THE VARIETIES OF GOODNESS

by Georg Henrik von Wright

LONDON ROUTLEDGE & KEGAN PAUL NEW YORK : THE HUMANITIES PRESS
-iii-
First published 1963 by Routledge & Kegan Paul Ltd Broadway House, 68-74 Carter Lane
London, E.C.4
Printed in Great Britain by Richard Clay and Company, Ltd Bungay, Suffolk
Georg Henrik von Wright 1963
No part of this book may be reproduced in any form without permission from the publisher,
except for the quotation of brief passages in criticism
-iv-
PREFACE
IN 1959 and 1960 I gave the Gifford Lectures in the University of St. Andrews. The lectures
were called 'Norms and Values, an Inquiry into the Conceptual Foundations of Morals and
Legislation'. The present work is substantially the same as the content of the second series of
lectures, then advertised under the not very adequate title 'Values'. It is my plan to publish a
revised version of the content of the first series of lectures, called 'Norms', as a separate book.
The two works will be independent of one another.
I take this opportunity to express my thanks to the University of St. Andrews for honouring
me with the invitation to give the Gifford Lectures and to the members of staff and students at
St. Andrews, with whom I was able to discuss the content of the lectures when they were in
progress. Giving the lectures afforded me with an urge and opportunity to do concentrated
research, for which I am deeply grateful.
In the course of revising the contents of my lectures and preparing them for publication I have
had the privilege of regular discussions over a long period with Professor Norman Malcolm. I
am indebted to him for a number of observations and improvements and, above all, for his
forceful challenge to many of my arguments and views.
There is very little explicit reference to current discussion and literature in this book. I hope
no one will interpret this as a sign that the author wishes to ignore or belittle the work which
is being done by others. It is true, however, that the works of the classics have provided a
much stronger stimulus to my thoughts than the writings of my contemporaries. In particular
have I learnt from three: Aristotle, Kant, and Moore. I have been successively under the spell
of the Kantian idea of duty and the Moorean idea of intrinsic value. In fighting my way
against Kant I was led to reject the position sometimes called 'deontologist', and in resisting
Moore I became convinced of the untenability of value-objectivism
-v-
and -intuitionism. In this largely negative way I arrived at a teleological position, in which the
notions of the beneficial and the harmful and the good of man set the conceptual frame for a
moral 'point of view'. Perhaps one could distinguish between two main variants of this
position in ethics. The one makes the notion of the good of man relative to a notion of the
nature of man. The other makes it relative to the needs and wants of individual men. We
could call the two variants the 'objectivist' and the 'subjectivist' variant respectively. I think it
is right to say that Aristotle favoured the first. Here my position differs from his and is, I
think, more akin to that of some writers of the utilitarian tradition.
From what has just been said someone may get the impression that this is a treatise on ethics.
It is not. (See Ch. I, sect. 1.) But I think that it contains the germ of an ethics, that a moral
philosophy may become extracted from it. This philosophy will hardly strike one as novel in
its main features. What may be to some extent new is the approach to ethics through a study
of the varieties of goodness. I think that this approach is worth being pursued with much more
thoroughness than I have been capable of. I hope others would find it inviting to work out in
greater detail things, which are here presented in the form of a first sketch.
GEORG HENRIK VON WRIGHT
-vi-
CONTENTS
PREFACE page v
1. The idea of the conceptual autonomy of morals -- a Kantian
tradition in ethics. A philosophic understanding of morality must be
based on a comprehensive study of the good in all its varieties 1
2. The idea of a sharp distinction between is and ought and between
fact and vatue -- a Humean tradition in ethics. Normative ethics and
meta-ethics. Doubts as to whether the two can be sharply
distinguished 2
3. Our inquiry is conceptual. Remarks on the nature of conceptual
investigations. Moral words in search of a meaning. The moral
philosopher as a moulder of concepts. The importance of ethics to
our orientation in the world as moral agents 4
4. Division of ethically relevant concepts into three main groups, viz.
value-concepts, normative concepts, and anthropological
(psychological) concepts. Concepts between the groups. The
narrow and the broad approach to ethics. The broad approach and
the idea of a Philosophical Anthropology. The broad approach and
a General Theory of Norms and Values 6
5. The Varieties of Goodness. Illustration of the multiplicity of uses of
the word 'good' by means of examples. Distinction of some
principal forms of goodness 8
6. The forms of goodness are not species of a generic good. Note on
the concept of form 12
7. The multiform nature of goodness is not due to an ambiguity of the
word or a vagueness of the concept. The variety of forms of the 13
good not a variety of analogical meanings. Is goodness a family-
concept? The meaningpattern of 'good' as a problem for
philosophical semantics
8. Affinities between the forms of goodness. So-called moral
goodness not an independent form of the good 17
1. Preliminary explanation of instrumental and technical goodness.
The instrumentally or technically good thing is often, but not
necessarily, good of its kind 19
-vii-
2. Instrumental goodness is primarily goodness for a purpose.
Instrumental goodness of its kind presupposes an essential
connexion between kind and purpose. Functional and
morphological characteristics of kinds. The question of unity of the
kind 20
3. In the realm of instrumental goodness the opposite of 'good' is
'poor'. Difference between 'poor' and 'bad'. There is no instrumental
badness 22
4. 'Good' and 'poor' connote contradictories rather than contraries.
Poorness a privation 23
5. Judgments of instrumental goodness or betterness are objectively
true or false. The subjective setting of any such judgment. The
notion of a good-making property. Vagueness and instrumental
goodness 24
6. Sentences expressing judgments of instrumental goodness have a
descriptive content or sense. Why they should not be called
'descriptive sentences'. Distinction between the sense of a sentence
and its use 30
7. Instrumental goodness and commending. Why does goodness
appeal? 30
8. Instrumental goodness and preferential choice. Distinction between
the real and the apparent good. A man necessarily prefers, with a
view to a given purpose, the thing which he judges better for that
purpose 31
9. Technical goodness is primarily goodness at something. Technical
goodness of its kind presupposes an essential connexion between
kind and activity. Technical goodness is acquired and not innate 32
10. In the realm of technical goodness the opposite of 'good' is called
'poor' or 'bad'. The opposition is between contradictories rather than
contraries. Technical badness is a privative notion 34
11. Tests of technical goodness. Competition and achievement tests.
Tests by symptoms and by criteria. The technical goodness of
professionals is secondary to instrumental goodness. Note on the
creative arts. Goodness and greatness 35
12. Technical goodness, commending, and praising. Technical
goodness and keenness on activity. A man necessarily wants to
practise an art on which he is keen as well as possible 39
1.
