1. The word ‘norm’ in English and the corresponding word in other languages is used in many senses and often with an unclear meaning. It can hardly be said to be a well-established term in the English philosophic vocabulary. This can be said however of the adjective ‘normative’.

‘Norm’ has several partial synonyms which are good English. ‘Pattern’ ‘standard’ ‘type’ are such words. So are ‘regulation’ ‘rule’ and ‘law’. Directions of use and prescriptions are perhaps not often called ‘norms’ but we should not hesitate to call them ‘normative’.

Since the field of meaning of ‘norm’ is not only heterogeneous but also has vague boundaries it would probably be futile to try to create a General Theory of Norms covering the whole field. The theory of norms must be somehow restricted in its scope.

When constructing a restricted theory of norms however it is as well to remember that the various meanings of ‘norm’ are not logically unrelated. The word is not ‘ambiguous’ in the ordinary sense. A restricted theory of norms runs the risk of being defective if it does not pay due attention to conceptual affinities and logical relationships between the various parts of the whole field of meaning.

In this chapter I shall try to single out and briefly characterize some of the chief meanings of the word ‘norm’ or as we could also say species or types of norms.

2. We have said that one of the meanings of ‘norm’ is law. The word ‘law’ however is used in at least three typically different senses. First we speak of the laws of the state. Secondly we speak of the laws of nature. Thirdly we speak of laws of logic (and mathematics).

Obviously the laws of nature and the laws of the state are very different. Yet the identity of name is no pure coincidence.

Thus with the Greeks the conception of the world as a kosmos or harmonious order seems to have been connected historically with their conception of the city-state as a just and lawful order for a human community. The natural philosophy of the pre-Socratics has been called a projection of ideals of a social order on to the entire universe. In the philosophy of Plato we could say this idea of the world as a kosmos is projected back on to human conditions and made a pattern or standard of the good life.

With the Greek conception of law as the conditions of equilibrium and harmony may be contrasted the Hebrew (Old Testament) conception of it as the expression of a commanding sovereign will. The idea of God as lawgiver may be regarded as an analogy or a projection on to a supernatural plane of the idea of a sovereign chief or king in a human community. As the king gives laws to those over whom he is set to rule so in a similar manner God rules the whole universe by His law or ‘word’. The Christian idea of a king ‘by the grace of God’ is a projection back on to human affairs of this idea of a supreme lord of the universe. The idea of the worldly kingdom is given a foundation in the same supernatural idea for which it originally set the pattern.

As we tend to see it the laws of nature and the laws of the state are toto coelo logically different in spite of affinities in the origins of the ideas of the two ‘laws’. The difference can be briefly characterized as follows:

The laws of nature are descriptive. They describe the regularities which man thinks he has discovered in the course of nature. They are true or false. Nature does not except metaphorically ‘obey’ its laws. If a discrepancy is found to exist between the description and the actual course of nature it is the description and not the course of nature that must be corrected.—This is a superficial characterization of what the laws of nature are. But I think it is basically correct.

The laws of the state are prescriptive. They lay down regulations for the conduct and intercourse of men. They have no truth-value. Their aim is to influence behaviour. When men disobey the laws the authority behind the laws tries in the first place to correct the behaviour of men. Sometimes however the authority alters the laws—perhaps in order to make them conform more to the capacities and demands of ‘human nature’.

The contrast ‘prescriptive/descriptive’ can be used for distinguishing norms from things which are not norms. The laws nature are descriptive not prescriptive—and therefore they are not norms. That is: we thus delineate the use of the word ‘norm’; we draw the boundaries of the concept. Under another use of the term the laws of nature can perfectly well be called ‘norms’.

Someone may think that the attribute ‘prescriptive’ gives the clue to a general characterization of norms. Normative discourse is prescriptive discourse it is often said. With prescriptive discourse is then contrasted descriptive and sometimes also evaluative discourse.

