1. IT is a widely entertained opinion that value-concepts are intrinsically normative notions. This opinion is reflected in a certain philosophic jargon which tends to confuse or to mix value-terms with normative terms. When for example some writers insist upon the value-free nature of science (Wertfreiheit der Wissenschaften) they often give as a reason that science can tell us how things are but not how they ought to be. In the spirit of Hume philosophers talk of a Great Divide between fact and norm between the is and the ought—but also of a Great Divide between fact and value as if the two divides were the same. (Cf. above Ch. I sect. 2.)
Which is then the alleged normative nature of value? When one tries to give a clear answer to the question one immediately runs up against difficulties.
To say that the good is something which ought to exist or ought to be pursued is not only very vague but can easily be seen to be an untenable opinion unless stated with heavy qualifications. Ought apples to be good? Ought good apples to be eaten? Must one choose the better of two instruments? Whom and in what way does the goodness of a good runner oblige?
A supporter of the idea that goodness is intrinsically normative would perhaps when faced with these questions wish to qualify his opinion and restrict it to ‘moral’ goodness only. Are morally good acts then morally obligatory? This is not at all obvious. It may on the contrary be argued that moral goodness is ‘over and above’ obligation and that no man is or does good merely on the ground that he does not neglect his moral duties.
But is it not clear at least that the morally bad or evil must not be done? The normative flavour of the notion of moral evil is certainly more obvious than the normative flavour of the notion of moral goodness. Yet according to the opinion of G. E. Moore to whom we owe one of the most serious efforts ever made at a formulation in precise terms of the relationship between ‘good’ and ‘ought’ evil-doing is permissible (right) provided the evil done is ‘outbalanced’ by the good consequences of one's action. To me this seems morally most objectionable. I am not however suggesting that Moore's position could be refuted by an appeal to moral intuitions. But the fact that Moore who was an intuitionist himself arrived at his challenging opinions at least shows how difficult it is to tell what the connexion is between norm and value. It may even make us doubt whether there is any immediate intrinsic connexion between the two at all.
Beside the problem how norms and values are related there is also the problem of which of the two norms or values is more fundamental. One sometimes distinguishes some main types of ethical theory according to the solution which the theories propose to these two problems. By an ethics of value (‘Wertethik’) one can conveniently understand a view according to which value is fundamental and norms are somehow to be extracted from or established on the basis of value-considerations. The views of Moore and Brentano are clearcut examples of this type of ethics. (But not any ethics of value needs as theirs be an ethics of consequences.) By an ethics of duty (‘Pflichtethik’) again one may understand a view which regards duty (ought) as a basic idea. The ethics of Kant is the standard example. As well known Kant thought that the good will was the only unconditionally valuable thing in the world and that this will was a will to do one's duty for duty's sake. Yet it would be rash to maintain that Kant wanted to derive value concepts from normative notions. It may even be doubted whether there has been any serious attempt in the history of ethics to do this. What has been maintained however is that there is a ‘moral ought’ which is sui generis and not founded on value. This is sometimes called the deontologist view. As one of its best and most forceful champions we may regard the late Professor Prichard.
It seems to me that the discussion of the relation between norms and values even in recent times has suffered from the narrowing and obscuring implications of the term ‘moral’. If we want to get to know what values as such have to do with norms as such or to know the general nature of the connexion if there is one between norms and values we must disentangle the two from their associations with morality and study them in the widest possible generality. We must try to link a General Theory of the Good with a General Theory of Norms. The usefulness of the ‘general theories’ it seems to me shows itself most convincingly in the approach to the old problem of the relation of Good and Ought. In the new light on this relation we shall also I would maintain come to a better philosophic understanding of the nature of morality.
2. A General Theory of Norms is not an objective of the present inquiry. A few general remarks on norms will nevertheless be necessary in this place. I hope to be able to substantiate these remarks with fuller arguments in another investigation.
Norms in the sense here contemplated may be said to be prescriptions for human action. One may distinguish three main aspects of such prescriptions. I shall call the three aspects norms as commands norms as rules and norms as practical necessities. (The reasons for calling them ‘aspects’ rather than ‘kinds’ of norm or prescription I shall not try to state in this place.)
One may also distinguish between three main ways of formulating norms in language. These three ways may with some caution be called ‘linguistic counterparts’ of the three main aspects of norms which I mentioned. The correspondences however are by no means rigorously maintained in ordinary usage.
The first linguistic counterpart of norms are sentences in the imperative mood. For example: ‘Open the window’ ‘Never tell a lie’ ‘Honour your parents.’ Imperative sentences may be said primarily to answer to the command-aspect of norms.
The second linguistic counterpart of norms I shall call deontic sentences. These are roughly sentences employing the auxiliary (‘deontic’) verbs ‘ought to’ ‘may’ and ‘must not’. ‘You must not open the window’ ‘Children ought to honour their parents’ ‘You may play in my garden.’ Deontic sentences could perhaps be said to correspond primarily to the rule-aspect of norms.
