behavior; they can only be used to evaluate states of affairs as right (in accordance with the norm) or wrong (in violation of the norm). Norms of the ought-to-do type, on the contrary, can both be used to guide behavior and to evaluate it. The norm that house owners should clear away the snow from the pavement before their houses directs house owners to clear away snow. Looking backwards it can be used to evaluate the snow clearing act of a house owner as right. Looking forward, it can be used in justifying the judgment that it would be wrong if the house owners in a particular street would not clear away the snow. Notice that although both ought-to-be and ought-to-do norms can be used to evaluate, the former will be used to evaluate states of affairs, while the latter will be used to evaluate behavior.[5] Ought-to-be and ought-to-do norms have in common that they underlie binary evaluations in terms of right and wrong, and not grading evaluations in terms of better and worse.

Because the emphasis of this contribution will be on norms that guide behavior and since only norms of the ought-to-do type can guide behavior, ought-to-be norms will be left out of consideration from here on.

2.2      Influencing and guiding behavior

One function of norms is to guide human behavior. To guide behavior is not the same thing as to influence behavior, although there is an important connection between the two. If Adrian influences the behavior of Bernadette, Adrian does something that exerts a causal influence on what Bernadette does. For instance, because the traffic is heavy, Adrian clutches his 6-year old daughter Bernadette to prevent her from crossing the street. In this way Adrian influences his daughter’s behavior by making it impossible. Bernadette has no choice whether she will cross the street.

Another way to withhold Bernadette from crossing would be to warn her. If Adrian warns his daughter not to cross the street, he leaves the choice to Bernadette, but tries to influence the choice that she will make. This influence is a causal relation between performing the speech act of warning and the motivation of Bernadette.

It would be quite similar if Adrian got frightened when he saw his daughter approaching the busy street and yells ‘Stop’ to her. Again this is a speech act aimed at exerting a causal influence on Bernadette’s motivation. The type of the speech act might be described as ‘giving an order’, but it should be noted that this order should not be seen as the imposition of a duty on Bernadette not to cross, but merely as an attempt to causally influence Bernadette’s behavior.

Let us assume that Adrian, being Bernadette’s father, has some authority over his daughter, and that he can impose duties on her. Suppose moreover that Adrian exercises this power and forbids Bernadette to cross the street. The ultimate purpose of this prohibition is to withhold Bernadette from crossing the street, and in this respect the speech act of prohibiting is similar to the issuing of a warning, or a mere order. However, there is an important difference between on the one hand the prohibition, and on the other hand the order (and the warning, for that matter). The order is merely an attempt to causally influence Bernadette’s behavior, with the causal relation being between the performance of the speech act and the motivation to refrain from crossing the street. The prohibition is a way to impose a duty upon Bernadette, and this duty also exists if Bernadette is not motivated to comply and crosses the street nevertheless. Moreover, the relation between the speech act of prohibiting and its consequence, the existence of a duty, is conventional, not causal. The causal influence between the prohibition and Bernadette’s behavior, if it exists, goes via Bernadette’s knowledge that she has a duty not to cross the street to her being motivated not to cross. Duties themselves do not motivate, but the awareness of an existing duty may.

The existence of a duty not to cross the street is a reason that applies to Bernadette -  and in that sense is a reason for Bernadette – not to cross.[6] Reasons directly guide behavior by telling what is the right thing to do, and one can indirectly guide behavior by creating reasons. In our example Adrian would guide the behavior of Bernadette indirectly by prohibiting her to cross the street. In doing this, Adrian creates a duty and therewith a reason for Bernadette to refrain from crossing and it is this reason that directly guides Bernadette’s behavior. Notice that this guidance by the reason is not a causal influence. It is still possible that Bernadette ignores the reason and is not at all motivated by it. If that happens, the existence of the reason does not causally influence Bernadette. Still the guidance exists; it consists in an indication of what is the right or the good thing to do, and if Bernadette does not act on the reason most likely she did something wrong.[7]

2.3      Guidance by norms; the second-person point of view

Norms are not the same things as reasons for action, but they are closely related. Suppose that the norm exists that pedestrians are not allowed to cross the street if the traffic lights for pedestrians are red. In that case the fact that the traffic lights are red is a reason for Adrian not to cross the street. The norm generates reasons in all cases to which it applies, and in that sense the norm guides behavior in a general way.

The presence of reasons for action is typically expressed by the use of ‘normative’ words, such as ‘shall’, ‘should’, ‘must’, ‘obliged’, ‘obligated’, ‘duty’, ‘obligation’, ‘forbidden’, ‘prohibited’, ‘permitted’, ‘allowed’ and ‘ought’. Some of these words express a situation in which not only an agent  should do something, but also somebody (else) is entitled to claim from the agent that he[8] acts in a particular way. The entitlement to such a claim, which can sometimes be enforced, is characteristic for moral and legal norms.

Norms are characterized by what Darwall called ‘the second-person standpoint’. Darwall described this second-person standpoint as ‘... the perspective you and I take up when we make and acknowledge claims on one another’s conduct and will’ (Darwall 2006, p. 3). This standpoint is characteristic for both legal and moral norms, but seems to be lacking for many prudential reasons. For example, if Bertie is thirsty she has a reason to take a drink, but – barring exceptional circumstances – nobody, not even she herself, can claim from her that she takes a drink. Law, morality and prudence all provide agents with reasons for action, but law and morality are normative in a sense in which prudence typically is not.

The normativity of norms does not only involve that norms indicate what should be done, but also that they can indicate how things should be done. Assume by way of example that there is a norm prescribing that one should eat asparagus with one’s fingers, rather than with fork and knife. Of course there is no duty to eat asparagus, but somebody who eats asparagus should do so with his fingers. If he uses fork and knife he does something that is wrong. These ‘how-to norms’ should be distinguished from the ‘technical norms’ that specify how something should be done in order to succeed in bringing about a particular result. Somebody who does not eat asparagus with his fingers still succeeds in eating asparagus, but somebody who tries to make a last will without witnesses will normally not succeed in making  last will. We will return to ‘how-to norms’ in section 5.4.

2.4      Norms and facts

Perhaps this is the right moment to briefly address the distinction that is traditionally made between norms and facts, a distinction that is often traced back to the work of Hume. A typical use of norms, in particular ought-to-do norms, is to evaluate acts. For example, the norm that house owners are to clean away the snow before their houses can be used to evaluate the cleaning as right or correct. Norms can only fulfill this function if they are somehow different from the acts that they are used to evaluate. The fact that house owners tend to clear their pavements does not coincide with the duty for house owners to do so, and it does not even have to be evidence for the existence of such a duty. If this distinction between the norm and the behavior that does or does not conform to this norm is meant by the distinction between norm and fact, it is obvious that the distinction between fact and norm exists.

However, it sometimes seems that the distinction between norm and fact is considered to be much more prominent, as if it were a major ontological divide between what is ‘out there’ independent of human beings and their minds, and what is added by human minds to what ‘really’ exists. This major ontological difference does not exist, or – if it is assumed nevertheless (see the discussion of ‘objective facts’ in section 4.1) – it is of limited importance. The distinction between deontic (normative) and non-deontic (non-normative)[9] is real, but has no major ontological relevance.[10] It is comparable to the epistemic  difference between what is certainly the case and what is the case. However, in section 3 a different distinction will be made which does have major ontological relevance. This is the threefold distinction between facts, constraints on facts, and descriptions of facts. This distinction, which has nothing to do with the distinction between deontic and non-deontic, is highly relevant for a proper understanding of rules and therefore also of norms as deontic rules.

3.      Rules as Soft Constraints on Possible Words

Often the notion of a rule is connected to the guidance of behavior: rules would indicate what we should do. On this interpretation the meanings of the terms ‘rule’ and ‘norm’ would practically coincide. And yet there are ‘rules’ whose primary function does not seem to be to guide behavior, and which can therefore only be ‘followed’ in a broad interpretation of that term. Examples would be rules that confer competences and rules which make that something also counts as something else.[11] A proper account of the nature of rules would explain why some rules can be complied with and other rules cannot. Such an account would also clarify the nature of norms as a special kind of rules. The first step in providing such an account is to go into some detail concerning ‘directions of fit’.

3.1      Directions of fit

Perhaps the best way to introduce the distinction between directions of fit is by means of an example of Anscombe's (Anscombe 1957, 56). Suppose that Elisabeth makes a shopping list, which she uses in the supermarket to put items in her trolley. A detective follows her and makes a list of everything that she puts in her trolley. After Elisabeth is finished, the list of the detective will be identical to her shopping list. However, the lists had different functions. If Elisabeth uses the list correctly, she places exactly those items in her trolley that are indicated on the list. Her behavior is to be adapted to what is on her list. In the case of the detective it is just the other way round; the list should reflect Elisabeth’s shopping behavior. The two different functions of the list with regard to Elisabeth’s behavior reflect the two different directions of fit that we are looking for.

The two items involved in Anscombe’s example are a linguistic one, the list of items, and the world. The directions of fit distinction can also be applied to other items than purely linguistic ones, but let us focus on the purely linguistic case first.

The relation between language and the world goes in two directions. If the linguistic entities are to be adapted to the world, as when the detective writes down which groceries are in the trolley, the fashionable expression is ‘word-to-world direction of fit’ (Searle 1979, 1-30). If the world is to be adapted to the linguistic entities, as when Elisabeth puts those items in her trolley that are mentioned in her shopping list, the fashionable expression is ‘world-to-word direction of fit’.

However, expressive as these expressions ‘world-to-word direction of fit’ and ‘word-to-world direction of fit’ may be, they are also difficult to keep apart. Therefore it is proposed to use the different expressions, ‘down’ and ‘up’. (See figure 1) The basic idea is that descriptive sentences consist of words that aim to fit the world. The propositions expressed by them are true and the speech acts in which they are used are successful in the sense of ‘truthful’, if and only if the facts in the world correspond to (‘fit’), what these propositions express. This is the up direction of fit.

Figure 1: Directions of fit














For the down direction of fit we must distinguish between three kinds. For all three kinds holds that somehow the facts in the world are adapted, in order to ‘fit’ what is expressed by the words. One case is when the words function as a directive, as when Adrian shouts ‘Bernadette, stop!’ when he fears that Bernadette will cross the busy street. This order aims at making its addressee stop, and if the order is successful in the sense of ‘efficacious’, Bernadette will stop and the facts in the world fit the content of the order. In this case the relation between the utterance of the order (the performance of the speech act) and the facts in the world is causal by nature. We might therefore speak of the ‘causal down direction of fit’.