To identify the meaning of ‘normative’ with that of ‘prescriptive’ and ‘norm’ with ‘prescription’ would however be too narrowing. Besides ‘prescriptive’ and ‘prescription’ are words with a vague meaning and must be made more precise in order to be useful. As we shall soon see there are things which we may without hesitation wish to call norms but to which the attributes ‘prescriptive’ and ‘descriptive’ both appear equally inappropriate.

3. Let us briefly consider the meaning of ‘law’ in the phrase ‘laws of logic (mathematics)’. The laws of logic were often in the past also called the Laws of Thought.

On closer inspection we find that there are in logic and mathematics several types of proposition which are or may be called ‘laws’. We need not here inquire into these distinctions. As examples of laws of logic we shall instance the Law of Excluded Middle in the formulation ‘Every proposition is either true or false’ and the Law of Contradiction in the formulation ‘No proposition is both true and false’.

Are such laws ‘descriptive’ or ‘prescriptive’? If the first what do they describe? The way people think? This suggestion is not very satisfactory. For first of all it is not clear in itself what it weans to think according to the law for example that no proposition is both true and false. Secondly the idea that the laws of logic describe how people think seems difficult to reconcile with the notion that these laws are a priori and thus true independently of experience—including experience of how people think.

The a priori nature of the laws of logic seems easier to reconcile with a view of them as prescriptive laws. Shall we then say that the laws of logic prescribe how we ought to think and how we may and must not think? Perhaps we can say this but it is also obvious on reflection that the sense in which the laws of logic ‘prescribe’ (order permit prohibit) is a different sense from that in which the laws of the state prescribe.

Here the idea suggests itself that the laws of logic and mathematics prescribe how one ought to think and calculate in order to think and calculate correctly. The laws of logic do not aim at making people think correctly as the laws of the state can be said to aim at making people behave in a certain way. The laws of logic provide a standard whereby to judge whether people think correctly or not. This seems to be a good way of characterizing the difference between the two types of law and the different senses in which they ‘prescribe’.

Yet to say that the laws of logic prescribe how people have to think in order to think correctly is a challenging and dangerous way of talking. It suggests that the ‘prescriptive’ function of the laws of logic is secondary to a ‘descriptive’ function of them as stating principles of correct thinking. Primarily the laws of logic and mathematics state truths about the logical and mathematical entities—propositions relations inferences numbers etc. This they also do overtly when formulated in the usual way as e.g. when we say ‘Every proposition is either true or false.’

Thus the view of the laws of logic as prescriptive of the way people ought to think leads to a view of these laws as being primarily descriptive. What on this new view the laws of logic describe is not however how people think but how the logical entities are constituted.

This view of logic (and mathematics) is connected with great difficulties. It seems to presuppose a peculiar ‘ontology’ of the logical (mathematical) entities. This ontology is sometimes called Platonism or Realism in the philosophy of logic (mathematics). On this view the laws of logic (mathematics) are at the same time very much like and yet significantly different from the laws of nature. Both types of law have a truth-value. But laws of the first type are necessarily true; laws of the second type contingently so. Both types of law describe the properties and relations of some entities. But the entities with which laws of the first type deal are eternal and imperishable whereas the entities with which the laws of the second type deal are mutable and contingently existing. This is a superficial characterization. But I think it catches hold of something typical.

The main alternative to a realistic (Platonistic) position in the philosophy of logic (mathematics) is sometimes called a nominalist or conventionalist position. It has many variants. Some of them seem just as implausible and difficult to defend as some radically Platonistic view. I shall here refrain from giving even a superficial characterization of the conventionalist position as such. I shall only hint at the status which the laws of logic (mathematics) will acquire if we reject a Platonistic philosophy.