-viii-
a sub-category of the useful. The beneficial presupposes the notion
of the good of a being 41
3. Relations between instrumental and utilitarian goodness.
Instrumental goodness as a degree of usefulness 43
4. The opposites of the useful and the beneficial. The broad and the
narrow sense of 'harmful'. Two senses of 'evil' 45
5. Remarks on the logic of causal efficacy in the sphere of utilitarian
goodness. The meanings of 'favourable' and 'adverse' 47
6. Are judgments of utilitarian goodness objective? The axiological
and the causal components of judgments of the beneficial and the
harmful 48
7. Which kind of being has a good? Good and life 50
8. Medical goodness -- the goodness of organs and faculties. Relations
to instrumental and technical goodness. Essential functions of a
being. The concept of normalcy 51
9. The tripartite division well-weak-ill and the bipartite division good-
bad. Illness as basic notion. Weakness as potential illness. 'Good'
means 'all right'. Note on aitia 54
10. Pain and frustration logically constitutive of badness of organs.
Badness and sub-normal performance 56
11. Goodness of faculties. The social aspect of mental illness 58
12. 11. Are judgments of medical goodness objective? 60
13. 12. Health, illness, and the good of a being. The ethical significance
of medical analogies (Plato) 61
1. Pleasure insufficiently discussed in literature. Distinction between
passive and active pleasure and the pleasure of satisfaction 63
2. Passive pleasure. The good-tasting apple as example. Primary and
secondary hedonic judgments 65
3. Criticism of the view that pleasantness is a sensible quality 67
4. Pleasure and its contraries. The concept of pain 69
5. Analysis of secondary hedonic judgments. An analogy to the
emotive theory in ethics. Distinction between third and first person
hedonic judgments. The first person judgments express valuations
and lack truth-value. The third person judgments are about
valuations. They are true or false -- but no value-judgments 71
6. The logical form of primary hedonic judgments. How is mention of
a valuating subject to be worked into the overt formulation of the
value-judgment? The liking-relation 75
7. Active pleasure. Liking to do and wanting to do 77
8. Discussion of Psychological Hedonism. The doctrine misinterprets
the necessary connexion between pleasure and satisfaction of desire
as being a necessary connexion between desire and pleasure as its
object 79
-ix-
9. Can a man desire the unpleasant? 84
10. Remarks on Ethical Hedonism. Pleasure is not the sole good, but all 84
forms of the good may have an intrinsic relationship to pleasure
1. Welfare the good of man. Distinction between welfare and
happiness. Happiness the consummation of welfare 86
2. Happiness, welfare, and ends of action. Discussion of the position
of Aristotle. Refutation of Psychological Eudaimonism. A man can
pursue his own happiness as ultimate end, and the happiness of
others as an intermediate or an ultimate end of his action. Welfare
only 'obliquely' an end of action. Welfare as ultimate end and
beings as 'ends in themselves' 88
3. Ideals of happiness. Happiness and passive pleasure. Criticism of
Epicurean ideals. Happiness and contentedness. Criticism of ascetic
ideals. Happiness and active pleasure 92
4. The conditions of happiness. Happiness as conditioned by luck,
internal disposition, and action 94
5. The mutability and permanence of happiness. The analogical pairs:
pleasant-unpleasant, glad-sad, happy-unhappy 96
6. Analogy between eudaimonic and hedonic judgments. To be happy
is to like one's circumstances of life 97
7. First person judgments of happiness express valuations, third
person judgments are true of false statements about the way men
value their circumstances of life. Insincere first person judgments.
Ultimately the subject is judge in his own case 99
8. The causal component involved in judgments of welfare.
Consequences and causal prerequisites of changes, which affect the
good of man 101
9. Things wanted and unwanted in themselves -- an analogue to the
concept of intrinsic value 103
10. Ends of action and things wanted in themselves both fall under the
category 'goods'. How they are mutually related 104
11. A good and its price. The constituents of a man's good determined.
Things beneficial and harmful defined as the nuclei of the positive
and negative constituents of a man's good. The concept of need 105
12. Distinction between the apparent and the real good of a man.
Judgments of welfare depend upon knowledge of causal
connexions. The limitations of man's capacity of judging correctly
in matters relating to his welfare 108
13. The problem of the 'objectivity' of judgments concerning that which
is good or bad for a man 110
-x-
1. The notion of regret. Repeatable and non-repeatable choices, which
are relevant to the good of man. The choice of a life. Note on
akrasia 112
1. Technical goodness and 'good' as an attribute of actindividuals 114
2. Instrumental goodness and acts. The notion of a 'way of doing'
something. The conncxion between way of doing and thing done is
intrinsic 115
3. Utilitarian goodness and acts. The connexion between an act as
means and an achievement as end is extrinsic 115
4. The beneficial and the harmful as attributes of acts. The notion of
doing good (bad, evil, harm) to some being 117
5. The moral goodness and badness of acts is not an independent form
of the good, but has to be accounted for in terms of the beneficial
and the harmful. A sense in which moral goodness (badness) is
'absolute' and 'objective' -the beneficial (harmful) again is 'relative'
and 'subjective' 119
6. On the possibility of judging of the overall beneficial or harmful
nature of an act from the point of view of the good (welfare) of a
community of men 120
7. Criticism of a suggested definition of moral goodness and badness
in terms of the beneficial and harmful nature of action. The moral
quality of an act essentially depends upon the agent's intention in
acting and his foreseeing of good and harm to others 121
8. Intention. The relation between the intention in acting and the
foreseeing of consequences. Foreseen consequences of action are
not necessarily intended results of action. Distinction between the
intentional and the not unintentional 123
9. 'Good' as an attribute of intentions. 'Good intention' and 'intention
to do good'. The notion of the intended good (bad). Distinction
between the factual and the axiological object of intention. The
utilitarian value of good intcntions for the promotion of good 125
10. A suggested definition of morally good and bad intention in acting
and of morally good and bad acts 128
11. The concept of unavoidable bad. Discussion of the conditions,
under which an act, from which some bad is foreseen to follow, is
not a morally bad act 130
12. The good man. Instrumental, technical, and utilitarian goodness of
men. Benevolence and malevolence as attributes of character. The
good man and the virtuous man 133
1. Virtue -- a neglected topic in modern ethics 136
-xi-
2. Note on the words arete,virtus, and 'virtue'. The logical
inhomogeneity of the concept. Distinction between the concept of
virtue and the concept of a virtue. Here we are interested only in the
second 137
3. A virtue is neither an innate nor an acquired skill in any particular
activity. Comparison between a virtue and a technical excellence or
the goodness of a faculty 139
4. Acts in accordance with a virtue do not form an actcategory. The
genus of a virtue is neither that of disposition nor that of habit 141
5. The genus of virtues is state of character. Comment on Aristotle's
division into moral and intellectual virtues. Note on the concept of
character 143
6. Virtue is concerned with the choice of a right course of action in a
particular situation, when the good of some being is at stake 145
7. Virtuous choice the outcome of a contest between 'reason' and
'passion'. The virtuous man has learnt to conquer the obscuring 146
Instead of saying that something is good for a being or thing, we often also say that it does
good to the being or thing in question. To take a holiday or to get married will do him good,
we say.