The third linguistic counterpart of norms I propose to call anankastic sentences. These are sentences employing the auxiliary (‘anankastic’) verbs ‘must’ (‘has to’) ‘need not’ and ‘cannot’. ‘I must be at the station in time for the train’ ‘You need not help me with this’ ‘He cannot travel to-morrow.’ As expressions of norms anankastic sentences primarily answer to the aspect of norms as practical necessities.
As already said ordinary language does not uphold a sharp distinction in meaning between the three types of sentence. The meanings of imperative and deontic sentences shade into one another. ‘Open the window’ and ‘You ought to open the window’ can mean exactly the same. The border between deontic and anankastic sentences is also vague. ‘You may go for a walk’ and ‘You need not stay at home’ can be alternative ways of stating the same permission. The classification of sentences (and auxiliary verbs) as ‘deontic’ or ‘anankastic’ is therefore to some extent arbitrary. (We have classified ‘must’ as anankastic but ‘must not’ as deontic.)
Anankastic sentences have a common use for expressing beside practical also logical and natural (causal physical) necessities. Deontic sentences too are not infrequently used in connexion with these latter modalities. ‘If the wire is to carry the weight it ought to be at least one inch thick’ we may say. Instead of ‘ought to’ we could also have said ‘must’ or ‘has to’. Of that which is logically or physically contingent it is just as natural to say that it may be thus or otherwise and to say that it need not be thus or otherwise.
Statements of natural necessity are not by themselves prescriptive of action. They are therefore not norms in our sense though they are commonly called ‘laws’ of nature. But they are nevertheless often relevant to human action and thus indirectly to norms.
We say for example that if the house is to be habitable it has to (ought to) be heated. This is natural necessity—about the living conditions of human beings. It has as such nothing to do with a command or an obligation or even with a practical necessity to heat the house. But it may become connected with a norm. It does so in the first place by engendering a practical necessity. How this happens we shall soon have occasion to study.
3. First however some main distinctions and points relating to the aspect of norms as commands must be made.
One can distinguish between positive and negative commands. Positive commands are orders to do negative commands are orders to forbear. Negative commands are also called prohibitions.
A command is issued by somebody—we call it a norm-authority. It is further addressed to some agent or agents—we call them norm-subjects. The command has a content: this is the category of act or activity the doing or forbearing of which is prescribed. The norm as command finally is issued for one or several occasions on which the subject or subjects have to do or forbear the prescribed thing.
Normally when a orders s to do p a can be said to want s to do p. In commanding the authority makes or tries to make the subject do the commanded thing. The issuing of commands may therefore be characterized as efforts to make agents do or forbear things.
The issuing of commands I shall call normative action. The result of a normative act is the coming into existence of a certain relationship among agents. I shall speak of this as a normative relationship or relationship under norm.
An important aspect of normative action is the promulgation of the command or norm. The authority promulgates the norm to the subject or subjects i.e. he makes known by means of symbols (usually language) what he wants the subjects to do or forbear.
To the promulgation of the norm is further in many cases attached a threat of punishment or sanction in case the subjects should not comply with the order. It may be argued that unless there is at least an implicit threat of punishment i.e. of some evil consequent upon disobedience then there is properly speaking no command either.1 I shall accept this view.
If the norm-authority is different from the norm-subject(s) the command or norm is called heteronomous. This may be regarded as the normal case. If authority and subject are one and same the command or norm is called autonomous. It must not be taken for granted that the notion of a self-command makes sense. I hope presently (sect. 8) to be able to show that the conception of autonomous norms as self-commands is an analogical supplementation of the command-aspect of norms.
4. With the statement of natural necessity ‘If the house is to be habitable it must (ought to) be heated’ may be compared the form of words ‘If you want to inhabit the house you must (ought to) heat it.’
In the second sentence there is mention of something wanted i.e. of a state of affairs which is or may be desired as an end of action. Such sentences express that which I propose to call technical norms. Technical norms are roughly the same as that which Kant called ‘technical imperatives’ and which he regarded as a sub-class of the broader category of ‘hypothetical imperatives’.2
The formulations of technical norms ordinarily employ a deontic or an anankastic vocabulary. But there is no strong objection to the use of the imperative mood in them. Instead of saying ‘If you want to inhabit the house you ought to heat it’ we could also say ‘If you want to inhabit the house then heat it.’
What is the logical nature of technical norms? Do they express propositions? How are they related to commands? And how are they related to statements of natural necessity? These are questions of considerable complexity and difficulty.
In order to facilitate understanding of the nature of technical norms let us first ask what could be the use of the sentence ‘If you want q you must do p.’ The sentence could for example be used to inform a person of the existence of a causal tie between a certain human act and a certain consequent state of affairs which may be desired as an end of action. Or it could be used to remind him of the existence of this tie. Or it could be used to give him a piece of advice or a recommendation.
The person who reminds or gives advice is probably not indifferent to the acts and aims of the other person. But saying to another person ‘If you want q you must do p’ or even ‘Since you want q you must do p’ would not normally signify an attempt or effort on the part of the speaker to make the addressee of his words do p. Therefore speaking thus would not normally be an act of commanding. No threat of sanction on the part of the giver of the advice is implied.