A second case concerns constitutive speech acts, such as ‘I hereby forbid you to cross the street’. If such a prohibition is successful, the facts in the world come to match the content of the speech act and Bernadette has from that moment on the duty not to cross the street. In this case the relation between the performance of the speech act and the facts in the world is constitutive by nature; the performance of the speech act constitutes the duty. We might therefore speak of the ‘constitutive down direction of fit’.

Notice that this down direction of fit relates a speech act to a duty, not to the compliance with the duty. Efficacy is here the coming about of the duty which the speech act aimed to create. The duty itself can also be efficacious in the sense that it is complied with, but that would be an example of the causal down direction of fit.[12]

The third kind of down direction of fit concerns the effects of ‘constraints’. Constraints will be discussed more extensively in section 3.3, but here we will use one kind of constraint as example: the conceptual rule (rule of meaning) that makes that the bachelors are unmarried man. Given this rule, if somebody is a bachelor, he must be unmarried. This ‘must’ depends on the conceptual rule that defines the relation between being a bachelor and being married. Given this rule it cannot be otherwise than that a person who happens to be a bachelor is also unmarried.[13] The facts in the world adapt themselves to the constraint, and that is what is meant by the down direction of fit of constraints.

In section 6.2 we will see that the constitutive down direction of fit is a special case of the down direction of fit of constraints, and more in particular of dynamic rules.

3.2      Possible worlds

We are all familiar with the distinction between what the facts actually are and what the facts might have been. The sun is shining, but it might just as well have been raining. In West-Europe there is peace, but there might have been a war. In the common law, judge-made law plays a paramount role, but its role might have been subordinate to statute-based law.

Logicians use possible worlds-terminology to deal with this distinction between what the facts are and what they might have been. They say for instance that in the actual world the sun is shining, but that in some other possible world it is raining. Intuitively, a possible world is a set of facts which makes some descriptive sentences true, and others false. The set of facts that defines a possible world is complete in the sense that it determines for every non-modal descriptive sentence[14] whether it is true or false. The actual world is one of the many worlds that are possible and in the actual world the sun is shining. However, in some other possible world, it is raining. That is another way of saying that although actually the sun is shining, it might have been raining.

Some things are necessarily the case. For example, five is necessarily bigger than three. If something is necessarily the case, there is no possible world in which it is not the case. In all possible worlds, five is bigger than three. And in all possible worlds, if Janet is either in Berlin or in London, and she is not in London, then she is in Berlin. Being necessary may just as well be circumscribed as being the case in all possible worlds. What is necessary is the case in all possible worlds, while what is impossible is not the case in any possible world. What is contingent is the case in some, but not in all possible worlds.

What makes that a world is possible, and how can we distinguish possible worlds from impossible ones? To answer these questions we need to apply the idea of constraints to possible worlds. Constraints on possible worlds are limitations on which facts can go together and which facts exclude each other. Possible worlds satisfy these constraints, while impossible worlds violate one or more of them.[15] Examples of such constraints are for instance physical laws. A physically possible world is a world that satisfies all physical laws, including the law that metals expand when heated. Since all physically possible worlds satisfy this constraint, in all these worlds pieces of metal expand when heated. The same thing, stated in terms of facts that go together, is that in all physically possible worlds the facts that M is a piece of metal and that M is heated go together with the fact that M expands. So if we only look at physically possible worlds, it is necessarily the case that a piece of metal will expand if it is heated. It is also the case that a piece of metal would have expanded if, counterfactually, it would have been heated.

3.3      Constraints

Necessity and possibility are not absolute phenomena. Something is always necessary or possible relative to some set of constraints.[16] If all constraints, including those of logic, are left out of consideration, everything is possible, even what is logically impossible. Not all constraints are physical or logical. There are also conceptual constraints, such as the constraint that bachelors are unmarried males, that a rectangle has straight corners,  and – perhaps more controversial – that gold is a metal. Some constraints seem to defy any category, such as the constraints – if they are actual constraints -  that all colored objects have a surface, that an item cannot simultaneously be at two different places, that every event has a cause, and that causality does not operate backward in time.

We can try to imagine a world that is not constrained in any way. In that world, all facts are independent of each other, as are the truth values of propositions that purport to describe these facts. The truth of one proposition has no connection at all to the truth of any other proposition, and the relations between the truth values of all propositions would be like the relations between the truth values of atomic propositions in propositional logic. Although it may be hard to imagine, the proposition ‘It is now five o’clock and it is raining’ might be true, while at the same time the proposition ‘It is raining’ would be false. A world in which this is the case is logically impossible, but it is still a possible world if the logical constraints are left out of consideration.

A logically possible world is a possible world in which logical constraints determine (not necessarily exclusively) which combinations of facts always hold, and which other combinations of facts never occur. Exactly which combinations of facts are necessary or impossible depends on the precise nature of the logical constraints. One such a constraint is that a fact and the ‘opposite’ of this fact cannot go together. For instance, if the fact that it is raining obtains in a logically possible world, then the fact that it is not raining does not obtain in that world. Or – to say the same thing in term of truth values of propositions – a logically possible world does not make both the propositions ‘It is raining’ and ‘It Is not raining’ true. More in general, a logically possible world does not allow that a proposition and its negation are both true. Nor does it allow that a proposition and its negation are both false. These are both constraints on logically possible worlds.[17]

Constraints may seem mysterious entities, and the question is justified in which manner they exist. An attempt to explain necessity and possibility in terms of constraints seems a bit like the explanation of the sleep inducing nature of opium by pointing to the vis dormitiva, the sleep inducing power, of opium in Molière’s play The Imaginary Invalid.  That constraints somehow exist must be concluded from the fact that some things are necessary and others impossible. If we know that circles are necessarily round, we know something not only about actual circles, but also about the characteristics something would have if it were a circle, that is knowledge about possible circles. Knowledge about necessity is knowledge about hypothetical situations. This knowledge must be a priori (no dependent on sensory perception) and must be based on reasoning. The only feasible explanation that not only actual circles, but also possible circles are round is that the world is constrained in a manner that disallows non-round circles, and that this constraint also applies to the world that contains the possible circles about which we know that they must be round.

At first sight it does not make much sense to search further for the nature of constraints and their mode of existence, but a little bit more can still be said. Take again the roundness of circles. This is often considered to be a conceptual truth. Unless one is a conceptual realist who assumes that concepts exist ‘out there’, to be discovered by intelligent beings, one can assume that concepts are being created by human beings. In particular with regard to artificial concepts, such as ‘computer’, it is plausible that they are human creations and could, at the time of their creation, be arbitrarily defined. It is still possible to modify the concept of a computer, to make it, for instance, include or exclude smart phones.[18] What a computer is, is a matter of convention, and the convention might have been slightly different from what it actually is. However, given the convention as it actually is, computers have some characteristics essentially and necessarily - for instance having one or more processors - and other characteristics - for instance their color - contingently and only possibly. The necessary characteristics of computers are based on a convention that functions as a constraint on what a computer can and cannot be.[19] Apparently at least some constraints are man-made, and rules belong to this category of man-made constraints.

3.4      Rules as soft constraints

Why are rules a kind of constraints? Because they behave in many ways as other constraints. In the world in which a rule exists, the rule imposes itself on the facts of that world with the down direction of fit that other constraints also have. So if some possible world contains the rule that thieves are punishable, then in this world thieves are punishable. Moreover, the rule also supports conditional and counterfactual judgements: if John had been a thief, he would have been punishable.

In a world that contains the rule that thieves are punishable, it  is not merely a contingent matter of fact that thieves are punishable, but a necessary one, because being a thief makes one punishable. In this connection something remarkable is the case. Rules allow for exceptions in the sense that sometimes the consequences of a rule do no hold, even though the conditions of the rule are satisfied. For instance, John, who is a thief, is also a minor and therefore the rule about the punishability of thieves cannot be applied to John. This possibility of exceptions seems hardly compatible with the necessary connection between being a thief and being punishable. And yet the necessity and the exceptions have the same ground, which is that they are based on a constraint. Otherwise than descriptive sentences (see section 4.6), constraints can have exceptions.[20] However, constraints also cover hypothetical and counterfactual situations, and that explains why judgments based on a constraint can express necessary relations such as the relation that thieves are necessarily punishable. Strangely, necessity and exceptions go hand in hand.

Rules have a lot in common with more traditional constraints such as the logical and physical ones, but there are also major differences. One such a difference is that rules only apply locally: the laws of one country are for example different from the laws of another country. The necessity of rule-based judgements seems therefore to be merely local necessity. This is different for logical and physical laws, which seem to have a universal scope of application.[21]

The scope of rules is not only limited in space, but also in time. Many rules can be created or derogated and in that sense they differ from the more traditional constraints which somehow seem outside the scope of human manipulation. When the rule that thieves are punishable is introduced, suddenly all thieves become punishable. And when the rule is repealed again, the punishability of thieves disappears with the rule.

As a consequence of these differences, there can be some logically and physically possible worlds in which a particular rule exists, and other possible worlds in which the same rule does not exist. In a sense it might be said that logical and physical constraints create necessities that are themselves necessary, while rules create contingent necessities. For this reason, rules will be categorized as ‘soft constraints’, as opposed to the hard constraints that do not depend for their existence on human decision making or social practices.[22]

4.      Kinds of Facts

If the notion of a fact is taken broadly as that aspect of reality which makes a true descriptive sentence true, there are many different kinds of facts: facts that exist ‘objectively’, facts that depend on recognition, facts that are the results of rules or the use of reason, facts that are independent of all other facts, facts that ‘supervene’ on other facts, ‘neutral’ facts and facts involving evaluation, ‘inert’ facts and facts that motivate or guide behavior. For a proper understanding of norms, the distinction between ‘inert’ facts and facts that guide behavior may be the most important distinction between kinds of facts, but this distinction cannot be seen separate from many other kinds of distinctions between kinds of facts. Therefore the present section and its subsections are devoted to a number of distinctions between kinds of facts. Their main purpose is to open up conceptual space for ‘deontic facts’, facts that involve that something should, or ought to be the case, of that somebody should or ought (not) do something. These deontic facts are crucial to understand norms.

4.1      Objective facts

Some facts seem to be objective. They include the facts that Mount Everest is a mountain and that it is higher than 8000 meters, that there are lions and other kinds of animals, and that there are N suns, with N being some as yet unknown natural number. The objectivity of these facts lies in their being mind-independent, that is independent of anyone’s beliefs, linguistic practices, conceptual schemes, and so on.