That which, in the sense of the above examples, is good for or does good to some being or
thing, I shall call beneficial.
Related to the category of the beneficial is the category of the useful and its sister-categories
the advantageous and the favourable. For the last three we shall coin the term utilitarian
goodness. I shall, however, also frequently employ 'useful' as a common name of the three
sister-categories and 'usefulness' as a synonym for 'utilitarian goodness'.
There are typical uses of the phrase 'good for somebody', under
-41-
which we would not translate 'good' by 'beneficial' but by 'useful'. For example: to know the
language of the country in which he is travelling, is a good thing for the tourist -- knowing the
language he will get more out of the journey. Such knowledge is useful under the
circumstances. But we do not normally call it beneficial.
When 'good' in 'be good for somebody' means 'useful', then the whole phrase cannot -- as
when 'good' means 'beneficial' -- be replaced by 'do good to somebody'. But of that which is
useful, we often say that it is 'good to be' or 'good to have'. For example: to be courageous or
to have courage is a good thing for a man, it is useful in situations, when he is facing danger.
Many other states of character, traditionally called virtues, such as temperance or industry, are
useful too. But virtues are not normally called beneficial.
Which is the logical relation between the category of the useful (and its sister-categories) on
the one hand and the category of the beneficial on the other hand? I think it is right to say that
the second is a sub-category of the first, or that beneficiality is a sub-form of utilitarian
goodness. Everything beneficial is also useful, but not everything useful is also beneficial.
Things which are useful without also being beneficial I shall call 'merely useful'.
On the view which is here taken of the relation between beneficiality and utilitarian goodness,
the two must have some feature in common. The beneficial, in addition, must possess some
distinguishing feature, which marks the beneficial as that which is in some special sense
useful, advantageous, or favourable.
The common feature seems to be this: Something is useful or beneficial by virtue of the way
in which it causally affects something else. I shall call this causal relevance of the useful or
beneficial a way of affecting favourably the thing in question. The 'logical mechanism' of
affecting things favourably will be briefly examined in section 4. This form of causal
relevance will be seen to have two main branches. I shall refer to them using the words
promote and protect respectively.
In the case of useful things, which are also called beneficial, that to which the useful is
favourably causally relevant -- ultimately if not immediately -- may be characterized as the
good of some being. Instead of 'good' we can here also say 'welfare'. Physical exercise, for
example, is beneficial, because good for the health. 'Health' is another name for the good
(welfare) of the body. The good of the
-42-
human body is an aspect of the good (welfare) of man. Thus to say that exercise is beneficial
is to say that it affects favourably, immediately the good of the body, and ultimately the good
of man. Generally speaking: everything which is beneficial affects favourably the good
(welfare) of some being.
It appears to be the case that a thing which affects the good of some being protectively rather
than promotively, we call (merely) useful rather than (also) beneficial. The reason, for
example, why we call the causal relevance of virtues to the good of human beings usefulness
rather than beneficiality, seems to be that virtues are protective rather than promotive of this
good.
Things are thus sometimes useful without also being beneficial on the ground that they affect
favourably the good of some being. In most cases, however, things which are merely useful
are this on the ground that they are favourably causally relevant to (the attainment of) some
end of action. It is convenient, we have said, to call anything which is an end of action a good.
Accepting this terminology, one could also say that a thing is useful when it is favourably
causally relevant to some good; and that it is beneficial when it is promotive of that peculiar
good which we call the good of a being.
2. Things which are useful for the attainment of ends are often also said to be instrumental to
the attainment of those ends.
In view of this common use of the word 'instrumental', it may be asked whether we ought not
to have reserved the term 'instrumental goodness' for the category of the useful. This, I think,
would correspond to the way in which many philosophers actually have used the term.
The dispute about terminology is futile. It is important, however, to realize that 'good' in 'a
good knife' or 'a good watch' or 'a good car' does not ordinarily mean the same as 'useful'.
Even a poor knife can, under circumstances, be useful. It is useful, whenever the use of this
knife is a good thing. But this usefulness of the knife does not necessarily mean that it is a
good knife.
The distinction between usefulness and that which I have called instrumental goodness can be
illuminated by considering the difference in meaning between the phrases 'be good for a
purpose' and 'serve a purpose well'. To say of something that it is good for a purpose
ordinarily means that it can be used to serve this purpose. If we are in pursuit of the purpose
in question, then this thing is
-43-
useful, a good thing to have. But to say of something that it is a good so-and-so, e.g. knife, is
to presuppose that it can be used for a purpose essentially associated with so-and-sos, and to
say that it serves this purpose well. (See Ch. II, sect. 2.) The opposite (contradictory) of 'good
for this purpose' is 'no good for this purpose', i.e. 'cannot be used to serve this purpose' or
'useless for this purpose'. The opposite of 'serves this purpose well' is 'serves this purpose
poorly'. Nothing can serve a purpose even poorly unless it can serve it, i.e. unless it is in a
sense good for this purpose.