In these features technical norms differ sharply from that which I shall call hypothetical norms. The words ‘If it starts raining shut the window’ would normally express a hypothetical command. The person who utters them is anxious to influence the conduct of the person whom he addresses in such a way that should a certain contingency arise this person will perform a certain act. There is no causal connexion however between this act and the state of affairs.
I think it is correct to say that Kant much to the detriment of his ethics did not distinguish technical norms from hypothetical norms. The ‘if-then’-form of words used for expressing norms of either kind seems to have influenced him into thinking that the two were essentially the same sort of rule. This is by no means the case.
Have technical norms then nothing at all to do with commanding? Consider a man who wants to attain an end q e.g. to make an empty hut habitable for himself. He is deliberating about the means to this end i.e. about that which he—as we commonly express ourselves—ought to do or has to do or must do in this situation. He might then say to himself ‘If you want q you must do p.’ Such soliloquies even in the second person occur. The ‘if’ is not here the conditional ‘if’. It does not mean ‘in case’. It could be replaced by ‘since’. And that which he says to himself could also be explicated and given the form of a ‘syllogism’. As follows:
You want q.
Unless you do p you will not get q.
∴ You must do p.
The first premiss is a want-statement. The second premiss is a statement of natural necessity. It is ‘purely objective’; there is no mention or even hint of wants in it.
Is the conclusion a command? One could call it a quasi—command which the person who conducts the argument addresses to himself. I shall however prefer to call it a statement of practical necessity. In it the person who reasons reaches the conclusion that his wants plus a certain natural necessity impose upon him the practical necessity of acting in a certain manner.
I shall call the whole argument a Practical Syllogism. The form of words ‘If you want q you must do p’ used in soliloquy could be regarded as a contracted form of the syllogism which skipping the second premiss proceeds directly from the first premiss to the conclusion. The technical norm on this analysis splits into the three components of a want-statement a statement of natural necessity and a statement of practical necessity. The last we noted shows a certain resemblance to a command.
Not in the technical norms themselves but in the conclusions of practical syllogisms are we confronted with that aspect of norms which in section 1 I called the aspect of norms as practical necessities.
The name ‘practical syllogism’ can be used and has been used to mean different things which it is important to keep sharply apart. The great obscurity of Aristotle's doctrine of the practical syllogism is partly due it seems to me to a failure to uphold the relevant distinctions here.
By a Practical Syllogism one can understand a pattern of reasoning in which both premisses and conclusion are norms. For example:
It is permitted to do p.
One must not leave q undone if one does p.
∴ It is permitted to do q.
Such patterns of reasoning are studied in the branch of logic which is nowadays commonly known under the name Deontic Logic.
By a Practical Syllogism one also often understands a pattern in which one normative and one factual premiss yield a normative conclusion. For example:
All thieves ought to be hanged.
This man is a thief.
∴ This man ought to be hanged.
This pattern too may be relegated to the province of Deontic Logic. But the practical syllogisms which illustrate how wants and natural necessities engender practical necessities of action require a theory of their own. They have so far been very little studied by logicians. A formal study of them will not be attempted in this work.
5. The second premiss of our practical syllogism on p. 161 ‘Unless you do p you will not get q’ we called a statement of natural necessity. q is something which is or may be an end of action. p is the result of an act which is causally connected with this state of affairs q. Considering this we can also say that the second premiss is a statement about means to an end.
There are at least two typical uses of the word ‘means’ (to an end).
Under the one use the term signifies a thing or a kind of thing. Such a thing is often also called an instrument. For example: To cut a cloth I may need a pair of scissors. ‘By means of’ the instrument I achieve an end or goal. By ‘means of production’ or ‘means of transportation’ we usually understand means in this sense of the word or in some closely related sense.
Under another use of the term ‘means’ signifies an action. For example: By turning the key I open a door. Turning the key is an act by means of which I achieve an end viz. the opening of the door. The key too is here a means viz. a means in the sense of instrument.
There is an obvious logical relationship between the two uses of the term ‘means’. Means in the second sense is very often but not always the same as the use of a means in the first sense. The turning of the key as a means of opening the door is the use of the key as an instrument.
Instruments would not be means to ends unless they were used i.e. unless there was human action aiming at certain goals. Because of this we can say that ‘means’ in the sense of action is primary to ‘means’ in the sense of instrument. (But there may be other reasons for saying the reverse.)
Means in the sense of instruments we have discussed before. Instruments it will be remembered are called good when they serve well a purpose for which they are used. This is what we called ‘instrumental goodness’.
Here we are concerned only with means which are acts. The relation which we wish to study between means and ends is thus a relation between certain acts and certain states of affairs. Our question is now: Which conditions must those acts and those states of affairs satisfy in order that we shall say that the former are related to the latter as means to potential ends? The question is a complex and difficult one and our treatment of it here cannot pretend to be exhaustive.