We could then compare these laws to the rules of a game. Playing a game is an activity and so is thinking and calculating. The rules of say chess determine which moves are permitted and which not and sometimes require a certain move to be made. In a similar sense it may be suggested the rules of logic determine which inferences and affirmations are ‘possible’ (correct legitimate permitted) in thinking. Of a person who does not play in accordance with the rules of chess we would say either that he plays incorrectly or that he does not play chess. We would say the first e.g. if he wanted to follow the rules but did not know or understand what they demanded of him. Or we would say it if he is trying to cheat his opponent. We would say the second e.g. if he did not care about following the rules or consciously and consistently played according to different rules. In a similar sense the suggestion runs we say of a person who does not infer according to the rules of logic either that he infers incorrectly or that he does not ‘infer’ at all. And we say the one or the other on roughly the same grounds as those which determine our reactions to the player.

The ‘Platonist’ would argue that the above analogy breaks down at this point: Whereas the man who plays against the rules of a game sins only against the rules the man who thinks against the rules of logic is in conflict with truth. The rules of a game are man-made and can be altered by convention or at will. The standards of truth are not conventional. That there is some truth in this argument is obvious. What this truth is and what implications it has for the analogy between the laws of logic and the rules of a game is however not obvious.

We raised the question whether the laws of logic and mathematics are descriptive or prescriptive. We have found that neither characterization appears quite to the point. These laws may be called descriptive but not in the same clear sense in which the laws of nature are descriptive. They may also be called prescriptive but in a rather different sense from that in which the laws of the state are prescriptive. The comparison of the laws of logic (mathematics) to the rules of a game suggested a new characterization of these laws. According to this new characterization the laws of logic (mathematics) neither describe nor prescribe but determine something. Irrespective of what we think of the comparison in other respects we can agree to the usefulness of this characterization. It suits the laws of logic (mathematics) better than either the attribute ‘descriptive’ or the attribute ‘prescriptive’.

4. The rules of a game are the prototype and standard example of a main type of norm. We shall here reserve the name rule as a technical term for this type.

Playing a game is a human activity. It is performed according to standardized patterns which can be called moves in the game. The rules of the game determine as I shall say these moves or patterns—and thereby also the game ‘itself’ and the activity of playing it. We could say that when viewed from the point of view of the game itself the rules determine which are the correct moves and when viewed from the point of view of the activity of playing the rules determine which are the permitted moves. It is understood that moves which are not correct are prohibited to players of the game and that a move which is the only correct move in a certain situation in the game is obligatory when one is playing the game.

The rules of grammar (morphology and syntax) of a natural language are another example of the same main type of norm as the rules of a game. To the moves of a game as patterns correspond the set forms of correct speech. To play or the activity of playing a game corresponds speech or the activity of speaking (and writing) a language. Of a person who does not speak according to the rules of grammar we say either that he speaks incorrectly or that he does not speak that language. The grounds for saying the one or the other are very much the same as the grounds for saying of a person either that he plays a game incorrectly or does not play it at all. But the rules of grammar have a much greater flexibility and mutability than the rules of a game. They are in a constant process of growth. What the rules are at any given moment in the history of a language may not be possible to tell with absolute completeness and precision.

The rules of a logical and mathematical calculus are in some respects even more like the rules of a game (such as e.g. chess) than are the rules of grammar of a natural language. (Games and calculi have a much poorer ‘history’ than natural languages.) In at least one important respect however the rules of a calculus are more like rules of grammar than like rules of a game. Calculating like speaking a language is a play with symbols. Calculi and languages have a semantic dimension which games on the whole lack.

5. A second main type of norms beside rules I shall call prescriptions or regulations. We have already met with one sub-type of such norms viz. the laws of the state.

I shall regard the following features as characteristic of norms which are prescriptions:

Prescriptions are given or issued by someone. They ‘flow’ from or have their ‘source’ in the will of a norm-giver or as we shall also say a norm-authority. They are moreover addressed or directed to some agent or agents whom we shall call norm—subject(s). The authority of a norm can normally be said to want the subject(s) to adopt a certain conduct. The giving of the norm can then be said to manifest the authority's will to make the subject(s) behave in a certain way. In order to make its will known to the subject(s) the authority promulgates the norm. In order to make its will effective the authority attaches a sanction or threat of punishment to the norm. In all these respects the norms which we call prescriptions differ characteristically from the norms which we call rules.