Instrumental goodness is typically an excellence or a notion of rank and grade, whereas
usefulness is not. This is reflected in the different rle, which absolute and comparative
judgments of value play in connexion with instrumental goodness on the one hand and in
connexion with the useful or the beneficial on the other hand. We are more often interested in
knowing whether one knife is, within a given subjective setting of purpose, better than
another, than in knowing whether it is, in that setting, good simpliciter. This is so because of
the intimate connexion which there is between judgments of instrumental goodness and
preferential choice. When judging of usefulness again we are primarily interested in the
question whether something can be used or is no good for a certain purpose, and in judging of
beneficiality whether something will or will not do a being good -- for example whether it
will be good for him to go into business. This is not to say that we do not also sometimes
grade things as being more or less useful or beneficial. But whereas even the less good of two
medicines, which both work to cure an illness, does the patient good and in this sense is
beneficial, useful for restoring his health, not even the better of two knives, which can both be
used for carving meat, needs be a good carving knife.
To call something an instrumentally good K is to say that it is a K which is good as such, i.e.
as a K. When'good'in a phrase 'a good K' means useful or beneficial, then, for all I can see, we
are not attributing a goodness of its kind to the K. A good habit, for example, is not good as
habit. Habits have no special excellence of their kind, as knives or watches or cars have. (Cf.
Ch. II, sect. 1.) The logical picture, however, is complicated here by the fact that many kinds,
the members of which are judged useful or beneficial in some respect, are also essentially
associated with purposes, as e.g. medicines with purposes of curing illness. When this is the
case,
-44-
the useful or beneficial thing can also be judged from the point of view of its instrumental
goodness. Then the instrumental goodness of the thing can be said to measure its degree of
usefulness.
3. We must now consider the opposites of the useful and its sub-category the beneficial.
One opposite of the useful is the useless. To call something useless can mean several things. It
can mean that the thing in question has no use for any purpose whatsoever, or that it cannot be
used for a certain purpose. When 'useful' means 'good for this purpose', i.e. 'can be used for
this purpose', and 'useless' means 'no good for this purpose', i.e. 'cannot be used for this
purpose', then the useful and the useless are opposed to one another as contradictories, and not
as contraries. This seems to be the way the two are normally opposed.
If one had to name the opposite of the beneficial, the first name to suggest itself would
probably be the adjective 'harmful'. Others which would also come to the mind are
'detrimental', 'damaging', and 'injurious' or 'injuring'.
The beneficial, we said, is something which is good for or does good to a thing or a being, i.e.
serves the good of the thing or being in question. Clearly, no thing is harmful merely on the
ground that it does no good or does not serve the good of any being. 'Harmful', unlike
'useless', is not a privative term. The harmful which is the opposite of the beneficial is that
which affects the good of a being unfavourably, adversely. And to affect the good of a being
adversely is not the same as not to affect it favourably. If it were, then to be harmless (=not
harmful) would be the same as beneficial, which it obviously is not.
Good habits, for example, have beneficial effects on a man, good laws and institutions
beneficial effects on a community of country. Bad habits have detrimental effects on the good
of a man, e.g. because they ruin his bodily health. Similarly bad laws and institutions do harm
to the community. But habits can also be harmless without being good, and institutions and
laws useless or 'pointless' without being harmful, positively obnoxious. Thus whereas the
useful and the useless, and the harmful and the harmless are opposed to one another as
contradictories, the beneficial and the harmful are opposed to one another as contraries. They
exclude one another, but between them there is a neutral zone.
-45-
The harmful, when it is the contrary of the beneficial, is also said to be bad, or sometimes
evil, for the being whose good it affects unfavourably. It is further said to do bad of evil to
this being. It is called a bad thing of maybe even more frequently an evil thing. Often it is
called simply an evil. The institution of ostracism, one could say, was an evil to the Athenian
republic; the reign of Ahab was an evil to Israel, that of Hitler to Germany and Europe. I find
it convenient to adopt the substantive form an evil for anything -- be it an institution, an act, a
state of affairs or of character -- which is harmful in that sense of 'harmful', which is the
opposite of beneficial, i.e. for that which is bad for the good of a being.
The notion of evil, thus defined, is a sub-category of the notion of the harmful. Anything
which frustrates or hampers the attainment of some end of human action is harmful or
detrimental, viz. to the attainment of that end. But not of everything, which is in this sense
harmful, do we say that it does evil to the being in pursuit of the end. Sometimes we say that
it was good for a man that this of that plan of his became ruined -- but that which ruined his
plans was nevertheless a harmful thing, viz. to his plans. This observation shows that the
adjective 'harmful' is commonly used also to mean the opposite of utilitarian goodness in
general, and not only to mean the opposite of its sub-category the beneficial.
There are thus two senses of 'harmful' which must be distinguished: 'harmful' in the broad
sense, which names the contrary of utilitarian goodness, and 'harmful' in the narrower sense,
which names the contrary of the beneficial.
There are also two senses of the word 'evil' to be noted in this connexion. When 'an evil'
means something which affects the good of a being adversely, then 'an evil' means something
which is a cause of harm (to the being involved). But instead of 'a cause of harm' we can here
also say 'a cause of evil'. Thus the word 'evil' sometimes means the cause of harm and
sometimes the harm caused. The word 'harm', it would seem, has not the same double
meaning. It nearly always means the thing (damage) caused or suffered.
The term 'poor', be it noted in conclusion, is also sometimes used in connexion with assessing
utilitarian goodness. Then 'poor' does not mean the same as that which is the ordinary
meaning of 'bad' in such contexts. The poor thing, in a utilitarian sense, is to
-46-
some low degree favourably relevant to some end or purpose. It is thus not entirely useless,
still less is it harmful, for the end or purpose in question. The poor is some, but very little, use.
A poor medicine, for example, still does some good. If it affects the patient adversely, it is not
called poor. But it may be called bad in the utilitarian sense of harmful or obnoxious.
4. The property of being useful or harmful, and a fortiori also the property of being beneficial
or evil, belongs to a thing by virtue of a causal relationship between this thing and something
else (cf. sect. 1). In the case of things useful or beneficial this causal relationship is one of
affecting favourably or being favourably relevant to the other thing in question. In the case of
things harmful or evil this relationship is one of affecting adversely or unfavourably this other
thing.