Can an act be a means to its own result as an end? If for example the desired state of affairs is that the window be open is then the act of opening the window a means to achieving this end? It seems to me that calling it a ‘means’ is one of the uses of the word—though perhaps not a very common one. When individual acts are viewed as means to their own results as ends the acts are often classified as falling under different ways of doing this thing i.e. of achieving this result in acting. These ways may then be rated as good or bad or one as better than another. For example: Let ‘going there by bus’ and ‘going there by train’ be two ways of reaching a certain destination. I can compare the two with regard say to expense time and comfort. On the basis of this comparison I may rate the one as a better way of travelling to the destination than the other. Such an attribution of goodness to a way of doing something is we have said before closely related to the attribution of goodness to instruments.
The means-end relation if we call it by that name between an act and its result is an intrinsic relation. It has therefore nothing to do with the natural necessities referred to in the second premisses of practical syllogisms. It is accordingly not a means-end relationship of the kind in which we are now interested.
The means-end relationship between acts and states of affairs to which the second premisses of practical syllogisms refer is a causal or extrinsic relation. As previously observed the relation between an act and its consequences is extrinsic. (See Ch. VI sect. 3.) So is also the relation between an act and its causal requirements or prerequisites. (Cf. Ch. V sect. 8.)
Not everything which is a consequence of action needs be an end of action. Some consequences are not foreseen at all. Others are foreseen but not yet ends i.e. consequences for the sake of the production of which the act was undertaken. To those consequences which are ends and to them only can the act rightly be said to stand in the relation of means to end.
Often the attainment of an end requires or presupposes the performance of an act the result of which is not the same as the end. In order to fetch a book from the top-shelf of my bookcase it may be necessary for me to step on a ladder. My climbing the ladder is here a ‘necessary preparation’ or a ‘causal requirement’ of my fetching the book. It can also without twisting ordinary usage be called a means to an end the end being the fetching of the book. This is true generally of every act which is a causal requirement of the attainment of a given end.
There are thus two basic types of causal or extrinsic relationships between acts as means and states of affairs as ends:
The one is a relation between an act and its consequences. If doing p produces a state of affairs q different from p and if q is an end of human action then the doing of p is a means to this end.
The other type is a relation between acts and their causal requirements. If the production of a state of affairs q requires the doing of p and if q is an end of human action then the doing of p is a means to this end.
I shall call means of the first type productive means and means of the second type necessary means.
Productive and necessary means are in the following sense inter-definable:
Let the doing of p be a productive means to the end q. Then the forbearing of p will be a necessary means to the end ∼q. For example: If opening the window is a productive means to cooling the room then keeping the window closed is a necessary means to keeping the temperature in the room from sinking.
Let the doing of p be a necessary means to the end q. Then the forbearing of p will be a productive means to the end ∼q. For example: If turning the radiator on is a necessary means to increasing the temperature in the room then keeping the radiator closed is a productive means to preventing the temperature from rising.
A means to an end can be both productive and necessary. When this is the case we say that the means is the only means to the end in question. In a room where the windows are closed the only means of altering the temperature at will may be by regulating the radiator.
It Is important to observe that the causal relation between productive or necessary means and an end of action need not be a relation of so-called universal implication between an act-category and a generic state of affairs. It can also be a ‘laxer’ relationship of probability. When for example we say that unless water is boiled there is a risk that contagion shall spread or that if there are more than 40 passengers on board the vessel the passage will not be safe the words ‘risk’ and ‘safe’ indicate a causal relationship of a probabilistic nature. Yet boiling the water can quite appropriately be termed a necessary means or precaution to prevent contagion from spreading and not allowing more than 40 persons on board the vessel a necessary means of making the passage safe.
Productive and necessary means are both that which in Ch. III sect. 4 we called favourably causally relevant to the attainment of an end. They therefore possess qua means to ends utilitarian goodness. This value-aspect of theirs however is not relevant to the discussion in this chapter of the relation of value to norm.
The relations of productive means to ends are natural necessities just as much as the relations of necessary means to ends. The means-end relationship however to which the second premisses of practical syllogisms (on our understanding of the term) refer are causal or extrinsic relations between ends of action and necessary means to those ends. It is therefore with the notion of a necessary means to an end that we are primarily concerned in the present discussion.
6. Reference to causal means-end relationships is often made when we wish to explain (interpret understand) human action.
Suppose we are anxious to get to understand why a person behaves in a certain way or does something which is exceptional or surprising or such that a man would not ordinarily do this ‘for its own sake’. For example: We see a man running in the street and wish to know why he does this. We are told that he is running in order to catch a train.
If we accept this reply as an explanation of the person's behaviour we assume that the agent regards his action running as a productive means towards this end.
It is however not certain that the reply ‘He ran in order to catch the train’ will be accepted as an explanation of the person's behaviour. Why did he not take a taxi? May be our rejoinder. The man must know that this is another productive means towards his end and probably a better (more efficient) one. What we wish to know is why he used that very means which he actually did use.
Suppose now that we are told that the man had to run because this happened to be the only means for him to reach the train. There was say no taxi available in the street or he had no money on him to pay the taxi-fare. Saying that the man bad to run in order to catch the train is to say that running was a necessary means to his end.
To explain an action by giving reference to productive means to a desired end is to leave open a question to which we may want to have the answer before we regard the explanation as complete. This is the question: Why the agent used these very means? If we are told that the means were necessary as well as productive we regard the explanation as complete or exhaustive of the case.