The idea that there are objective facts stems from a distinction that we make within our beliefs. To some beliefs we ascribe a counterpart that somehow exists independently of what we humans think about them. This counterpart would consists of objective facts and the facts mentioned above typically belong to this category. They are distinguished from other facts, which we take to depend on the human mind in some way. For example, many people take it that value judgments do not reflect an independently existing world of values, but rest on the way we humans evaluate things. No doubt evaluation is typically also based on objective characteristics of things, for instance the sharpness of the picture offered by the computer monitor, but these objective characteristics must be combined with a man-made standard to lead to the evaluation that this monitor is a good one. This dependence on a man-made standard makes that the value judgment is not objective and that the fact expressed by it is not objective either. Objective facts are different, however. They are taken to obtain independently, and our knowledge, if it is objective, reflects these objective facts as they really are.

It may be argued that there are no objective facts in the sense of ‘objective’ that is presently at stake. The reason is that every fact is the fact that ..., where the dots are to be completed by some descriptive sentence. Facts depend on language and since language is not mind-independent, facts are not mind-independent either, not even the ‘objective’ ones. Many people would object against this conclusion because there must be something ‘out there’ that precedes human categorization. That would be the ‘real’ facts, and we humans try to develop concepts that fit this pre-linguistic substrate as well as possible. Whether such a pre-linguistic substrate really exists is an open question, but it certainly does not consist of the conceptualized reality in which we humans live. The assumption of a pre-linguistic substrate is the result of theorizing, not a precondition of it.

4.2      Brute social facts

When we are satisfied with a very coarse categorization, social facts may be described as facts which exist because the members of some group collectively recognize or accept them as existing. There are two variants on collective recognition. In the case of what we will call ‘brute social facts’ the facts themselves are recognized by the members of some social group, while in the case of ‘rule-based facts’ the facts are the result of some rule.[23]  The facts based on rules that exist because of social recognition are perhaps better known as ‘institutional facts’.

Brute social facts are the result of collective recognition. Important aspects of collective recognition are that sufficiently many and/or sufficiently important members of a social group believe the fact to be present, believe that the sufficiently many and/or sufficiently important other members also believe the fact to be present, and believe that these mutual beliefs constitute the believed fact.

Suppose that about 20 persons together make a foot trip to the top of a mountain. They believe that they are as a group walking to the mountain top, they believe that the others also believe that they are walking as a group to the top, and they all believe – minimally in the sense of not denying it when asked – that their mutual beliefs about acting together make that they are walking as a group to the mountain top, rather than as a set of individuals. In this case the people in the group make a foot trip to the mountain top as a group. This is a brute social fact.

The above example deals with collective recognition of acting together, but collective recognition does not always deal with collective action. Suppose that one member of the group, say Henriette, utters strong opinions about which path to take to the top of the mountain and that most of the group members tend to act on these opinions. After having several times chosen a particular path because Henriette proposed to take it, most group members recognize the leading role of Henriette. They believe that Henriette has become the group leader, that most other group members hold the same belief, and that Henriette is the leader of the group because she is recognized as such by most group members. In this example, the brute social fact concerns the possession by Henriette of the status of leader of the group.

In the two above examples, recognition took the form of believing. However, sometimes mere believing does not suffice. If the leadership of Henriette in the group of mountain climbers has been sufficiently established, the group members may collectively recognize an order from Henriette as a reason for acting. Suppose that Henriette ordered Susan to walk on the back of the group to see to it that nobody stays behind. Then Susan is considered to be obligated to walk on the back on the basis of collective recognition. This involves she is liable to be criticized by group members, including herself, if she does not walk on the back. Susan is obligated to walk on the back as a result of collective recognition, but Henriette’s competence to create such a duty for Susan is also based on collective recognition. The group members collectively recognize an order from Henriette as creating a duty for the person who was ordered. They do this by collectively recognizing the duties that ensue from the orders. In combination, this amounts to the recognition of a power to create duties by giving orders. In its turn, the power actually exists by being recognized. This is an example of a fact – the existence of a power - that exists as the result of collective recognition, where the recognition does not consist in a belief, but in a complex set of dispositions to act.

4.3      Social rules

If in our example about the mountain climbers the group members normally recognize the duties imposed by the leader of the group, whoever that may be, it may be said that the group has the rule that the group leader can impose duties. This rule exists through being recognized and therefore as a matter of social fact. The difference between having this rule and the recognition of the power of Henriette is the abstraction from the actual person having the power. When a power is not anymore ascribed to a particular person, but to a role – in this case the role of group leader – the acceptance of an ordinary social fact has become the acceptance of a social rule.

 It is tempting to follow Hart (2012, p. 57) in assuming that the existence of a social rule involves the existence of a critical reflective attitude with regard to behavior covered by the rule. This characterization of social rules is quite adequate for rules that prescribe behavior, but less so for other kinds of rules such as power-conferring rules. A broader, and therefore more adequate, characterization of a social rule is that a social rule exists within a group if sufficiently many (sufficiently important) members of the group recognize the consequences of the rule when the rule is applicable. For a mandatory rule this means that sufficiently many group members assume the presence of a duty or obligation if the rule attaches this duty or obligation to an actual fact situation. If the duty or obligation applies to a specific group member, this recognition typically involves that the group member is motivated to comply with the rule. For a power-conferring rule this means that sufficiently many group members recognize the power of a person to whom the rule conferred the power. This recognition typically consists in the recognition of the effects of the exercise of the power. In our example this was illustrated by the group members recognizing the duty that Henriette imposed on Susan.

4.4      Rule-based facts

If the group of mountain climbers has the rule that its leader has the power to create duties for group members, the power of Henriette to impose the duty on Susan to walk on the back is an example of the application of this rule. Henriette’s power exists because of the rule and does normally not require separate recognition of Susan’s duty by the other group members. Because the group has this rule about the powers of its leader, Susan has the duty to walk behind as soon as Henriette has imposed that duty on her.

The fact that Susan has this duty exemplifies a rule-based fact. Rule-based facts are those facts which exist because they were attached by a rule to some other fact, including the occurrence of some event.[24] Law provides telling examples of rule-based facts. Suppose that the parents of Joan own the Blackacre ranch. When they die, Joan inherits the Blackacre ranch and becomes owner of the ranch, at the moment that her parents die, even though it may still take some time before people, including Joan herself, receive the information that this is the case and before people are in a position to recognize that Joan has become the owner.

When Joan becomes the owner of the ranch, she also becomes competent to mortgage the ranch, and to transfer the ownership of the ranch to somebody else. Most likely, the ownership of the ranch also brings for Joan the duty to pay real estate taxes. All these facts obtain solely because of the application of rules to the existing facts. Rules can attach consequences to facts and to events, and these consequences are new facts, rule-based facts.

One kind of rule-based fact is the existence of another rule. Again this phenomenon is particularly important in law where most rules exist because they were explicitly created. That the rule-creating events actually lead to new rules is because other rules attach the consequence that a rule exists to these events. A properly created rule immediately exists, even if its consequences do not receive any recognition yet. However, if the recognition of the rule consequences never occurs, or – in other words – if the rule is completely inefficacious, the rule stops existing. If the term ‘valid’ is used for the existence of rules, this means that rules that belong to a legal system are valid if the system as a whole is efficacious; efficacy of the individual rule is in first instance not required. However, if an individual rule is or becomes inefficacious for a longer stretch of time (desuetudo) this may take away the rule’s validity (Kelsen 1960, p. 215-219). 

Also outside the law rule-based facts play an important role. It becomes easier to recognize this when one sees that standards at the hand of which value judgments are given are also a kind of rules.[25] Suppose that a group uses the standard that a soccer match is good if the play is aggressive but not foul. Then, if some match has aggressive but not foul play, this match is good. This situation is not very different from that of the group that recognizes that Susan has the duty to walk on behind because the group leader said so. The fact that Susan has this duty is just as ‘real’ as the fact that the soccer match is good. Obviously, the existence of evaluative facts depends on a presupposed standard, but this holds for all rule-based facts.

Still another example of rule-based facts are the facts expressed by the theorems of some branch of mathematics, systems of formal logic included. The theorems express facts, they are true propositions, but they derive their truth from the axioms and the semantic rules (the  ‘valuation function’) of the formal system to which they belong.

4.5      Creation and derogation

Because the existence of rules is often rule-based facts, there is a risk that rules are confused with the events by means of which they were created. Legislation is then, for instance, seen as a collection of rules, rather than as a means to create rules.

A similar misunderstanding underlies the view that legal norms are a kind of commands.[26] Such a characterization of norms would be wrong because a command is a speech act and therefore a kind of event, while a norm is a rule and therefore not an event.  The temptation to see norms as a kind of commands may be explained from the shared normativity of commands and norms; it almost seems as if the norms command to act in a particular way. However, a proper understanding of the mode of existence of norms should focus on norms being a kind of rules, rather than on the specific kind of rules that norms are.

Basically the same kind of mistake is made if the existence of norms is somehow connected to a legislator.[27] Many norms have been created intentionally by some authority, but being a norm, not even being a legal norm, is not the same as being created by a state authority or any other kind of authority.

The speech acts by means of which some norms are created should not be identified with the norms created by them, or with any other kind of norm. Mutatis mutandis this also holds for speech acts by means of which norms are derogated or repealed. By passing a bill, a legislator can derogate existing norms, but that does not make the bill into a norm. The idea that there are derogating norms therefore rests on a mistake.[28]

4.6      Factual and descriptive counterparts of rules

The rule that thieves are punishable makes it impossible that thieves are not punishable.[29] Or, to state the same thing affirmatively, the rule necessitates that thieves are punishable. The (existence of) the rule that car drivers must drive on the right makes that car drivers have the duty to drive on the right. And the rule that cars count as vehicles for the Traffic Law, makes that cars are (count as) vehicles.

If some rule – or, more in general, a constraint – exists, this means that some general descriptive sentence will be true. This sentence has more or less the same formulation as the rule, but it is not the rule formulation but a sentence that aims to provide information about the facts. Since these facts obtain because of the constraint, this sentence will be true.

Such sentences are open generalizations. An open generalization is a generalization over potentially infinitely many items. Examples would be that pedestrians wear shoes (a false open generalization), and that atoms have a nucleus (a true open generalization). Examples of closed generalizations are that all desks in this class-room are brown, and that all instances of the Olympic Games lasted less than four months. Open generalizations can have counter-instances and still be true. For example, the open generalization that birds can fly is true, notwithstanding the existence of ostriches. A counter-instance to a closed generalization falsifies this generalization.