There are two principal ways in which something can be causally favourable to the attainment
of an end. Either this thing is favourably relevant to the end by taking us, metaphorically
speaking, nearer or even up to this end. Or it is favourably relevant by preventing us,
metaphorically speaking, from being taken farther away from the end. We have already (p. 42
) coined the terms promotive and protective for these two forms of favourable causal
relevance to ends.
In a similar manner one may distinguish between two main forms of affecting the attainment
of ends unfavourably. Something is unfavourably relevant by taking us farther away from the
end or goal, or by preventing us from getting nearer to it. When the adverse effect is of the
first kind, it is also called deteriorative.
The promotive effect of some useful or beneficial thing is sometimes described by saying that
it makes bad better and sometimes by saying that it makes good better. The former type of
promotive effect is often also called an effect of curing or healing.
Take, for example, the case of a beneficial medicine or cure. It serves the good of the being to
whom it is administered, by curing some illness and thus restoring health. An illness is an
evil, something which affects the good of the body -- ultimately the good of a man --
adversely. A beneficial medicine or cure thus works for the good of a being by working
against an existing evil. Generally speaking: it promotes the good by making bad better.
-47-
In a similar manner, the deteriorative effect of some harmful or evil thing is sometimes
described by saying that it makes good worse and sometimes by saying that it makes bad
worse.
When, for example, the wrong medicine is administered to the ill man making him sicker still,
the effect of the harmful consists in making bad worse. When excessive use of tobacco or of
alcohol ruins the health of a person, the harmful makes good deteriorate. Bad habits, generally
speaking, promote evil by ruining the good of the being who practises them.
It would be an interesting and worthwhile task to investigate the formal logic of these various
forms of the causal efficacy of utilitarian goodness and its opposite. The investigation would
show, for one thing, in how many different senses something can be a 'cause' of good or of
evil. To observe these different senses is essential to any ethics, which measures the moral
value or rightness of acts in terms of the consequences of action. Yet it is an observation
which traditional ethics has habitually neglected to make.
5. Are value-judgments, of the kind which we are now considering, objectively true or false?
In discussing this question it is important to separate judgments which involve the notion of
the good of a being, from judgments which do not involve this notion. The case of the latter is
much simpler.
In judgments of usefulness and harmfulness, which do not involve the notion of the good of a
being, we judge the causal relevance of something for some purpose or end which we pursue.
The existence of the purpose or end is presupposed in the judgment. An end, we said, can also
be called 'a good' simply by virtue of being desired. The question, however, whether an end is
good or bad in some other sense, i.e. the question of the value of ends, is totally irrelevant to
judgments of usefulness and its opposites. Such judgments are purely causal.
For example: A man wants to train himself to become a good runner and deliberates whether
he should give up smoking. Is the habit of smoking detrimental, harmful, obnoxious to the
attainment of his end? He has to consider the causal effects of smoking upon excellence in
running. The problem can be viewed under the various aspects of causal relevance, which we
distinguished in the preceding section: does smoking prevent him from improving his talents,
or does it positively ruin them? Which aspect of the causal
-48-
problem that will interest him most, will largely depend upon the state which he has already
reached with a view to the end.
Suppose our man arrives at the conclusion that smoking is a bad habit in the sense that it has
adverse effects on the attainment of his end. The conclusion is true or false -- and in this sense
'objective'. Assume that it is true. Does it then follow that smoking is a bad habit with every
man who pursues the same end as he?
Obviously, identity of aim or end or purpose is not enough to make the judgment concerning
the badness of smoking generalizable. Another man may pursue the same end, but have a
different constitution or otherwise be differently circumstanced and therefor 'immune' to the
harmful effects of tobacco. What is a bad habit for one man in pursuit of a certain end, need
not be a bad habit for another man in pursuit of exactly the same end. With a view to this one
may call the value-judgment passed on the habit 'subjective'. But this subjectivity -- if it be
called by that name -- does not remove the judgment from the province of truth and
falsehood.
Suppose we generalized the case which we have been discussing, and said that smoking is bad
for any man who pursues the same end and is exactly similarly circumstanced. This would be
trivial, unless we specified the circumstances, in which case the generalization might easily be
false.
If we want to generalize in matters of usefulness and harmfulness, we shall on the whole have
to be content with rough generalizations of the following schematic type: For most men in
circumstances C the thing X is good or bad with a view to the end E. The corresponding holds
true also for the special form of the useful, which we call the beneficial, and the special form
of the harmful, which we call the evil.
Judgments of utilitarian goodness, which do not involve the notion of the good of a being, are,
we said, purely causal judgments, though of limited generalizability. Judgments of the
beneficial and the evil, however, are not purely causal. They split in two components. One is
causal. It concerns the consequences or effects of certain acts or habits or practices or
institutions or what not. The other component could conveniently be called axiological. It
concerns the relations of these effects to that which we have termed the good of a being. This
is not a causal relation. It is more like a relationship of belonging. We shall later have
occasion to investigate the nature of this relationship in more detail.
-49-
6. What kinds or species of being have a good? What is the range of significance of 'X' in the
phrase 'the good of X'?
Can, for example, artefacts and other inanimate beings have a good? It is not unnatural to say
that lubrication is beneficial or good for the car, or that violent shocks will do harm to a
watch. The goodness of a car or a watch is itself instrumental goodness for some human
purposes. Therefore that which is good for the car or watch is something which will keep it fit
or in good order with a view to its serving a purpose well. It may be argued that, since the
goodness of the car or watch is relative to human ends and purposes, that which is good for
the car or watch cannot properly be said to serve the good of the car or watch themselves. If it
serves anybody's good at all, it will be the good of the human being to whom the instrument
belongs or who uses it.
A being, of whose good it is meaningful to talk, is one who can meaningfully be said to be
well or ill, to thrive, to flourish, be happy or miserable. These things, no doubt, are sometimes
said of artefacts and inanimate objects too -- particularly when we feel a strong attachment to
them. 'My car does not like the roads of this district very much, as shown by the frequent
overhaulings which it needs', we may say. But this is clearly a metaphorical way of speaking.