The explanation which makes reference to necessary means can also be exhibited in the form of a practical syllogism as follows:
x wants to reach the train in time.
Unless x runs he will not reach the train in time.
∴ x has to run.
What this syllogism is supposed to explain is x's action. It explains this by making it plain that the action was a practical necessity incumbent upon the agent.
A practical syllogism when offered as an explanation of action which has taken place in the past or is taking place in the present I shall call a third person practical syllogism. When the explanation is given by the agent himself he will normally be speaking in the first person. ‘I want(ed) to reach the train in time etc.’ This I nevertheless call a third person syllogism. The agent is here speaking about himself as it were viewing himself from the outside.
Is a third person practical syllogism a logically conclusive argument? This is not a futile question to ask. On the one hand our syllogism above seems absolutely conclusive flawless watertight. On the other hand one would look in vain through all the text-books and works on logic for the inference scheme of which this syllogism could be said to be an instantiation. Its pattern is not one of the categorical or modal syllogisms of course. It has a certain resemblance to a modus ponens or perhaps rather a modus tollens argument. But certainly it is not an ordinary modus ponens or tollens either. Shall we try to supplement it with suppressed premisses and thus make it conclusive according to the laws of ‘ordinary logic’? I believe that this would be a blind alley. Nor would it help if we expanded the second premiss into ‘x believes that unless he runs he will not reach the train in time.’ Shall we deny then that the syllogism is logically valid? This way out too has been suggested—but seems to me to be a mere evasion. We must I think accept that practical syllogisms are logically valid pieces of argumentation in their own right. Accepting them means in fact an enlargement of the province of logic. We cannot reduce the practical syllogisms to other patterns of valid inference. But we can indeed must say something more to elucidate their peculiar nature.
7. A man is walking in the street towards the railway station. He wants to be there in time for the train. Perhaps he is expecting his fiancée. He looks at his watch and realizes with a start that the train will be there in a few minutes time and that if he continues at walking pace he will be late. There is no taxi within sight. He has to run. And he puts himself in motion.
We could call the situation just described a first person practical syllogism—though not a syllogism in words. Let us try to give a schematic presentation of this wordless argument A man has a certain aim. (First premiss.) He realizes that unless he does a certain thing he will not reach his aim. (Second premiss.) But what is the conclusion?
Shall we say that the conclusion is his insight that he has to do this thing? But what is this ‘insight’? As far as I can see the only thing here which deserves to be called an insight is the agent's understanding that unless he does a certain thing he will be late for the train. But this we said is the second premiss. If the conclusion were the second premiss repeated the first premiss would be totally irrelevant to the argument. This it obviously is not.
Shall we say then that the conclusion of the syllogism is the man's doing of the thing the necessity of which he has realized in the second premiss? Saying thus would be to take a view of the matter reminiscent of one which Aristotle appears to have held. In one place Aristotle explicitly says that the conclusion of the practical syllogism is an act.3 In other places it is not quite clear how he regarded the conclusion. It should be observed that calling an act a ‘conclusion’ from something is not at all odd. We very often call that which a man does the ‘conclusions’ which he has drawn from a certain situation.
I would say myself that calling the conclusion of our wordless syllogism an act is almost right—but not quite right. Consider the man on his way to the station. If his running is the conclusion when shall we say that the conclusion has been drawn? Suppose that after he has run one hundred yards he has a stroke and dies? Shall we then say that he died before he had drawn the conclusion—or before he had finished drawing it? Must he ‘draw’ it all the way to the station? (This is almost like asking: How ‘long’ must the conclusion be?)
Let us go back to the case as we originally described it. The description ended in the words ‘he puts himself in motion’. This in my opinion correctly describes the conclusion. Neither the man's insight into the necessary connexion between an act and an end nor his doing of the act is the conclusion of the wordless piece of argumentation but the man's setting himself to do the act. Putting oneself in motion is setting oneself to run. This is what the syllogism ends in.
Is ‘setting oneself to do something’ action? The agent who sets himself to do something also ‘puts himself in motion’ i.e. embarks on the road to the actual performance of the act. Drawing the conclusion of a first person practical syllogism need not take the agent to the very end at which he is aiming. But it must put him on the road to this end. Aristotle would have been quite right had he said that the practical syllogism leads up to action. It ends not necessarily in doing something but in setting oneself to do something.
When the act is completed we can ask in retrospect ‘What made bim do it?’ And the answer then is that his wanted end of action in combination with the insight that unless he does the thing he will fail of his end made him do it.
Sometimes the conclusion of a practical syllogism refers to the future. I want say to leave the town not later than the day after to-morrow. I realize that unless to-morrow I do so and so I shall not be free to travel the day after. Therefore I must do this thing to-morrow. The question may be raised: In what sense can deciding to-day to do something to-morrow be called ‘putting oneself in motion’ or ‘embarking on the road to the actual performance of the act’? The answer will be: In the sense that the agent will from now on (‘is from now on set to’) forbear doing things which are likely to prevent him from completing his set task tomorrow and do things which are in their turn necessary or useful for making the decided thing come off. He will for example decline an invitation to lunch and finish some piece of work instead of meeting his friends. The agent who has set himself to do something in future is like a ship moving towards a destination.