A rule of thumb to distinguish open from closed generalizations is that a closed generalization requires the use of the word ‘all’, or some equivalent, while this word can be left away in case of open generalizations. For example, the sentence ‘Desks in this classroom are brown’ expresses an open generalization  which is almost certainly false, even if all desks in the classroom happen to be brown. (“Happen’ is another word that indicates a closed generalization .) The open generalization  requires for its truth a law-like connection (a constraint) between being a desk in the classroom and being brown. Interestingly, it is precisely this law-like connection which makes that open generalizations can have counter-instances and be still true.[30]

The open generalizations that describe the effects of rules typically have the same formulation as the rule the effects of which they describe, and they are true because that rule exists. They describe facts that will be called the ‘factual counterpart’ of the rule, and they may themselves be called the ‘descriptive counterparts’ of rules. These descriptive counterparts of rules describe facts that are based on the existence of rules, but mediated by the rule-based facts that are immediately based on the rules.

Where rules impose themselves on the world by way of their down direction of fit, but are not true or false, the descriptive counterparts of rules are descriptive sentences, which are true or false, usually depending on the existence of the rules of which they are the counterpart. In schema:



















4.7      Norm-propositions

The descriptive counterparts of norms have some resemblance with what are in the literature on deontic logic sometimes called ‘norm-propositions’. Von Wright used the term ‘norm-proposition’ for a proposition stating that a particular norm exists. Apart from the observation that Von Wright apparently saw norms as a kind of entities that can exist, that is as a kind of logical individuals (see section 6.1), these ‘norm-propositions’ are not at all similar to descriptive counterparts of rules. Where norm-propositions in the sense of Von Wright talk about norms, descriptive counterparts talk about the subjects of norms (rules). For example, the descriptive counterparts of the norm ‘No vehicles in the park’ is about vehicles, not about a norm.

Alchourrón and Bulygin (1981) and later Navarro and Rodríguez (2014) seem to follow Von Wright in calling statements about (the existence of) norms ‘norm-propositions’, but they add that norm-propositions are statements about what is mandatory, prohibited or permitted relative to some set of norms. Since these statements are statements about prescribed, prohibited or permitted states of affairs or actions, they are not statements about norms, so this identification seems to be based on a confusion. However, they are correct in pointing out that there are descriptive sentences, made true by existing norms (better: rules), stating that particular kinds of actions have a particular deontic status. Such statements describe, if they deal with individual acts, or with acts to be performed by a specific agent, rule-based deontic facts. Examples are the descriptive sentences ‘This killing was permitted’ (based on the license to kill for secret agents), and ‘John must clear away the now from the pavement before his house’ (based on the rule prescribing house-owners to clean away the snow before their houses). If the truth of these descriptive sentences is based on legal rules, these sentences would also express what Kelsen called ‘Rechtssätze’ (Kelsen 1960, p. 57).

If such statements describe general prescriptions, prohibitions or permissions, they are descriptive counterparts of rules. Examples would be ‘House-owners must clear away the now from the pavement before his house’, and ‘Secret agents are licensed to kill in the performance of her majesty’s secret service’.  Kelsen called these general descriptive sentences Rechtssätze too (Kelsen 1960, p. 85), but also ‘legal rules’ (Kelsen 1945, p. 45).

4.8      ‘Entailed’ norms

The recognition of descriptive counterparts of rules is important, because they are ‘ordinary’ descriptive sentences to which deductive logic is applicable, whereas the applicability of deductive logic to rules, and to norms as a species of rules, is dubious since rules are from a logical point of view individuals (see section 6.1). Many inferences which seem to have rules as their conclusions may well be interpreted as arguments with descriptive counterparts of rules as their conclusions. For example, the argument ‘Volkswagens count as cars. Cars owners must pay road tax. Therefore  owners of Volkswagens must pay road tax’ is dubious as an argument in which a rule is derived from two other rules. As an argument in which two descriptive counterparts of rules are used to derive another open generalization, it is valid.[31]

The possibility to derive open generalizations from other open generalizations also explains the phenomenon of ‘deontic inheritance’ (Hage 2001), ‘entailed norms’ (Navarro and Rodríguez 2015) or ‘normative consequences’ (Araszkiewicz and Pleszka 2015). An example would be that a prohibition for vehicles in the part would entail a prohibition for Volkswagens in the park. It is highly dubitable whether these entailed ‘norms’ are rules that can be traced back to some official legal source. However it is obvious that if vehicles are prohibited in the park, then – normally speaking – Volkswagens will be prohibited as special case of this general prohibition. The one deontic fact - vehicles being prohibited – encompasses the other – Volkswagens being prohibited - and there is no objection against deriving the proposition that expresses the latter fact from the proposition expressing the former fact. However, interpreting such a derivation as the derivation of one norm from, amongst others, another norm  would be misguided (Hansen 2013).

5.        Deontic Facts

5.1      Deontic facts and motivation

Norms are normative because they lead to duties or obligations.[32] Duties and obligations are entities that exist in time but not in space. They can be created and destroyed, and they can have all kinds of characteristics such as being a nuisance, or being suitable to deal with societal problems.

The existence of a duty or an obligation is a deontic fact. It is a fact about an immaterial ‘thing’, a duty or an obligation, and it is deontic (normative) in the sense that it guides behavior. Suppose that Susan and Thera have concluded a labor contract. After that event, Susan has an obligation toward Thera to pay her a monthly salary, while Thera has an obligation toward Susan to work the afternoons of all week days. The sentences describing the existence of these obligations are true just like any other descriptive sentences. Facts that involve the existence of a duty or an obligation, and also some other kinds of facts, including the existence of permissions, may be called ‘deontic facts’, after the convention that has arisen in logic to call logics that deal with duties, and obligations and with everything else that ought to be done ‘deontic logic’.

Even though the existence of a duty or an obligation is a fact, it is also deontic, normative, or behavior guiding, whatever you may want to call it. Although this is not the place to go into details with regard to the nature of normativity (however, see section 2), it may nevertheless be useful to say a little about why duties (and obligations) are normative.[33] The normativity of duties lies in the connection, however remote that may sometimes be, between the existence of a duty and the motivation of persons to act in a particular way.[34] Typically the acting person is the holder of the duty, and the behavior at issue is compliance with the duty. Agents tend to have a disposition to comply with their duties.[35] Although it is possible for an agent to recognize that he has a duty without being motivated at all, it is not possible that agents typically would not be motivated when they recognized to have duties. The reason is that if duty holders would not normally be motivated to comply with their duties, the very notion of a duty would not make sense.

The existence of a duty is not only based on the behavior or the disposition thereto of the duty holder;  the behavior of agents  in the environment of the duty holder is relevant too. This behavior consists of praise – in case the duty was complied with – or blame – in case the duty was violated, or – to use Hart’s phrase – of a critical reflective attitude.

Sometimes the connection between a duty and the motivation to act is indirect. That is for instance the case with duties based on rules which exist themselves as a matter of rule-based fact, such as legal duties. Then the disposition to comply with duties is the disposition of the addressees of the normative system as a whole to comply with duties based on the normative system. It is not the case anymore that every duty based on the system must lead to a disposition for compliance. Moreover, the efficacy of the system as a whole – because that is what we are talking about – must consist in recognition of the consequences of the system’s rules, and since the rules are not always mandatory, the required efficacy is not always compliance with duties. It can, for instance, also be recognition of the power to make rules.

The connection between a deontic fact such as the existence of a duty, and behavior may be quite complicated, but first, normativity cannot exist without such a connection, and second, there is nothing more to normativity than this connection. There is, for example, no such a thing as ‘binding  force’ apart from the disposition to motivate. That means that there is, for instance, no need to postulate the existence of a ‘norm’ next to the rule-based facts that persons have certain duties, and the constitutive rules on which these facts are based.[36]

5.2      Duties and obligations

In normative discussions, words like ‘ought’, ‘should’, ‘must’, ‘duty’, and ‘obligation’ are often used interchangeably. Although the meanings of these words in natural language are not fixed and overlapping, there exist different categories of deontic facts. The differences between these categories are important, although the words used to denote them are not. In the following, we will consider some distinctions and adopt particular words to denote the newly delineated categories. Let us start with some examples:

-     Everybody has the duty not to steal, but normally there is no obligation to that effect.[37]

-     A and B are under obligations towards each other, because they entered into a sales contract, but these obligations are not duties.

-     From the fact that P is under an obligation or a duty to do something, it follows, pro tanto – that is, if only this reason is taken into account - that P ought to do it, but not the other way round.

The first distinction to be made is between duties and obligations. Both the existence of a duty and of an obligation is a reason why somebody ought to do something, but neither the duty, nor the obligation coincides with the fact that this person ought to do it. A duty is often connected to a role or status. It is for instance the duty of house owners to pay real estate tax, and the duty of a Mayor to maintain the public order in a municipality. All human beings[38] are under a duty not to kill other human beings. However, as our example of Bernadette and Adrian (section 2.2) illustrated, it is possible to have a duty as the result of a command, and such a duty is not connected to a particular role or status.

Whereas duties are often connected to a particular status or role, an obligation is the outcome of an event and depends on that event having occurred. Typical examples of such obligation generating events are causing damage, making a promise, or contracting. Moreover, whereas a duty is not a duty with regard to somebody in particular, obligations are always ‘directed’, obligations towards somebody else.[39] This directedness of obligations still holds if this ‘somebody else’ is (as yet) unknown, as when for instance a car was unlawfully damaged but the owner of the car is still unknown. Duties are not directed in this way[40]

5.2      Being obligated and owing to do something

The term ‘obligated’ will be used here as a term of art to denote the common denominator of duties and obligations: a person who is under a duty to do B is obligated to do B; a person who is under an obligation to do B is also obligated to do B. Being obligated is not directed. If A has contracted with B to pay him €100, then A has an obligation toward B to pay him €100, and A is also obligated to pay B €100, but A is not obligated toward B to pay B €100.

By now we have encountered three normative concepts, ‘duty’, ‘obligation’, and ‘being obligated’. They all differ from the normative concept that is often used as a catch-all for all kinds of normativity, the concept of ‘ought’. In connection with duties, obligations and being obligated, the more relevant notion is ought-to-do. The word ‘ought’ as defined here stands for the outcome of the interplay of one or more reasons for acting, a kind of aggregate of these reasons. Examples are the legal ought, as the aggregate of legal reasons for action, and the moral ought as the result of the aggregate of moral reasons.