The attributes, which go along with meaningful use of the phrase 'the good of X', may be
called biological in a broad sense. By this I do not mean that they were terms, of which
biologists make frequent use. 'Happiness' and 'welfare' cannot be said to belong to the
professional vocabulary of biologists. What I mean by calling the terms 'biological' is that
they are used as attributes of beings, of whom it is meaningful to say that they have a life. The
question 'What kinds or species of being have a good?' is therefore broadly identical with the
question 'What kind or species of being have a life?' And one could say that it is metaphorical
to speak of the good of a being, to the same extent as it is metaphorical to speak of the life of
that being.
Artefacts, such as cars and watches, have a life and therefore a good, only metaphorically.
Plants and animals have a life in the primary sense. But what shall we say of social units such
as the family, the nation, the state? Have they got a life 'literally' or 'metaphorically' only? I
shall not attempt to answer these questions. I doubt whether there is any other way of
answering them except
-50-
by pointing out existing analogies of language. It is a fact that we speak about the life and also
the good (welfare) of the family, the nation, and the state. This fact about the use of language
we must accept and with it the idea that the social units in question have a life and a good.
What is arguable, however, is whether the life and a fortiori also the good (welfare) of a
social unit is not somehow 'logically reducible' to the life and therefore the good of the beings
-- men or animals -- who are its members.
It would seem that man, among beings who have a good, holds a position, which is peculiar in
two respects.
Man is not the only living being who has a social life. But there are a vast number of social
units, which are peculiar to man. These are the units which presuppose a normative order.
The supreme example of such units is the state. How the good of such 'covenanted', as they
may also be called, social units is conceptually related to the good of the human individual, is
a major problem of political philosophy.
The other respect in which man holds a peculiar position among beings who have a good, is
that he may be regarded as quasi composed of parts, of whose independent life and good one
can speak. Man is body and mind, one sometimes says. Or one makes a tripartite division of
man into body, mind, and soul. It makes sense to speak of the life of the body and also of the
welfare of the body. The same holds true for the mind and the soul. The welfare of the body
and mind is called (bodily and mental) health.
In the rest of this chapter I shall discuss a form of goodness which is connected with that
aspect of the good of man, which we call health. We may call this form medical goodness.
7. Goodness of the form here called medical is referred to when we speak about a (in the
medical sense) good heart, good lungs, or good eyes -- but also about a good memory or
understanding.
The heart and the lungs are organs of the body; memory and understanding are often called
faculties of the mind. That which is here called medical goodness could also be spoken of as
the goodness of organs and faculties.
Are the eyes an organ of the body or of the mind? Perhaps we could say that they are an organ
of the body, which serves as the bodily substratum of a faculty of the mind, viz. sight. The
same holds good of the other so-called sense-organs.
-51-
Organs resemble instruments or tools in that they have both morphological and functional
characteristics. Faculties again have no morphological features proper to them; in this and
other respects they resemble abilities and skills. The functional characteristics of organs, too,
resemble abilities; they consist in things which the organs themselves do, such as pumping
blood or breathing air, rather than -- as is the case with tools -- in their usability for various
assigned purposes.
In speaking of the goodness of various organs we are primarily concerned with their
characteristic functions and not with their morphological features. A deformation of the heart
may be a cause why it performs badly and is thus called a bad heart. But if the deformation
did not impede the performance of the organ, we should hardly call it bad merely because
deformed.
When 'good' in the phrase 'a good K' refers to medical goodness and K thus is some organ or
faculty, we have another case of a goodness of its kind. In this respect medical goodness
resembles instrumental and technical goodness, but differs from utilitarian goodness. (Cf.
sect. 2.)
Instrumental goodness of its kind presupposes, we found, an essential connexion between the
kind and some purpose; technical goodness again an essential tie between the kind and some
activity. In a similar manner, medical goodness of its kind can be said to presuppose the
existence of an essential connexion between the kind (of organ or faculty) and some function.
There is a conspicuous resemblance between medical and technical goodness by virtue of the
fact that both forms of goodness manifest themselves as a certain excellence of performance.
But there are noteworthy differences as well. I shall here mention the following three:
The technically good man is good at something. But the phrase 'good at' is seldom used in
connexion with good organs or faculties. A good heart is not ordinarily said to be good at
pumping blood or good lungs to be good at breathing. Good eyes are not said to be good at
seeing, nor even the man who has good eyes. But the man of good sight may be good at, say,
discerning landmarks at a great distance. The man of good memory may be good at, say,
remembering dates or telephone numbers or knowing poems by heart. In these last cases the
things at which the man is good is some special activity in which one can train
-52-
oneself, and not the function, as such, of the organ or faculty concerned.
This observation takes us to a second difference between technical goodness and the goodness
of faculties and organs. Activities at which men are said to be good, are for the most part
acquired rather than innate. Teaching and training are normally needed to reach technical
excellence. The functions of organs and faculties again are substantially innate. By training
one's body and mind one can, to some extent, develop those functions to greater perfection,
and by caring for one's body and mind one can, to some extent, prevent them from decaying
and deteriorating. But substantially the goodness of faculties and organs is not dependent
upon what a man does with a view to perfecting them, but upon what might be called the
graces of nature and fate.
The term 'innate', incidentally, must not be misunderstood. It does not mean that all faculties
and organs perform their proper functions from the birth of the individual. It may take time
for them to mature. Man can quite properly be said to be innately endowed with sight, and yet
a newly born baby cannot see.
A third major difference between technical goodness and the goodness of organs and faculties
is constituted by the way in which organs and faculties serve the good of the being who has
them. A man can use his acquired skills and special talents to promote his good. He need not,
however, do this. But good faculties and organs he needs. To have a bad heart or bad lungs or
a bad memory is bad for -- an evil to -- the man who has it, and it is because of the
detrimental effects on the good of the whole that we judge the organs or faculties in question
bad. (See below, sect. 9.)
Because of the intrinsic connexion which holds between the goodness of organs and faculties
and the good of the being to whom they belong, I propose to call the functions, which are
proper to the various organs and faculties, essential functions of the being, or rather, of the
kind or species of which the individual being is a member. The 'essentiality' of the functions
does not entail that every individual of the species can actually perform all those functions.
But it entails that, if an individual cannot do this 'at the time when by nature it should' -- to
quote Aristotle -- we call it abnormal or defect or faulty or, sometimes, injured. It follows by
contraposition that the essential functions of the species are functions which any normal
individual of the species can perform.