Generally speaking prescriptions are commands or permissions given by someone in a position of authority to someone in a position of subject. Military commands are an example of prescriptions. So are the orders and permissions given by parents to children. Traffic-rules and other regulations issued by a magistrate largely have this character too. The decisions of a law-court may be said to have a prescriptive aspect or component.

6. A group of norms which are in some respects like rules and in other respects like prescriptions are customs.

Customs may be regarded as a species of habits. A habit is primarily a regularity in an individual's behaviour a disposition or tendency to do similar things on similar occasions or in recurrent circumstances. Habits are acquired and not innate. Customs may be regarded as social habits. They are patterns of behaviour for the members of a community. They are acquired by the community in the course of its history and imposed on its members rather than acquired by them individually.

Customs have to do with the way people greet each other eat dress get married bury their dead etc. Ceremony fashion and manner are sister-categories of custom. It is a custom of my country but not of the Anglo-Saxon countries to thank the hosts or the heads of a family when the meal is finished. This is regularly done. A member of the community who—either exceptionally or habitually—does not do this is regarded with disapproval. A ‘foreigner’ to the community may be excused for not knowing or not adopting the custom.

Habits and customs qua regularities of behaviour show a certain resemblance to the regularities of nature which natural scientists study. Social anthropology is largely a science des mœurs. It is ‘descriptive’ in much the same sense in which natural science is descriptive.

Yet there is a difference ‘in principle’ between regularities of behaviour such as customs and laws of nature. This difference is not that the former regularities are ‘statistical’ and admit of exceptions the latter regularities ‘nomic’ and exceptionless. There seems to be no objection to calling some statistical regularities ‘laws of nature’. It is not the bare existence of exceptions to a rule that constitutes the difference ‘in principle’ between customs and regularities in nature. The difference lies in the way in which exceptions may occur. There is a sense in which the human individual can ‘break’ the rule of custom and in which the course of nature cannot ‘break’ its (causal or statistical) laws.

We can characterize this difference between customs and laws of nature by saying that the former present a genuinely normative or prescriptive aspect which the latter lack. Customs are ‘normlike’ in the sense that they influence conduct; they exert a ‘normative pressure’ on the individual members of the community whose customs they are. The existence of this pressure is reflected in the various punitive measures whereby the community reacts to those of its members who do not conform to its customs. In this respect customs are entirely unlike laws of nature and resemble not so much norms which are rules as norms which are prescriptions.

Yet there are important differences too between customs and prescriptions. Customs first of all are not given by any authority to subjects. If we can speak of an authority behind the customs at all this authority would be the community itself including both its past and present members. Customs could aptly be characterized as anonymous norms or prescriptions. But this characterization must not encourage any mysticism about the nature of the community as norm-giver.

Another difference between customs and prescriptions is that the former do not require promulgation by means of symbolic marks. They need not be ‘written down’ anywhere in so many words. On this ground they could also be called implicit prescriptions. It is an interesting problem whether within an animal or other community without a language customs which exert a normative pressure on the members are (logically) possible.

There are some respects in which customs are more like rules than like prescriptions. Customs determine or as it were ‘define’ ways of living which are characteristic of a certain community. A member who does not live in accordance with custom is seldom sought out for punishment in the same way as he who breaks the laws. The awkwardness of his position is more like that of a child who stands aside and does not want to join in the games of his playmates. He becomes a ‘stranger’ to his community rather than an ‘outlaw’.

7. A third main type of norms beside rules and prescriptions are those which I shall call directives or technical norms. They are approximately speaking concerned with the means to be used for the sake of attaining a certain end.

‘Directions for use’ are examples of technical norms. In them is presupposed that the person who follows the directions aims at the thing (end result) with a view to the attainment of which those directions are laid down.