An ought itself is not a reason for acting, but merely the outcome of one or more reasons. So, where the fact that X is under a duty to pay real-estate tax is a  reason why X ought to pay real-estate tax, the fact that X ought to pay the tax is not a reason for paying it, although it presupposes the existence of such a reason (the duty, for example). An ought is comparable to being obligated in the sense that it abstracts from the precise reasons for acting, but nevertheless indicates (through presupposition) that such reasons exist. Where being obligated is tied to precisely one such a reason, owing to do something is based on a set of reasons, even though this set may contain one reason only. Being obligated can therefore be seen as a pro tanto ought.[41]

The difference between, for instance, an obligation, and the ought based upon it becomes clear if one considers what happens in case it is impossible to perform one’s obligation. For instance, if Antony contracts with Giovanni to transfer his car to Giovanni, and if he also contracts with Guido to transfer his car to him, then Antony both has an obligation towards Giovanni and towards Guido. It is impossible for Antony to comply with both obligations, and therefore[42] it is not the case that Antony both ought to transfer the car to Giovanni and to Guido. The law has a simple solution for such cases. Both obligations have an equal status (paritas creditorum). If Antony complies with his obligation to Giovanni, he must default on his obligation toward Guido, and – because the obligation still exists – Antony must compensate the damage of Guido. The question which obligation supersedes the other has a clear answer: neither one of the obligations as such supersedes. However, in determining what Antony ought to do, the reasons for acting have to be balanced. If the above account of the legal situation is correct, the outcome will be that Antony is legally permitted to deliver the car to any one of his creditors[43], and that he will have to financially compensate the other creditor.[44]

5.3      Permissions

Traditionally, permissions have been treated as the opposite of prohibitions. For example, if P is forbidden to do A, this means (is the same as) that P is not permitted to do A. However, it has turned out that the relation between on the one hand permissions and on the other hand the other deontic notions, such as ‘ought’, ‘obligated’, ‘duty’, and ‘obligation’, is not straightforward (Hansson 2013). Characteristic in this connection is the distinction, popularized by Von Wright (1963, p. 85-87), between weak and strong permissions. A weak permission would be nothing else than the absence of a prohibition, while a strong permission would involve a prohibition to interfere with an agent’s freedom in a certain respect.

We will have a brief look at several possible interpretations of permission, and start with the possibility that act tokens, acts that have already been performed, were permitted. In this brief discussion, the agents performing actions will mostly be left out of consideration and the talk will be about action types or act tokens that are, or are not, permitted. The evaluation of act tokens may be undertaken from several points of view, such as the legal and the moral point of view. Here we confine ourselves to the legal point of view.

What does it mean that a particular act token was legally permitted? Basically it means that this act, as performed by this agent, does not belong to a legally prohibited action type. Either there was no legal norm prohibiting this type of action, or in the concrete case there was an exception to the norm. Suppose that Ellen takes a break by making a walk on the lawn. Taking a bread was allowed, but walking on the lawn was not. Therefore what Ellen did was not permitted, because her act can be subsumed under at least one prohibited action type. However, if Ellen would have received a special permission to walk on the lawn, het act was permitted, because the granted permission makes an exception to the general prohibition to walk on the lawn.

An action type is permitted if there is no norm which directly or indirectly prohibits that type of action. We will return to this distinction between direct and indirect prohibition soon, but first it is necessary to say something about default deontic status. Most of us live in a society where everything that has not been prohibited is permitted. Being permitted is the default deontic status of all action types, and it requires a prohibitive norm to change that status for a particular kind of action. Things might have been different, however. Logically it would have been possible that everything that has not been permitted is prohibited. The reader should keep this in mind and be prepared to turn the following account of prohibited and forbidden action types around for the theoretical case that a society would have prohibition as its default deontic status.

An action type is directly prohibited if and only if there is a norm that explicitly prohibits that type of action. So, if the norm exists that prohibits lying, the action type lying is directly prohibited. If this norm does not exist, it might still be possible that there is a norm that explicitly forbids cheating. Then cheating is directly prohibited. Suppose now, for the sake of argument, that lying involves cheating.[45] Then, barring exceptions, every instance of lying is also an instance of cheating. In that case the direct prohibition of cheating makes that lying, the action type, is indirectly forbidden by the norm that prohibits cheating. Using this distinction between directly and indirectly forbidden action types, we can say that an action type is permitted if this type is neither directly nor indirectly prohibited.

Thus far we discussed permissions in the sense of absence of prohibition. An act token was permitted if it could not be classified as belonging to a prohibited type. An action type is permitted if there is no norm that directly forbids this type of action and if the performance of an act of this type does not involve the performance of a prohibited action. However, as was already pointed out by Von Wight, some action types are explicitly permitted by a permissive norm. Such a permissive norm is typically – but not logically necessarily – connected to a freedom right. For example, the right to vote includes a permission to vote, and the freedom of expression includes a permission to utter one’s opinions.[46] The permission that is included in some rights should not be confused with other elements that are also included in these rights. The idea, for instance, that strong permissions include Hohfeldian immunities -  the legislator would for instance not have the power to forbid a citizen to vote - seems to be based upon this mistake. In a right, a permission may be combined with an immunity, but this combination does not mean that the permission somehow includes the immunity. It is the right that includes both the permission and the immunity.

It is possible that an act token can be classified as belonging to two types, one type being explicitly permitted and the other type being forbidden. This would, for instance, be the case if it is forbidden to set people against each other, while it is permitted to expression one’s political opinions. Then there is a norm conflict, which should be treated in the same way as other norm conflicts.

5.4      The anatomy of ought-to-do

To understand the nature of deontic facts and of norms, it is useful to distinguish the elements of deontic facts. We will enumerate these elements for states of affairs of the ought-to-do type, but most of the discussion can mutatis mutandis be applied to duties and obligations as well.[47]

Figure 3: The anatomy of ought-to-do













An ought-to-do state of affairs involves that somebody is either permitted, required or prohibited to do something, or to do something in a particular way, or at a particular time or place. An ought-to-do state of affairs consists of three or four elements, the deontic modality, the addressee, the act specification, and – occasionally – the specification of the act modality.

Take the following examples:

a.   It is forbidden to murder.

b.  Car drivers ought to carry a driver’s license.

c.   Leon is allowed to eat asparagus with his fingers.

In example a, everybody is an addressee; the modality is a prohibition, and the object of the deontic state of affairs is the performance of an action type (to murder).

In example b, the addressees are the members of the open class[48] of car drivers, the modality is an ought, and the object of this ought is the performance of an action type (carrying a driver’s license).

In example c, the addressee is a single agent, the modality is a permission, and the object is an action mode (using one’s fingers to eat asparagus).

It may be tempting to treat example b as expressing a conditional sentence: if somebody is a car driver, then he ought to carry a driver’s license. This temptation is even strengthened if one considers how the deontic fact described in b can be used to argue why a particular car driver, say Lenny, ought to carry a driver’s license. The following modus ponens style argument seems to do the job well: all car drivers ought to carry a driver’s license; Lenny is a car driver; therefore Lenny ought to carry a driver’s license. Still, it seems a better idea to follow Von Wright (1963, p. 82) by distinguishing conditions under which a deontic fact obtains, and the agents for which this deontic fact holds.

Notice, by the way, that this distinction presupposes that there are no conditional deontic facts, but only conditions for the existence of a deontic fact. If car drivers should place a warning sign before their cars if the car has broken down while it is dark, the deontic fact is that car drivers should put a warning sign before their cars, and this fact is present if the cars break down while it is dark. Of course the sentence in which this relation between darkness and the duty to place a warning sign is expressed is itself conditional. However, that does not make the deontic fact conditional.

Only deontic facts where the addressee is a single agent (or a closed group of agents) concern the existence of a duty or obligation, and only these can be immediate rule-based facts. If it is a fact that car drivers ought to carry a driver’s license, this is a fact only because every individual car driver, actual or merely hypothetical, ought to carry a driver’s license. These individual oughts are most likely based on the rule (norm) that car drivers ought to carry a driver’s license. A similar argument shows how it can be a fact that it is (for everybody) forbidden to murder. From this perspective it can be seen that the sentences a and b are ambiguous: they may be read as rule formulations, but also as descriptions of deontic facts. In the latter case they are the descriptive counterparts of the rules (see section 4.6).

6.      Of Norms and Other Rules

Given the facts that there is no fixed terminology concerning norms, and that norms are closely related to rules and to normativity, it seems worthwhile to explore the idea that norms are rules -  a kind of constraints – that lead to deontic facts. They lead in the first place to the existence of duties and obligations and - derived from these duties and obligations – to facts of the obligated- and ought-type. We will explore that idea in this section and to that purpose we start with distinguishing three kinds of rules, dynamic rules, fact-to-fact rules and counts-as rules. Then we consider the relevance of these kinds of rules for the constitution of deontic facts. This section is closed with  some remarks on competence-conferring rules and other rules that confer status.

6.1      Rules as individuals in the logical sense

Rule formulations such as ‘Thieves are liable to be punished’ and ‘The Mayor of a municipality is competent to issue emergency regulations for that municipality’  have, as far as their formulation is concerned, much in common with general statements. Moreover, rules can be used in rule-applying arguments which look like arguments of the modus ponens type. Nevertheless there are important differences between rules and statements. Otherwise than statements, rules exist in time: they can be created and repealed (derogated; abolished). They can become outdated, and can stop being used (desuetudo). Moreover, in contrast to statements which have the up direction of fit, they have the down direction of fit (of constraints). It is possible to predicate something of rules, such as in the sentence ‘This rule has been studied by legal historians for dozens of years’. Rules can also be part of a relation, as can be stated in the sentence ‘The rule that thieves are liable to be punished exists longer than the rule that gives Mayors the competence to create emergency regulations’.

Because of these latter reasons, there is much to be said for treating rules as a kind of things, rather than as statements describing what is the case. In the terminology of logic, such ‘things’ are called individuals. If rules are from a logical point of view individuals it is easy to see why rules as such cannot be parts of deductive logical derivations.[49] Deriving something from a rule would be comparable to deriving something from a chair. Norms do not figure in deductive arguments, but the reason is not that they are deontic or like imperatives (pace Jörgensen 1937/8), or that they have the down direction of fit of constraints[50], but that they are from the logical perspective individuals.

6.2      Dynamic rules

All rules connect facts to each other. These facts may be simultaneous, or they may succeed each other in time. The latter is the case with dynamic rules: they create new facts, or modify or take away existing facts as the consequence of the occurrence of an event.