-53-
The essential functions are needed for that which could conveniently be called a normal life
of the individual.
The notion of normalcy, as we shall soon see, is of great importance to the understanding of
the special form of goodness which belongs to organs and faculties.
8. Medical goodness may be said to be related to the notion of the good of a being thanks to
the intermediary rle of the notions of health and illness.
An organ which performs its proper function well is said to be good or well. It is also often
said to be healthy and sometimes said to be in good health. This last, however, is more
commonly said of the man, or being, whose organs are concerned.
An organ which does not perform its proper function well is sometimes called ill and
sometimes weak. It is also called bad or poor. One can distinguish between the meanings of
'ill' and 'weak'. 'Bad' and 'poor' again are used pretty much as synonyms in the field of medical
goodness; perhaps one could maintain that 'bad' is more often used for the ill than for the
weak, and 'poor' more often for the weak than for the ill organ.
That a man has bad health usually means that he suffers an illness of some or several of his
organs. Bad health is also called poor health. If a man is said to be of weak health, this
weakness of health must be reflected in the weakness of some or several of his bodily parts.
Weak health too is sometimes called poor health.
What is the difference between illness and weakness of organs, or between bodily illness and
weakness of health? Ordinary usage can hardly be said to maintain a rigorous distinction
between the concepts. But it may be said to hint at a distinction which can be made and
maintained with a certain rigour.
Illness, we shall say, is an actual and weakness a potential evil or cause of harm to the being
concerned. The ill or diseased organ causes harm or suffering to the being whose organ it is.
The weak organ may cause suffering. The meaning of 'may' here is not that of mere physical
possibility. The 'may' has to be explained in terms of probability. Roughly speaking:
weakness as a potential cause of harm to the body is a probability of illness. How this
probability is estimated need not concern us here. Weakness could also be called a disposition
of tendency to deteriorate into illness. This holds good both of weakness of organs and of
weakness of health.
-54-
Poor or bad organs thus function in a way which affects health adversely, by being either an
actual or potential cause of illness. Shall we say contrarywise that good organs function in a
way which affects health favourably, i.e. promotes the physical wellbeing of the being or at
least prevents it from deteriorating? It seems to me that it would not be correct to say thus.
For, let us ask, what is ordinarily meant by 'good eyes', 'good lungs', etc. Primarily, it seems,
organs which are not bad (poor), i.e. neither ill nor weak. 'Good' as an attribute of an organ of
the body means very much the same as 'all right'. Organs are good when they are in order, fit,
as they should be, normal. Organs which are better than normal are called 'exceptionally
good'.
Thus in the case of organs, badness (poorness) appears to be logically primary to goodness.
'Good' is here a privative term. It means 'not bad', 'all right', 'no source of complaint'. If I am
not mistaken, this is a logical feature in which the form of goodness, which we are now
discussing, differs basically from both instrumental and technical goodness. In their case,
'poor' (for technical goodness also 'bad') is a privative term, 'good' a positive term. The
positive is logically primary to the privative.
It follows from the definition of medical goodness as an absence of weakness and illness that
'good' and 'poor' as attributes of organs or of bodily health denote contradictories rather than
contraries. But from the definition of weakness as likelihood of illness it also follows that the
border separating the good from the poor will be vague. (The meaning of 'likely' is inherently
vague.) Because of this vagueness it will frequently be impossible to pronounce definitely on
the question whether a certain organ of some individual being should be considered good or
not.
If a being is said to be in good bodily health, when its organs perform their proper functions
well, and if organs are good, when there is nothing wrong with them, i.e. no illness or
weakness, then the notion of (good) bodily health too is a privative notion. From this privative
notion of health one may distinguish a positive notion of health, which is present when the
being positively 'enjoys' good health, feels fit, thrives or flourishes physically. But of the two
notions, the privative seems far more important. It would be correct to say of an individual
that he is in perfectly good bodily health merely on the ground that there is nothing wrong
with his body and bodily functions.
-55-
The privative notion of goodness of organs and of bodily health has an interesting connexion
with certain ideas relating to causation. If an organ functions unsatisfactorily, i.e. suffers from
some weakness or illness, we think of this bad as having a cause -for example some
constitutional morphological defect of the organ or some injury, which has befallen it in the
course of the life of the individual. But the normal state of the organ, i.e. the state in which it
is when it functions satisfactorily, is not in the same sense 'caused'.
The idea that a cause is primarily a cause of harm and thus is an evil, i.e. a disturbance of an
equilibrium or normal state or good order (a kosmos in the literal sense of the Greek word)
seems to be the very root, from which the idea of causation as we know it to-day has
originated. The Greek word for cause is aitia, which also means guilt, i.e. responsibility for
harm or evil. It is an interesting observation that the Finnish word for cause, syy, has precisely
the same double meaning as the Greek aitia. The received meaning of the term aetiologyis the
study of the causes of diseases, but the literal meaning of the word is science of causes in
general.
9. Organs are bad, i.e. weak or ill, by virtue of their adverse effects on the being whose organs
they are. Briefly speaking: organs are bad because of some bad or evil of which they are the
cause. But, as noted in the preceding section, organs are not good, because of some good they
cause, but simply by not causing harm.
An organ may be judged diseased on the ground that it exhibits a certain deformation. But, as
said before (sect. 7), morphological anomalies can at most be symptoms of badness, not
defining criteria. For the deformations are bad only to the extent that they impede the
function.
The relation between the functioning of an organ and its effects on the body is a causal and
thus extrinsic relation. But the relation between the badness of the effects and the badness of
the organ, whose functioning is responsible for those effects, is a logical and thus intrinsic
relation. The evil which bad organs cause is constitutive of the badness of the organs
themselves, one could also say.
For example: That the discomfort and fatigue, which a man feels each time he has to climb
stairs, should be due to an insufficiency of his heart, is a fact about causation in the human
body. But when the heart, because of this and similar effects, is said to be weak or
-56-
to perform poorly, then the badness of the effects is not 'symptomatic', but 'constitutive' of the
weakness of the organ and poorness of its performance.
Which then is the evil or harm, which bad organs cause, actually or potentially?
One basic form of such harm is pain or pain-like sensations such as discomfort, ache, nausea.