His forbearances and acts are to some extent preformed by his decision. He may become detracted from his set route but not without compelling reasons.
We have so far regarded the first person practical syllogism as a wordless argument. It could be accompanied by words and sometimes is. The agent might say to himself aloud or in thoughts only ‘I want to be at the station in time’ thus verbalising the first premiss. But the essential thing is not whether he says this but that he wants this as an end of his action. The agent may also say to himself: ‘Unless I run I shall be late’ thus verbalizing the second premiss. Yet the essential thing is not whether he says this but that he thinks thus. The agent may finally say: ‘I ought to run’ or which would do equally well: ‘Hurry up there is no time to be lost’ thus verbalizing the conclusion. But again it is not the words which matter but that he sets himself to do it.
In first person practical syllogisms words are ‘decorations’ ‘frills’ ‘inessential accompaniments’ only. Essential to it are the wanting of an end the understanding of a natural necessity and the decision to act. The way in which these three components unite in the syllogism I shall call practical necessitation—necessitation of the will to action through want and understanding.
‘Want’ and ‘understanding’ these are words related to ‘passion’ and ‘reason’. We are here in the neighbourhood of the problem discussed with so much fervour and ingenuity by Hume whether reason alone can move man to action. Hume said reason cannot move but passion is the mover. With this view he coupled an in many respects deep-cutting attempt to found obligations on interests. But I think it is right to say that Hume did not fully realize the role which reason plays in the origination of obligations. He did not see clearly through the logical mechanism of the practical syllogism. To show why something is an obligation founded on interest is not to show that it is something we (‘really’ ‘innermost’) want to do but that it is something we have to do for the sake of that which we want (to be to do to have to happen).4
We said at the end of the preceding section that practical syllogisms are logically conclusive arguments—but that it was difficult to see how they are this. The non-verbal first person syllogism provides the clue. If practical syllogisms are logically conclusive then the practical necessitation of the will to action must at the same time be a logical necessitation. This I think it is. Assume that the man in our example wants to be at the station to meet his fiancee that he quite clearly sees that unless he runs he will be late (and that if he runs he has at least a good chance of arriving in time) but that he does nothing to hurry up. What shall we then say about the case? One thing which we might say is this: Our man was so unwilling to run (perhaps he finds it undignified) that having realized that only by running he could arrive in time he altered his aims. ‘I would rather be late than run’ he says. He of course wanted to be at the station in time. He still wishes he could have been. Arriving in time to meet the train is moreover an in itself wanted thing to our man; if a God appeared and took him to the station in time he would welcome this. But he no longer wants this as an end of his action. Therefore he is not moved to act.
To regard practical syllogisms as logically conclusive arguments is I would say to knit or tie together in a peculiar way the concepts of wanting an end understanding a necessity and setting oneself to act. It is a contribution to the moulding or shaping of these concepts. The justification of this moulding procedure is partly its conformity with actual usage and partly that which it does to meet the philosopher's craving for clarity in these matters.
8. In the conclusions of first person practical syllogisms I shall maintain we encounter the autonomous norms. (Strictly speaking we encounter a concept of autonomous norm. For there are other phenomena too which qualify as candidates for the name. We shall not discuss them however in this work.)
The autonomous norms on this view are necessitations of the will to action under the joint influence of wanted ends and insights into natural necessities.
The concept of autonomous norm of which we are now treating is rather unlike Kant's. It has for one thing no specific connexion with so-called moral matters. Yet it is not in every relevant respect unlike the Kantian notion.
One point of resemblance has to do with a contrast which in Kantian phraseology could be called a contrast between ‘inclination’ and ‘duty’. There is a truth hidden in the idea which is so prominent with Kant and for which he has been severely censured by others that action under autonomous norm or rule must in some sense go against inclination. I think the truth is that action under such rule is never undertaken because we want to do it but is forced upon us by natural necessities and the remoter objects of our wants and likings. If the act is something which for its own sake we want to do then no autonomous norm is needed to move the will to action. When the act is in itself indifferent to us the movement towards action follows smoothly upon want and insight—we do not have to say to ourselves ‘I must do this.’ But when there is a practical necessity of doing something which is in itself unwanted or shunned by us then the incumbent nature of the autonomous obligation acquires prominence. It is chiefly in such cases that we resort to soliloquy and by keeping our goal and the insight into the necessity of the act clear before our minds urge ourselves on to action. We could say: the wider the gap between the must and the want to the more prominent the must; and if there is no gap at all—meaning that we do the act from sheer inclination—then there is no autonomous necessitation of the will either.
One crucial difference again between Kant's and our notion of an autonomous norm has to do with Kant's distinction between categorical and hypothetical imperatives. The notion of a hypothetical imperative with Kant is not altogether clear. It is anyhow closely related to our notion of a technical norm. Of the technical norms we said (sect. 4) that they may be regarded as contracted forms of practical syllogisms. The conclusions of practical syllogisms in the first person finally are what we (here) call autonomous norms. Even though they are categorical in form it seems obvious that this relationship of theirs or dependence upon technical norms marks them as different from the Kantian categorical imperatives and therewith from the Kantian autonomous norms.