Examples of events to which rules attach consequences are that John promised Richard to give him €100, and that Eloise was appointed as chair of the French Parliament. John’s promise has the consequence that from the moment of the promise on John has the moral obligation to pay Richard €100. The appointment of Eloise as chair has as its legal consequence that from the starting point of the chair’s new term on, Eloise will be the chair of the French Parliament. Other examples in which a dynamic rule attaches legal consequences to an event are that a Bill was passed, with as consequence the existence of new rules, or that Lionel committed theft, with as legal consequence that Lionel is liable to be punished.

Like all rules, dynamic rules have an element of generality. They apply to events of a particular kind and attach to these events facts of a particular kind. Dynamic rules may be conditional, in which case their consequence is only attached to the event under certain conditions. An example is the rule that if it is dark, the occurrence of a car accident obligates the drivers to place a sign on the road before to the cars.

If a juridical act or some other constitutive speech act is performed and a dynamic rule attaches consequences to this act, this is both a case of the constitutive down direction of fit and of the down direction of fit of constraints (see section 3.1). The former focuses on the speech act by means of which the consequences were constituted; the latter focuses on the rule that attaches consequences to the performance of the act. The constitutive force of speech acts rests on the effects brought about by dynamic rules and therefore the constitutive down direction of fit of constitutive acts is a special case of the more general down direction of fit of constraints and more in particular dynamic rules.

6.3      Fact-to-fact rules

Where dynamic rules govern the succession of facts in time, static rules govern the co-existence of facts. One kind of static rules are fact-to-fact rules, rules which make that one kind of fact tends to go together with some other kind of fact, where the latter fact depends (supervenes) on the former. The relation between the kinds of facts is timeless, in the negative sense that the one kind of fact is not the occurrence of an event after which the second kind of fact comes into existence.[51]

A logical example of a fact-to-fact rule is that a conjunction is true if both conjuncts are true.

A moral example is the rule that spouses should be faithful to one another.

Legal examples of fact-to-fact rules are the rules that

1.   the owner of a good is allowed to use this good;

2.   the mayor of a municipality has the competence to issue emergency regulations for that municipality;

3.   house owners must keep the pavement before their houses clean;

4.   the king of Belgium is the commander in chief of the Belgian army.[52]

Characteristically, all the legal example rules attach consequences to the possession of a certain legal status. Important legal examples of fact-to-fact rules are rules that impose legal duties (example 3), rules that confer competences on people with a particular status (example 2), and rules that attach a specific status to the presence of some other status (example 4).

6.4      Counts-as rules

The second kind of static rules that will be discussed here are counts-as rules. They have the structure: Individuals of type 1 count as individuals of type 2. These ‘individuals’ may be human beings, as in the rule that the parents of a minor count as the minor’s legal representatives. Often, however, the ‘individuals’ that count as another kind of individual are events. For instance, under particular circumstances, causing a car accident counts as committing a tort, or offering money to another person counts as attempting to bribe an official.

Usually counts-as rules are conditional, meaning that individuals of type 1 only count as individuals of type 2 if certain conditions are satisfied. An example from Dutch law (art. 3:84 of the Civil Code) would be the rule that the delivery of a good counts as the transfer of that good if the person who made the delivery was competent to transfer and if there was a valid title for the transfer.

Counts-as rules cannot create deontic facts by themselves. However, they often make that something counts as something to which a norm attaches deontic facts. Causing a car accident may count as a tort to which a rule of tort law attaches the obligation to pay damages. Being a person against whom serious objections exist counts as being a criminal suspect, a fact that gives police officers permission to arrest you.

6.5      Norms

To focus our discussion, we have defined norms as rules that constitute deontic facts. Let us focus our discussion even more, by assuming that there are only two kinds of basic deontic facts that matter for norms, that is duties, which are not directed towards a corresponding right holder, and obligations, which are directed in that way.

This stipulation contains two clauses that need justification. The first clause is that we confine ourselves here to basic deontic facts. That excludes facts of the types that somebody is obligated to do something, or that somebody ought to do something. These latter facts are not basic, because they supervene on the existence of duties and obligations (see section 5.2).

The second clause is that we confine ourselves to deontic facts that matter for norms. This has to do with the second-person perspective of norms (see section 2.3): norms justify claims for compliance. Mere requirements of practical rationality, such as the requirement that everybody who is thirsty should drink something, do not justify such claims. These requirements of practical rationality are not duties, let alone obligations, and the constraints underlying them, such as the constraint that somebody who is thirsty should drink, are typically not called norms. The practical relevance of this second point is mainly semantic, however.

We assume therefore that norms are rules that constitute duties or obligations. Examples of such norms are the norms that:

A.  Nobody should steal (Everybody has the duty not to steal).

B.  Car drivers must (have the duty to) drive on the right hand side of the road (perform acts in a particular way).

C.  Paul must (has an obligation towards Patty to) compensate the damage of Patty.

As was to be expected, these norms have formulations that are identical to their descriptive counterparts. From the formulations it is not possible to detect whether we are dealing with a norm, or a descriptive sentence. The two major differences between norms and their descriptive counterparts are that norms have the down direction of fit of constraints, while their descriptive counterparts have the up direction of fit, and that norms are from the logical point of view individuals, and not part of language, while their descriptive counterparts are sentences and therefore part of language.

The account that was given above of norms emphasized the constitutive nature of norms: they create, rather than are, duties and obligations. This may look somewhat strange at first sight, since the notion of a norm is more often associated with guidance of behavior than with the constitution of facts. Still this finding should not surprise if the idea is accepted that there can be facts that guide behavior. Rules that constitute such behavior guiding facts – that is: norms - are both constitutive and behavior guiding (regulative).

6.6      Competence- , power- and other status conferring rules

The emphasis that is often placed on norms may draw our attention away from the rules that do not primarily aim at guiding our behavior. Still these other rules are crucially important for the functioning of more complex normative systems. To illustrate this, we will briefly pay attention to some rules that are not norms.

Dynamic rules attach new facts as legal consequences to the occurrence of some event. By performing some act that triggers the operation of a dynamic rule, an agent can bring about these legal consequences. For example, by committing a crime, a person can make himself liable to be punished, and by moving to a different municipality a citizen can change the amount of municipality tax he has to pay. The existence of a dynamic rule has as side-effect that persons can do things which they could not have done without the presence of these rules. In that sense they confer powers upon agents who through their acts are able to trigger the operation of dynamic rules and in that way bring about legal consequences.

The two examples given above of powers resulting from the existence of dynamic rules are not the most characteristic ones for law. The powers that can be exercised by performing a juridical act are much more characteristic. In this connection a juridical act may be defined as an act that is aimed at bringing about legal consequences, to which these legal consequences are typically attached by a dynamic rule for the reason that this was the aim of the act.[53] Typical examples of juridical acts are contracting, making a last will or an association, legislating, pronouncing a judicial verdict, and granting a license.

An agent who performs a juridical act and thereby creates legal consequences must be competent to create these legal consequences by means of this kind of juridical act. This competence is assigned by a legal rule, although not necessarily an explicitly created one. For example, in order to be able to transfer the ownership of a piece of land, the transferor should be competent to alienate the land. This competence is a legal status that is typically attached (by a fact-to-fact rule) to the ownership of the land. Being competent is a necessary condition for the successful performance of a juridical act, and therefore also for the existence of a legal power that must be exercised through the performance of a juridical act. However, as we have seen from the two example of the previous paragraph, legal powers, that is powers that exist because of the existence of legal rules, do not always require juridical acts for the exercise.

It is sometimes possible that somebody who lacks the competence for a particular juridical act nevertheless can succeed in bringing about the legal consequences of this act. For example, a public officer may succeed in providing somebody with a valid license, even though the officer lacked the competence to do so. Another example is that a non-owner may succeed in making somebody else the owner of a good. Both examples illustrate that legal consequences that could not be brought about through a valid juridical act may nevertheless occur for reasons of legal certainty. These examples illustrate than an agent can have the power to bring about legal consequences, while lacking the competence to do so by means of a valid juridical act.

Rules that confer an agent the competence to perform a particular kind of juridical acts can both be dynamic and fact-to-fact rules. An example of the former is the dynamic rule that governs contracts and that by and large involves that the legal consequences which the contract partners intended to bring about will actually hold. If one contract partner wanted to make the other party competent to perform juridical acts in his name, that is to act as a legal representative, the effect of the contract is that this latter party has become competent to perform juridical acts in name of the former contract partner. An example in which a competence is provided by a fact-to-fact rule is that the owner of a good is competent to alienate this good. The rule attaches the competence to the status of ownership.

Having a particular competence is a status assigned by a legal rule. There are many other examples of status assigned by legal rules. All legal counts-as rules assign a status to entities or events, such as the status of being a vehicle in the sense of the Road Act, being the president of Germany, being the commander-in-chief of the Belgian army, being a suspect in the sense of penal law, being the owner of a good, being wedded.[54] Very often norms attach deontic consequences to the presence of such a legal status.

7.      Summary

In this contribution an attempt was made to clarify the notion of a norm by elaborating the idea that norms are rules that lead to deontic consequences. The elaboration focused both on the nature of rules and on the nature of deontic facts.

Rules, it was argued, are a kind of constraints on possible worlds. They determine which kinds of facts necessarily go together or cannot go together. Three kinds of rules were distinguished: dynamic rules which attach consequences to the occurrence of events; fact-to-fact rules which attach one fact to the presence of some other fact, and counts-as rules, which make that some things (often evens) also count as something else. It was pointed out that the existence of a rule makes that some facts obtain: the factual counterparts of the rules. In this sense all rules are constitutive. The descriptive sentences that express these facts, the descriptive counterparts of rules, are open generalizations and they have often  the same formulations as the rules from which they derive their truth.

By distinguishing between objective, brute social, and rule-based facts, an attempt was made to overcome resistance against the idea that facts might be normative, that there might be deontic facts. That something is mind-dependent does not exclude that it is a fact. Deontic facts are mind-dependent, because they are facts that tend – often in an indirect way - to induce a motivation to comply in agents to which they apply. Deontic facts are most often the result of the application of fact-to-fact rules (duties) or dynamic rules (obligations). A distinction was made between two kinds of basic deontic facts - the existence of duties and of obligations – and two kinds of supervening deontic facts: being obligated and owing to do something.



Alchourrón and Bulygin 1981

Carlos E. Alchourrón and Eugenio Bulygin, ‘The Expressive Conception of Norms‘, in R. Hilpinen (ed.), New Studies in Deontic Logic, Dordrecht: Reidel 1981.