I shall regard as the standard formulation of technical norms conditional sentences in whose antecedent there is mention of some wanted thing and in whose consequent there is mention of something that must (has to ought to) or must not be done. An example would be ‘If you want to make the hut habitable you ought to heat it’.

Shall we say that the sentence quoted is ‘descriptive’ or ‘prescriptive’? The proper answer it seems to me is that it is neither.

Compare the sentence under discussion with the sentence ‘If the house is to be made habitable it ought to be heated’. This last sentence I would not hesitate to call (purely) descriptive. It says that heating the house is a necessary condition of making the house habitable. This is (or is not) true independently of whether anyone wants to make the house habitable and aims at this as an end. An equivalent formulation of the sentence would be ‘Unless the house is heated it will not be habitable’. We could say that the normal use of either is to make a statement about men's living conditions. The truth which the statement affirms is a kind of primitive ‘law of nature’.

A statement to the effect that something is (or is not) a necessary condition of something else I shall call an anankastic statement. A (type of) sentence the normal use of which is for making an anankastic statement I shall call an anankastic sentence. A sentence which is used for making an anankastic statement can also be said to express an anankastic proposition.

It would be a mistake I think to identify technical norms with anankastic propositions. There is however an essential (logical) connexion between the two. In giving the directive ‘If you want to make the hut habitable you ought to heat it’ it is (logically) presupposed that if the hut is not being heated it will not become habitable.

Another confusion to be avoided is that between technical norms and what I propose to call hypothetical norms. By the latter I understand approximately speaking norms concerning that which ought to or may or must not be done should a certain contingency arise. Hypothetical norms too are usually formulated by means of conditional sentences. For example: ‘If the dog barks don't run.’ This sentence would normally be used for prescribing a certain mode of conduct in case a certain thing should happen. The norm which the sentence enunciates is a prescription.

In the ‘background’ of a hypothetical norm (prescription) too there is often an anankastic proposition. Why must I not run if the dog starts to bark? If I run the dog may attack me. Therefore if I want to escape being attacked by the barking dog I must not run. Here the technical norm—or the underlying anankastic proposition—explains why the hypothetical prescription was given to me. But this connexion is accidental not essential. Neither the technical norm nor the anankastic relationship is (logically) presupposed in the giving of the hypothetical norm (prescription). Even if there existed no technical norm or anankastic relationship in the background the hypothetical order not to run if—could be given to a person. (Cf. Ch. IX Sect. 3.)

A man argues with himself: ‘I want to make the hut habitable. Unless it is heated it will not become habitable. Therefore I ought to heat it.’ I shall call this type of argument a practical inference. In it the person who conducts the argument extracts as it were a prescription for his own conduct from a technical norm. Such ‘autonomous’ prescriptions given by a man to himself are however very unlike the ‘heteronomous’ prescriptions categorical or hypothetical given by a norm-authority to some norm-subject(s). It is doubtful whether one should call the former ‘prescriptions’ at all. (Cf. Ch. V Sect. 8.)

8. What is the position of so-called moral norms (principles rules) in the division of norms into main groups?

An answer to the question might be easier if we could give obvious examples of moral norms. This however is not altogether easy. One example which appears relatively uncontroversial (as an example) is the principle that promises ought to be kept. It is however an example of a moral norm of a rather special character. Other examples would be that children ought to honour their parents that one must not punish the innocent or that one should love one's neighbour as oneself.

Are moral norms to be classified along with rules of a game i.e. do they determine (define) a practice? It seems to me that on the whole moral norms are not like rules (in the sense which we here give to the term). But some moral norms present this aspect too. It is an aspect of the obligation to keep promises that this obligation is inherent in or is a logical feature of the institution of giving and taking promises. ‘By definition’ one could say promises ought to be kept. But this is only one aspect beside others of the obligation in question.