The conclusions of practical syllogisms which I have called autonomous norms resemble heteronomous commands in that they both manifest efforts to make agents do or forbear. (Cf. above sect. 3.) This is an essential similarity which connects the aspect of norms as practical necessities with the aspect of norms as commands.
Are there features of autonomous norms which answer to promulgation and sanction? These two we said (sect. 3) were essential features of heteronomous commands.
There is an essential feature of autonomous norms which may be said to resemble promulgation by analogy. This is the agent's insight into or understanding of the causal connexion between the doing of a certain act and the attainment of a certain end. Such insight it seems is possible only for rational beings who are capable of conducting arguments and who possess some knowledge of the ways actions and events in nature are causally interwoven. It is plausible to think that such insight is possible only to beings who master a language. If this is true then another essential similarity between norms as practical necessities and norms as commands is that both presuppose language.
The conclusion of a first person practical syllogism we said need not be verbalized—not even ‘in thoughts’—and take the shape of a norm-formulation ‘do this’ or ‘you must (ought to) do this’ addressed as it were by the agent to himself. But it may become verbalized and a form of words may be used for soliloquy which strongly resembles the use of words for urging agents to action. This is an accidental feature of autonomous norms which may be said to resemble promulgation not by analogy but literally.
There is further an essential feature of autonomous norms which may be said to resemble sanction by analogy. If a man who is under a practical necessity of doing a certain thing fails to do this then he will also necessarily fail of his end of action. The end was a wanted thing. Not to get the wanted frustration of desire is intrinsically unpleasant we have said before. Failure to act in accordance with the norm is thus intrinsically connected with unpleasant consequences. This is an analogy of punishment. But it is only an analogy. Punishment properly so called is for disobedience and failure to carry an autonomous necessitation of the will into effect cannot properly be termed disobedience.
We are here touching upon an important difference between autonomous and heteronomous norms. With the second the agent is ‘in principle’ free to choose between obeying the rule and escaping the penalty and disobeying with a risk of being punished.
He may choose the second. But the autonomous norm comes into being with or consists in the agent's putting himself in motion as directed by his wants and insights. In a sense therefore autonomous norms are ipso facto ‘obeyed’. That is: the notions of obedience and disobedience do not apply to them.
If however failure to complete the act in conformity with the dictates of practical necessity is due to weakness of will or to negligence or thoughtlessness in action then it is due to something which is analogous to disobedience and which may for that reason be regarded as deserving punishment. The agent may invent various forms of self-chastisement e.g. with a view to strengthening his will-powers or to making him more careful in acting. When this happens autonomous norms are conjoined with an accidental feature which can be said to answer to sanction not by analogy but literally.
This will suffice as an answer to the question to what extent norms as autonomous practical necessities resemble norms as commands. Our next task will be to investigate to what extent and in what way norms as heteronomous commands may acquire the aspect of practical necessities.
9. Orders are sometimes issued for no particular reason. Then they have no foundation in the ends of the norm-authority. This case appears to be comparatively rare. Orders are sometimes issued in obedience to an order to issue them. Then the question why they are issued is the same as the question why the authority who issues them obeys the order of the higher authority. Finally orders are sometimes issued for the sake of an end.
A man x commands another man y to do a certain thing p. Why does x command y? The answer is often though certainly not always that x has some end in view towards the attainment of which his commanding y to do p is a necessary means. x is say engaged in some complicated task for the successful accomplishment of which he needs the assistance of y. y must do p if x is to succeed. But y does not do p out of his own inclination. y may have totally different ends of his action. Therefore x must make y do p if he is to succeed. Perhaps more gentle means than commanding are first tried but found fruitless. Therefore x must command y to do p if he is to succeed. That is: commanding y to do p is a necessary means of making y do p which in its turn is a necessary means of securing x's success. So x opens his mouth and gives the order to y ‘Do p’.
If the reasons why x ordered y to do p are such that in their light x's normative action assumes the character of a necessary means to his desired end then we shall say that x had a compelling motive or reason for giving the order to y. The normative act assumes this character when we come to think that had x not given the order to y he would not have attained his end.
That ordering y to do p is a necessary means for x either to p itself as an end or to some remoter end q is a statement of natural necessity. If x realizes this and if he aims at p or q then ordering y to do p becomes a practical necessity incumbent on x.
In this way the normative acts of issuing heteronomous commands or orders may become practically necessitated by the ends of the norm-authority in combination with considerations pertaining to means to those ends. When the normative act is thus necessitated we shall call the heteronomous norm which comes into existence as its result a well-grounded norm. A heteronomous norm is in this sense well-grounded when the norm-authority's normative action is guided by autonomous norm or is as we could also say it autonomously necessitated. Since autonomous norms are ipso facto practical necessities they may be said to be ipso facto well-grounded too.
Also obedience to (heteronomous) commands can be autonomously necessary.