Alvarez 2010

Maria Alvarez, Kinds of Reasons. An Essay in the Philosophy of Action, Oxford: Oxford University Press 2010.

Anscombe 1976

G.E.M. Anscombe, Intention, 2nd ed., Oxford Basil Blackwell 1976.

Araszkiewicz and Pleszka 2015

Michal Araszkiewicz and Krzysztof Pleszka, ‘The Concept of Normative Consequence and Legislative Discourse’, in Michal Araszkiewicz and Krzysztof Pleszka (eds.), Logic in the Theory and Practice of Lawmaking, Cham: Springer 2015, 253-297.

Austin 1954

John Austin, The Province of Jurisprudence Determined, Original version published in 1832. Republished in 1954 together with The Uses of the Study of Jurisprudence and a Preface by H.L.A. Hart, by Weidenfeld and Nicholson.

Bertea 2009

Stefano Bertea, The Normative Claim of Law, Oxford: Hart 2009.

Broome 2013

John Broome, Rationality through Reasoning, Chicester: Wiley Blackwell 2013.

Castañeda 1972

Hector-Neri Castañeda, ‘On the Semantics of Ought-To-Do’, in Donald Davidson and Gilbert Harman (eds.), Semantics of Natural Language, 2nd ed., Dordrecht: D. Reidel Publishing Company.

1972, p. 675-694.

Darwall 2006

Stephen Darwall,  The Second-Person Standpoint. Morality, Respect and Accountability. Cambridge: Harvard University Press 1996.

Frändberg typescript

Åke Frändberg, The Legal Order. Studies in the foundations of Juridical Thinking, typescript of a book in preparation.

Hage 1997

Jaap C. Hage, Reasoning with Rules, Dordrecht: Kluwer 1997.

Hage 2001

J.C. Hage, ‘Contrary to Duty Obligations. A Study in Legal Ontology’, in Bart Verheij, Arno R. Lodder, Ronald P. Loui and Antoinette J. Muntjewerff (eds.), Legal Knowledge and Information Systems. Jurix 2001: The Fourteenth Annual Conference. Amsterdam: IOS Press, 2001, 89-102

Hage 2005

Jaap C. Hage, Studies in Legal Logic, Dordrecht: Springer 2005.

Hage 2013

Jaap Hage, 'The deontic furniture of the world', in Jerzy Stelmach, Bartosz Brożek and Mateusz Hohol (eds.), The Many Faces of Normativity, Kraków: Copernicus Press 2013, p. 73-114.

Hage 2015

Jaap Hage, ‘Separating Rules from Normativity’, in Michal Araszkiewicz, Pawel Banaś, Tomasz Gizbert-Studnicki and Krzystof Pleszka (eds.), Problems of Normativity, Rules and Rule-Following, Cham: Springer 2015, p. 13-30.

Hansen 2013

Jörg Hansen, ‘Imperative Logic and Its Problems’, in Dov Gabbay, John Horty, Xavier Parent, Ron van der Meyden and Leendert van der Torre (eds.), Handbook of Deontic Logic and Normative Systems, College Publications 2013, 137-191.

Hansson 2013

Sven Ove Hansson, ‘The Varieties of Permission’, in Dov Gabbay, John Horty, Xavier Parent, Ron van der Meyden and Leendert van der Torre (eds.), Handbook of Deontic Logic and Normative Systems, College Publications 2013, 195-240.

Hart 2012

Herbert L.A. Hart, The Concept of Law, 3rd ed. Oxford: Oxford University Press 2012, 1st ed. 1961.

Hartmann 1962

Nicolaï Hartmann, Ethik, 4th edition, Berlin: Gruyter 1962.

Herrestad and Krogh 1995

Henning Herrestad and Christen Krogh, ‘Obligations directed from bearers to counterparties’

Proceedings of the 5th International Conference on Artificial Intelligence and Law (ICAIL’95), New

York: ACM, 1995, 210-218.

Hilpinen and McNamara 2013

Risto Hilpinen and Paul McNamara, ‘Deontic Logic; A Historical Survey and introduction’, in Dov Gabbay, John Horty, Xavier Parent, Ron van der Meyden and Leendert van der Torre (eds.), Handbook of Deontic Logic and Normative Systems, College Publications 2013, 3-136.

Jörgensen 1937/8

J. Jörgensen, ‘Imperatives and Logic’, in Erkenntnis 7 (1937/8), 288-296.

Kelsen 1945

Hans Kelsen, General Theory of Law and State. Cambridge: Harvard University Press 1945.

Kelsen 1960

Hans Kelsen, Reine Rechtslehre, 2nd ed., Wien: Franz Deuticke 1960.

Kelsen 1979

Hans Kelsen, Allgemeine Theorie der Normen, edited by Kurt Ringhofer and Robert Walter, Wien: Manzsche Verlags- und Universitatsbuchhandlung 1979.

Loux 1979

Michael J. Loux (ed.), The possible and the actual. Readings in the metaphysics of modality, Ithaca: Cornell University Press 1979.

MacCormick 1978

Neil MacCormick, Legal Reasoning and Legal Theory, Oxford: University Press 1978.

MacCormick and Weinberger 1986

N. MacCormick and O. Weinberger, An Institutional Theory of Law, Dordrecht: Reidel 1986.

Menzel 2015

Christopher Menzel, ‘Possible Worlds’, The Stanford Encyclopedia of Philosophy (Summer 2015 Edition), Edward N. Zalta (ed.), URL = <http://plato.stanford.edu/archives/sum2015/entries/possible-worlds/>.

Navarro and Rodríguez 2014

Pablo E. Navarro and Jorge L. Rodríguez, Deontic Logic and Legal Systems, Cambridge: University Press 2014.

Navarro and Rodríguez 2015

Pablo E. Navarro and Jorge L. Rodríguez, ‘Entailed Norms and the Systematization of Law’ in Michal Araszkiewicz and Krzysztof Pleszka (eds.), Logic in the Theory and Practice of Lawmaking, Cham: Springer 2015, 97-114.

Raz 1975

Joseph Raz, Practical Reason and Norms, London: Hutchinson 1975.

Sartor 2005

G. Sartor, Legal Reasoning, A Cognitive approach to the Law,  Dordrecht: Springer 2005.

Schauer 1991

F. Schauer,  Playing by the Rules, Oxford: Clarendon Press 1991.

Scheler 1954

Max Scheler, Der Formalismus in der Ethik und die materiale Wertethik: Neuer Versuch der Grundlegung eines ethischen Personalismus, 4th edition, Bern: A. Frankcke 1954.

Smits 2014

Jan M. Smits, Contract Law. A comparative introduction, Cheltenham: Edward Elgar 2014.

Toulmin 1953

Stephen Toulmin, The Philosophy of Science. An Introduction, New York: Harper Row 1953.

Ullmann-Margalit 1977

Edna Ullmann-Margalit, The Emergence of Norms, Oxford: Clarendon Press 1977.

Von Wright 1963

Georg Henrik von Wright, Norm and Action. A logical enquiry, London: Routledge and Kegan Paul 1963.

Williams 1981

Bernard Williams, ‘Internal and External Reasons’ in Bernard Williams, Moral Luck, Cambridge: Cambridge University Press 1981, 101-113.

Zimmerman 1996

Reinhard Zimmerman, The Law of Obligations. Roman Foundations of the Civilian Tradition, Oxford: Oxford University Press 1996.



[1]     For now, the term ‘norm’ will be used in a broad sense. The more specific use will be introduced in section 6.

[2]     It is also possible that norms guide behavior of non-human entities, such as computer programs and robots, but here we will not pay special attention to these possibilities.

[3]     The distinction is already quite old. Von Wright (1963, p. 14) mentions in a footnote the work of Max Scheler (1916) and Nicolaï Hartmann (1962). In the Handbook of Deontic Logic and Normative Systems, Hilpinen and McNamara (Hilpinen and McNamara 2013, p. 97) refer to Castañeda 1972.

[4]     It should be noted that, in particular in law, formulations that suggest an ought-to-be norm because they do not specify that something ought to be done, can nevertheless stand for ought-to-do norms, because it is clear from the context who is responsible for bringing about the right state of affairs.

[5]     Attempts to define ought-to-do norms in terms of states of affairs that ought to be the case (see Hilpinen and McNamara 2013, p. 97-112 for an overview) are in the eyes of the present author a major source of problems in formal deontic logic. See Hage 2001.

[6]     Reason-terminology has become dominant in ethical theory. See for instance Williams 1981, Alvarez 2010, and Broome 2013. In legal theory it still lacks the popularity it deserves, despite the efforts of Raz (1975),  Hage (1997 and 2005), and Bertea (2009) to promote it.

[7]     This is not the place to discuss the different functions of reasons for acting, and the distinction between guiding and explanatory reasons. However, it is worthwhile to point out that even if guiding reasons do not need to exert a causal influence on the person to whom they apply, the very notion of a guiding reason would not make sense if people in general would not be motivated by the awareness that a guiding reason applied to them. See also section 5.1.

      Legal philosophers will recognize the parallel with the relation between a legal system’s efficacy and the validity of the rules that belong to the system. Validity cannot be derived from efficacy, but it makes little sense to speak of the validity of norms that belong to a system that is completely inefficacious. Cf. Kelsen 1945, p. 42.

[8]     This contribution adheres to the convention that authors should use the pronouns for their own gender to refer to persons whose gender is not important for the argument or otherwise determined by the text.

[9]     From here on we will follow the custom amongst logicians and  use the terms ‘deontic’ and ‘non-deontic’ for the distinction between normative and non-normative.

[10]    The importance that is attached to the distinction between is and ought may be explained from the function of critical morality, that is to is to evaluate existing moral practices critically. The social practice of critically moralizing can only exist if the fact that some norms are actually used does not count as sufficient evidence for the claim that these norms should be used. Ignoring the difference between the norms people actually use and the norms people should use makes critically moralizing impossible. So there is a practical relevance to distinguishing between is and ought, but this relevance does not justify that the real difference is blown up to an ontological gap.

[11]    It is possible to construct these rules as elements of more complicated rules that do guide behavior, and that is why it was written that it is not their primary function to guide behavior. However, it is difficult to disagree with Hart (2012, 35-42), who wrote that the construction of such rules as parts of mandatory rules would be a distortion. Still there is a sense in which, for example, power conferring rules can be followed, and in section 4.2 an example will be discussed.