Are moral norms to be classified with the customs of a society (community)? It is noteworthy that the word ‘moral’ derives from the Latin mos which means custom. Some moral philosophers have sought to reduce ethics to a branch of a general science des mœurs. It seems to me that some moral ideas can be profitably viewed by the philosopher too against a background of the customs (traditions) of a community. This might be true for example of moral ideas in matters relating to sexual life. Other moral norms however seem to have no significant place in this perspective. To try to explain the obligation to keep promises for example in terms of the ‘normative pressure’ of customs seems utterly out of place.

Are moral norms prescriptions? If we think they are we must also be able to tell whose prescriptions to whom they are. Who gave the moral law?

A contract is a kind of promise. The legal obligations which people have under contract are therefore obligations to keep a kind of promise. The legal norms which institute these obligations are prescriptions. They can truly be called somebody's prescriptions to somebody—in spite of the fact that their authority is not a human individual or ‘physical’ person. But the moral norm to the effect that promises ought to be kept cannot become identified with the sum total of such legal prescriptions ‘supporting’ it. The laws of the state frequently have a ‘moral content’ or are concerned with ‘moral matters’. The same is true of the prescriptions which parents issue for the conduct of children. In the moral life of man prescriptions thus play a prominent role. This is no mere accident; it is a logical feature of morality. But this logical tie between moral norms and prescriptions does not so far as I can see reduce the former to a species of the latter.

Some think that moral norms are the commands of God to men. The moral law is the law of God. To take this view of morality is to regard moral norms as prescriptions. These prescriptions however are not only of a very special kind. They must perhaps be thought of as prescriptions in a special sense of the term. This is so because of the peculiar nature of the (supernatural) authority who is their source.

The chief alternative in the history of ethics to the view of morality as the laws of God is a teleological view of it. On the first view moral norms are a kind of prescription—or prescriptions in some special sense of the term. On the second view moral norms are a kind of technical norm or directives for the attainment of certain ends. But what end or ends? The happiness of the individual or the welfare of a community? Eudaimonism and utilitarianism are variants of a teleological ethics. It would seem that the ends relative to which certain modes of conduct are morally obligatory or permissible cannot be specified independently of considerations of good and evil. This holds true also of happiness and welfare as proposed ultimate ends of moral action.

In view of the difficulties encountered by both a law-conception of moral norms and a teleological conception of them it might be suggested that moral norms are sui generis. They are ‘conceptually autonomous’ a group of norms standing by themselves and not prescriptions for conduct in conformity with the will of a moral authority or directives for the attainment of moral ends. The view of moral norms as sui generis is sometimes called the deontologist position in ethics.

This is not the place for detailed criticism of deontologism in ethics. As a proposed way out of difficulties this position seems to me to be definitely unsatisfactory. The peculiarity of moral norms as I see them is not that they form an autonomous group of their own; it is rather that they have complicated logical affinities to the other main types of norm and to the value-notions of good and evil. To understand the nature of moral norms is therefore not to discover some unique feature in them; it is to survey their complex relationships to a number of other things.

9. The norms of various categories of which we have so far been talking are mainly norms concerned with that which ought to or may or must not be done. Laws of nature and other anankastic propositions are on the whole not concerned with action; but these we have decided not to call ‘norms’.

There is however a group of norms which are immediately concerned not with action but with things that ought to or may or must not be. German writers sometimes make a distinction between Tunsollen and Seinsollen.1 In Anglo-Saxon writings the distinction is not very often referred to.2

Following G. E. Moore3 I shall call norms which are concerned with being rather than with doing ideal rules. Ideal rules are referred to for example when we say that a man ought to be generous truthful just temperate etc. and also when we say that a soldier in the army should be brave hardy and disciplined; a schoolmaster patient with children firm and understanding; a watchman alert observant and resolute; and so forth.

We also say of cars watches hammers and other implements which are used to serve various purposes that they ought to have certain properties and should not have others. The question may be raised whether such statements should be counted as stating ideal rules or as anankastic propositions about the relations of means to ends. That question will not be discussed here.