It is an essential feature of the heteronomous norm we said (sect. 3) that it should be associated with a threat of penalty in case of disobedience. Threat of punishment for disobedience constitutes a motive for obedience. This is not to say that orders are always or even normally obeyed because the subject wants to escape punishment. But sometimes they are obeyed for this reason. Then the norm-subject considers obedience to the norm as a necessary and usually also productive means to escaping punishment. He wants to escape punishment. Unless he obeys he will be punished. Therefore he has to obey i.e. has to do the act which the norm says he ought to do.
Sometimes we are anxious to obey orders not for the sake of escaping punishment but for the sake of something else. The end could for example simply be our wish to please the person who gave the order. Perhaps he gave the order for the sake of attaining some end of his and we want to help him. Then our want in combination with the natural necessity of obedience to the norm as a means to his end engenders an autonomous practical necessity of obedience.
It can of course happen that the thing we are ordered to do is the very thing we want to do. Then there is no autonomous necessitation of the will to obedience. This case is therefore uninteresting from the point of view of the making-do-mechanism of commands. But there is a case which is related to it and which is of much interest. This is when the norm is addressed to a multitude of subjects and the state of affairs which prevails or comes into being thanks to the fact that all subjects obey the norm is an end of each one of the subjects individually. Then obedience is necessary as a co-operative step towards the attainment of the end. This is another case of autonomous necessitation to obedience to a heteronomous norm.
10. We may now go back to the problem of the relation between norms and values.
Ends are goods attainable through action we have said. To say as is sometimes done that ends demand or require us to pursue them is metaphorical speech. There is nothing normative about ends and goods as such. Ends are pursued and goods wanted or else they are not ends and goods. But the ordinances of ananke (that too of course is metaphorical talk) i.e. the natural necessities may when they become known to man in combination with his ends force upon him the practical necessities of doing things which for their own sake he would not do and of forbearing things from which considerations of ends apart he would not have abstained. In this way norms may be said to ‘hook on’ to values.
If by autonomous norms we understand the practical necessitations of the will to action under the joint influence of ends and insights into causal connexions then such norms are intrinsically value-directed one could say.
The heteronomous norms as commands or orders given by norm-authorities to norm-subjects have no intrinsic connexion with ends. But heteronomous norms may become grounded in ends and thus assume the appearance of practical necessities. If they are that which I here call well-grounded then the normative acts of issuing them are value-directed under autonomous norms.
Also the acts of obeying the norms can become thus value-directed. It will perhaps be objected to our treatment in this chapter of the problem how norms and values are related that we have been discussing only the relation of ‘good’ (or rather: ‘want’) to ‘must’. Traditional discussion of the problem has been concerned above all with the relation of ‘good’ to ‘ought’. (Cf. the presentation of the problem in sect. 1.) This is true—and also that the meanings of ‘ought’ and ‘must’ can be significantly distinguished. Our discussion of the relation of norms and values has been very far from exhaustive. In defence of the adopted course of treatment I would say that I tend to think that it is only the aspect of norms as practical necessities (the anankastic or ‘must’—aspect) which bears an intrinsic relationship to ideas of the good. Other aspects of the normative may become value-oriented only through the intermediary of the anankastic aspect. How this happens for the aspect of norms as commands I have tried to indicate in sects. 8 and 9. How it may happen for the aspect of norms as rules the deontic aspect as it could also be called (cf. sect. 2) will not be discussed here. Hints of this will be given in the last chapter.
- 1. Cf. Austin, The Province of Jurisprudence Determined, Lecture I, Section headed ‘The Meaning of the Term Command’: ‘A command is distinguished from other significations of desire, not by the style in which the desire is signified, but by the power and the purpose of the party commanding to inflict an evil or pain in case the desire be disregarded. If you cannot or will not harm me in case I comply not with your wish, the expression of your wish is not a command, although you utter your wish in imperative phrase.’
- 2. Kant divided the hypothetical imperatives, on the one hand into ‘problematic’ and ‘assertoric’, on the other hand into ‘technical’ and ‘pragmatical’. How the two grounds of division are related, according to Kant, is not quite clear. Problematic imperatives, broadly speaking, concern means to potential; assertoric imperatives again, means to actual ends. A man's own well-being or happiness (Glückseligkeit) Kant seems to have regarded as an actual and, moreover, necessary end of action. The pragmatic imperatives concern means to this peculiar end. They are therefore related to the notion of an autonomous self-regarding duty, which we shall discuss in the next chapter.
- 3. De Motu Animalium 701• 10–40. Cf. also EN 1147• 25–31. It should be remembered, however, that Artstotle's concept of a practical syllogism has a wider scope than ours.
- 4. It may, of course, be argued that, if an end is wanted, then use of the means, which are seen to be necessary for the attainment of the end, will a fortiori be wanted too. Cf. Kant, Grundhgung zur Metaphysik der Sitten (2nd ed.), p, 46: ‘Wer den Zweck will, will (sofern die Vernunft auf seine Handlungen entscheidenden Einfluss hat) auch das dazu unentbehrlich notwendige Mittel, das in seiner Gewalt ist. Dieser Satz ist, was das Wollen betrifit, analytisch.’ It is interesting to note that Kant regarded the proposition as analytical.