[12]    Seemingly the causal down direction of fit can also exist between duties and behavior, and not merely between speech acts and behavior. However, that would mean that non-material entities such as duties can exert causal influences, which does not sit well with our ideas about the nature of causation. It is therefore more coherent (with our views of causality) to say that the causal down direction of fit can exist between the belief that one has a duty, as realized by a brain state, and behavior.

[13]    It may be disputed whether the fact that bachelor are unmarried depends on a conceptual rule, and whether this conceptual rule does not depend itself on some ontological constraint (ontological nominalism or ontological realism). For the present purposes this does not matter, however.

[14]    For ease of exposition, we will ignore here exceptional descriptive sentences, such as the sentences ‘The king of France is bald’ and ‘This sentence is false’. The clause ‘non-modal’ was added to take into account that modal sentences which express necessity may be interpreted as dealing with more than one possible world.

[15]    The metaphysics of possible worlds is the central topic of an anthology edited by Loux (Loux 1979). To the present authors knowledge, however, the idea that possible worlds are relativized to sets of constraints is not treated in that ontology, nor in more recent overviews of the discussions about possible worlds, such as Menzel 2015.

[16]    The theory about constraints as exposed here has some remarkable similarities to the theory of modalities defended by Frändberg (typescript).

[17]    These constraints on logically possible worlds are typically represented in the semantics of logical theories by characteristics of the valuation function that assigns truth values to propositions. See for instance Navarro and Rodríguez 2014, p. 16.

[18]    As a matter of fact, the concept of a ‘planet’ has recently been redefined, taking away the status of a planet from the former planet Pluto. See https://en.wikipedia.org/wiki/IAU_definition_of_planet (last visited on December 24, 2015).

[19]    This relation between conventions and the necessity based on them is explored a bit more in Hage 2013.

[20]    Not only rules can have exceptions. There can also be exceptions to logical constraints (some descriptive sentences are not true or false), and to physical constraints (some physical laws are not applicable in extreme circumstances).

[21]    This difference should not be overestimated, however. The geometrical law that the three corners of a triangle add up to 180 degrees only holds for relatively small triangles and (which may be the same issue) for triangles in a flat plane. See also the discussion of the scope of physical laws in Toulmin 1953, p. 69 and 78.

[22]    This theme is elaborated in Hage 2015.

[23]    The facts based on rules are perhaps better known as ‘institutional facts’. Cf. MacCormick and Weinberger 1986, p.10.

[24]    In section 4.6 these rule-based facts will be called ‘immediate rule-based facts’, and they will be distinguished from ‘mediated rule-based facts’.

[25]    An important difference between these standards and, for instance, legal rules is that legal rules also generate exclusionary reasons (Raz 1975, Schauer 1991, Hage 1997), while evaluative standards typically do not. This difference has no fundamental consequences for the role of evaluative standards as underlying rule-based facts, however.

[26]    Famously, Austin defined laws in his first lecture in The Province of Justice Determined (Austin 1954, p. 24) as commands which oblige persons generally to acts or forbearances of a class.

[27]    Von Wright makes this mistake for a particular category of norms, the laws of the state. He calls such norms ‘prescriptions’, and defines prescriptions as having their source in the will of a norm-authority. See Von Wright 1963, p. 7. A similar mistake seems to be made by Alchourrón and Bulygin (1981), when they recognize an ‘expressive conception of norms’, according to which norms are essentially commands.

[28]    This mistake was made by Kelsen when he allowed the possibility of derogating norms. See Kelsen 1960, p. 57 and Kelsen 1979, p. 1.

[29]    Remember that this necessity is compatible with exceptions to rules. See section 3.4.

[30]    See also section 3.4.Because of the way open generalizations are often represented in formal logic, they have also become known under the misnomer ‘defeasible conditionals’. Generalizations are not conditional sentences, even though they tend to be represented in predicate logic by means of conditionals. Moreover, open generalizations are true or false and not defeasible - but conclusions based upon them may be defeasible; see Hage 2005, p. 14. However, the truth conditions of open generalizations differ from those of closed generalizations, because the former are not necessarily falsified by counter-examples, while the latter are.

[31]    If the word ‘rule’ is also used to denote open generalizations, there is no problem in deriving rules in this sense from other rules,  also in the sense of open generalizations, and facts. It is this kind of reasoning about ‘rules’ that seems to be at stake when MacCormick (1978, p. 100-108) writes about second-order justification of rules.

[32]    The difference between duties and obligations as it is made here will be discussed in section 5.2.

      Sometimes the duties and obligations themselves, or their contents, are called ‘norms’. As explained in section 1, we adopted a different terminology here.

[33]    The following paragraphs only discuss duties, but mutatis mutandis the argument also applies to obligations.

[34]    Sartor (2005, p. 454) seems to express the same idea when he characterizes obligations in terms of the intention to act on them.

[35]    Human agents often critically evaluate the ‘duties’ that they have according to a particular normative system, such as positive morality or positive law, and sometimes this evaluation leads to the conclusion that they should not comply with some ‘duty’. However, such a refusal to comply with a ‘duty’ which is not up to standard is often motivated by saying that the ‘duty’ turned out not to be a ‘real’ duty after all. In that case, the link between real duties and the motivation to act upon them remains intact. Obviously, much more can and needs to be said on this issue, but this is not the place to do so. Interested readers are referred to Hage 2013.

[36]    Such a postulation seems to be made in the account that Navarro and Rodríguez give of the relation between norms and normative propositions. Cf. Navarro and Rodríguez 2014, p. 78.

      The constitutive nature of the rules on which deontic facts are based is discussed in section 6.5 of the present contribution.

[37]    The term ‘obligation’ derives the technical meaning that is proposed here from the civil law tradition, according to which an obligation is a particular kind of bond between a debtor and a creditor. (For the historical roots of this word use, see Zimmermann 1996, p. 1) In the English literature the difference between duties and obligations is not drawn sharply, possibly under the influence of the common law.

[38]    Being a human being might be the most abstract status to which duties are assigned.

[39]    For a logical discussion of these ‘directed obligations’, see Herrestad and Krogh 1995.

[40]    An example of a duty without a person towards whom the duty exist is the duty to stop for a traffic light, even if nobody is approaching. However, even if a duty mentions persons, e.g. the duty not to kill prisoners of war, this is not a duty towards these persons. Other persons can also address the duty holder about compliance with the duty. This is different for obligations, where typically only the right holders can demand compliance.

[41]    The notion ‘prima facie ought’ is more fashionable, but is strictly speaking an epistemic notion: if A prima facie ought to do X, then for all we know A ought to do X.

[42]    Although the principle ‘ought implies can’ is in the eyes of the author not a logical constraint, there is from the moral and the legal point of view much to be said for it. In the law of obligations, for instance, impossibility is the main reason for assuming force majeure. That is why the principle is applied in the present argument.

[43]    Whether Antony is also permitted not to deliver the car to anyone of his creditors depends on the legal system. In the common law, where ‘specific performance’ is exception rather than the rule, Antony would be permitted to financially compensate both creditors rather than delivering the car to any one of them. In the civil law tradition, Antony would still be obligated to deliver the car to the creditor he does not compensate financially. (See Smits 2014, p. 194-202.) This example illustrates in the first place that the relation between the existence of an obligation and what a debtor legally ought to do depends on the law, not on logic alone, and in the second place that it is useful to study the law from a comparative perspective to see the respective roles of law and logic. Where legal solutions differ, they cannot be a matter of logic.

[44]    This account may not be correct for every legal system. In some systems obligations to transfer a good do have a priority, with the older obligation superseding the more recent one. In those systems the debtor legally ought to transfer the object to the oldest creditor. Also this example illustrates that the relation between legal obligations and what an agent legally ought to do are in the first place governed by law.

[45]    The precise nature of this involves-relation which may hold between action types is crucially important in this connection. A first approximation would be that action type A1 involves action type A2, if necessarily every token of A1 is also a token of A2 (Hage 2001). It should be noted in this connection that the approximation presupposes that one act token can belong to more than one action type, and that the constraints that determine what counts as necessary remain unspecified.

[46]    The inclusion of permissions -  and also of competences – in rights is somewhat analogous to the involvement of one action type by another action type. An adequate theory about the nature of rights should include an elaboration of this includes-relation, but this is not the place to address this topic.

[47]    The main difference is that whereas an ought and a duty contain three (or four) elements, an obligation is directed toward a creditor and therefore contains four (or five) elements, the extra element denoting the creditor.

[48]    That the class is ‘open’ means that the denoting expression refers to everybody who may happen to be a car driver, and not merely to the fixed set of actual car drivers. The ‘openness’ of the class makes that the deontic fact also deals with hypothetical car drivers, as in ‘If Thera would have been a car driver, she would have to carry a driver’s license’. See also the discussion of open generalizations in section 4.6.

[49]    This does not exclude that rule applying arguments are studied in logic. There are several ways to do so. One is to drop the demand that all elements of an argument are propositions. A second is to allow entities without truth values as propositions. And a third one is to use statements about the existence of rules as premises in rule-applying arguments. All of these options have consequences for the systems of logic that can be used to study rule-applying arguments that reach farther than accommodation for the defeasibility of rule-applying arguments.

[50]    That the down direction of fit of constraints does not preclude them from being parts of arguments becomes clear from the example that if the world is constrained in such a way that Volkswagens are vehicles and that vehicles are not allowed in the park, the world is also constrained in the sense that Volkswagens are not allowed in the park. However, from the fact that the former two constraints exist as rules, it cannot be derived that the latter constrains also exists as a rule. See also section 4.8.

[51]    Notice that this timeless relation between the conditions and the consequences of a fact-to-fact rule is compatible with the existence in time of the rule. Only as long as the rule exists, the condition-facts and the conclusion-facts go together in a timeless fashion.

[52]    This last rule may also be interpreted as a counts-as rule.

[53]    The formulation ‘aimed at’ has been chosen instead of the more natural sounding ‘performed with the intention to’ to include acts that are performed without a conscious intention, such as acts performed by implemented computer programs. A public officer who signs a license without even reading it also performs a juridical act, and it should not be precluded by definition that a computer program that buys and sells securities thereby performs juridical acts. For this reason, the aim of an act should not be identified with the intention with which the act was performed. Aims are ascribed to acts, and (ascribed) intentions are merely factors that play a role in ascribing aims. Thanks go to Hester van der Kaaij for pointing out to me how important this innocuous seeming difference between intention and aim is.

[54]    In a sense even having a duty or an obligation can be seen as having a particular status, but this stretches the idea of legal status to its limits.