Ideal rules are closely connected with the concept of goodness. The properties which we say a craftsman administrator or judge ought to possess are characteristic not of every craftsman administrator or judge but of a good craftsman administrator or judge. The person who has the properties of a good so-and-so in a supreme degree we often call an ideal so-and-so. The same holds true of watches cars and other things which serve various human purposes.

The features which ideal rules require to be present in good members of a class or kind of human beings can be termed the virtues characteristic of men of that class or kind. In an extended sense of ‘virtue’ roughly corresponding to the Greek arete the characteristic properties of good instruments are often called virtues also.

It is natural to call ideal rules concerning men in general as distinct from men of a particular class or profession moral rules or ideals. It is useful to distinguish between moral principles which are norms of moral action and moral ideals which set the pattern of a good man.

It may be thought that ideal rules are reducible to norms of action. The concepts of a brave generous just etc. act it may be argued are primary to the concepts of a brave generous just etc. man. The man who does brave acts is ‘by definition’ a brave man and so forth. This however would be to take a much too simple-minded view of the relationship in question. Yet it is also clear that ‘education’ (in the broadest sense) towards ideals will have to make use of prescriptions and other norms of conduct.

There is a certain similarity between ideal rules and technical norms. Striving for the ideal resembles the pursuit of an end. It would however be a mistake to think of the ideal rules as norms concerning means to ends. In order to be a good teacher a man ought to have such and such qualities. In order to fetch a book from the top shelf of his bookcase he ought to use a ladder. But those qualities of a man which determine his goodness as a teacher are not causally related to the ideal—as the use of a ladder may be a causal prerequisite of fetching a book from a shelf. The former relation is conceptual (logical). The ideal rules determine a concept e.g. the concept of a (good) teacher or soldier. In this they are similar to rules of a game. It is because of this similarity that we have given them here the name ‘rules’.

10. Our discussion in the preceding sections of the field of meaning of the word ‘norm’ has led us to distinguish between three major groups or types of norms. We have called them rules prescriptions and directives.

As a prototype of rules we instance the rules of a game. Rules of grammar also belong to this type of norm. Perhaps the so-called laws or rules of logic and mathematics should also be counted as belonging to it.

As prescriptions we count commands permissions and prohibitions which are given or issued to agents concerning their conduct. The laws of the state are prescriptions.

Directives we also call technical norms. They presuppose ends of human action and necessary relationships of acts to these ends.

In addition to these three main groups of norms we mentioned three minor groups of particular importance. They are customs Moral principles and ideal rules. It is characteristic of the minor groups that they show affinities to more than one of the major groups—they fall so to speak ‘between’ the major groups.

Thus customs resemble rules in that they determine quasi define certain patterns of conduct—and prescriptions in that they exert a ‘normative pressure’ on the members of a community to conform to those patterns.

On the nature of moral principles there has been much controversy and disagreement. Some philosophers regard them as a kind of prescription—say as the commands or laws of God to men. Others regard them as some sort of technical norm or directive of how to secure ends of a peculiar nature. Irrespective of which view one accepts as basically true one cannot deny that moral principles have important relationships both to prescriptions and to technical norms. The prescriptive aspect of morality moreover is related to custom. The ‘technical’ aspect of morality is related to ideals of the good life and man.

Ideal rules finally can be said to hold a position between technical norms about means to an end and rules which determine a pattern or standard.

• 1. See, for example, Nicolai Hartmann, Ethik (1925), Teil I, Abschnitt VI, Kap. 18–19. Max Scheler, in Der Formalisms in der Ethik und die materiale Wertethik (1916), uses the terms ‘ideales Sollen’ and ‘normatives Sollen’.
• 2. An exception is G. E. Moore, who draws the distinction very neatly in his paper ‘The Nature of Moral Philosophy’ (in Philosophical Studies, 1922).
• 3. Op. cit., pp. 320 f.