Obviously, the pain caused by a diseased organ need not be continuous, in order that we shall
call the organ (continuously) ill. These different relationships to change reflect differences in
the logic of the concepts of pain and of illness. For example: a diseased heart need not cause
pain when the individual is at rest, but it may cause discomfort when he moves.
Wherein does the evil of pain lie? To ask thus is not to ask a triviality. Pain is evil, I would
say, only to the extent that it is disliked or shunned or unwanted. It is a fact that pain is not
always disliked. The phrase 'a pleasant pain' is not a contradiction in terms. We shall have to
speak more about this in connexion with hedonic goodness.
Do all tests for judging, whether the function of an organ is impeded and the organ thus is
weak or ill, depend logically upon the notion of pain?
At first sight there seem to be ways of testing the functioning of organs and therewith their
goodness, which are independent of pain -- and even of harm generally. Consider, for
example, how a doctor tests a man's sight. A man, broadly speaking, is said to have bad or
poor eyes (sight), if he cannot, at the appropriate distances, discern things and movements
which most men can discern. Here the standard of goodness is set by something, which can be
called the normalcy of the function.
It would, however, be wrong to think of the test of normalcy as being exclusively a
performance-test. If the eyes of a man can see exactly those things which normal eyes are
supposed to see at the appropriate distances, but if the use of his eyes to see this causes him
pain, his eyes would not be normal. The same would hold true if he could look at things only
for a short time without his eyes getting fatigued. Thus the tests of normalcy of performance
include, or have to be supplemented by, considerations pertaining to pain or fatigue. (Fatigue
is discomfort and can for present purposes be counted a form of pain.) The fact alone that the
use of an organ is painful may disqualify it in the normalcy-test.
-57-
But how shall we judge of a case when the functioning of an organ falls decidedly short of the
normal performance, but when there is no pain or bodily discomfort of any kind? Such cases
are perhaps rare, but they may occur. Shall we then say that the organ is bad, because its
performance is sub-normal, or shall we say that it is all right, since it causes no discomfort?
At this point it is good to remember that a man uses his organs to satisfy various needs and
wants of his. If some organ of his performs sub-normally, he can be said to suffer
incapacitation. This means: there will then be things which he could do if the organ
performed normally, but which now he cannot do. Should he want to do these things, the fact
that he cannot do them may be a cause of annoyance, frustration, grief, and similar feelings.
These phenomena can, with some caution, be termed 'mental pains'. The man who suffers
them will then complain of the badness of the organ as a source of his mental discomfort.
But suppose he does not want to do things, which he could do, if the organ functioned
normally. He will then have no reason to complain of this organ of his either. Shall we
nevertheless call it bad? One may hesitate about the answer. It seems to me that, if we call the
organ bad, i.e. weak or ill, it is only because of the fact that men normally have the wants and
needs, which this particular man happens to be without, and therefore also normally would
suffer if they had this man's deficiency. A man who does not put some or other of his organs
to their normal use, may be said to be abnormal in a certain respect, or be said not to live a
normal life. This abnormality of the man or of his life could, for a variety of reasons, be
considered a bad thing and even be regarded as a sort of illness. But this badness is clearly
logically independent of that of the organ.
The evil or harm which bad organs cause, and which is logically constitutive of their badness,
is thus of two principal kinds, viz. pain and frustration of wants. Abnormality of performance
too is logically constitutive of badness of organs, but only indirectly, via the notion of
normality of the needs and wants of men. Due to the fact that some men may be lacking in
normal needs and wants, it may happen that an organ is 'objectively' diseased but that its
badness is not 'subjectively' noted.
10. We have so far been talking explicitly about the goodness (and badness) of organs only.
Much the same things can be said
-58-
about the goodness of faculties. But there are some noteworthy differences. I shall here touch
upon the subject very summarily.
In the case of the faculties too we can distinguish between the ill, the weak, and the good.
Calling a faculty bad (or poor) is usually to say that it is weak. A bad memory is essentially a
weak memory. When faculties completely deteriorate, the being whose faculties they are is
often said to be mentally ill or deranged. The faculties themselves are not commonly called
ill. It would also not be generally correct to say that a weakness of a faculty, like that of an
organ, is a likelihood of illness.
The harm caused by poor faculties cannot, for conceptual reasons, be pain in the primary
sense of bodily pain. The harm is here basically suffering due to incapacitation. It is the
annoyance, frustration, and grief which the individual will experience as a consequence of not
being able to satisfy needs and wants which a man, whose faculties function in the normal
way, can satisfy.
As in the case of the organs, the harm which bad faculties may cause to the being is logically
constitutive of their badness. If use of memory were not vital to the satisfaction of the needs
of a normal life, it is not clear why one should call sub-normal capacity of remembering
things 'bad memory'. Perhaps we can imagine a form of life under which that, which we call
bad memory, would be an altogether good thing to have and therefore would be called good
memory. We too sometimes consider it good to forget, though not on a scale which would
make us revise our notion of goodness of memory. Perhaps we can also imagine
circumstances, under which remembering would be a completely useless activity. To be good
at remembering would then be a technical excellence or something like it. People might be
keen on remembering, as on playing a game. Their notion, too, of a good or bad memory
would be different from ours, i.e. their notion would lack a feature which is essential to ours.
The harm caused by illness, mental or bodily, is not only a concern of the sick person himself.
The ill man may not be able to take care of himself. He has to be helped by others. Thus he
easily becomes a burden on his fellow-humans. He can, moreover, also be a menace to the
good of others: by spreading contagion, if he is bodily ill, and by harmful acts if he is
mentally deranged. The
-59-
question may be raised, whether the relationship between illness and the harm caused to
others by the sick is an intrinsic or extrinsic connexion, i.e. whether it is constitutive or not of
the notion of illness of a man -- as distinct from the notion of illness of an organ of faculty by
itself.
It seems to me that in this respect there is a conceptual difference between physical and
mental illness. It is part of our notion of the latter that it upsets the relations of the sick to his
surroundings in a way which, from a social point of view, is unwanted, undesirable. A man
could not be said to be bodily ill unless he, or at least any normal person in his state, suffers
pain or discomfort. But the feeble-minded or maniac, who is perfectly content and happy in
his state, we still consider ill. This we should hardly think of him if his state were not from a
social point of view a bad thing, an evil. We can imagine forms of lunacy -- there may even
be such forms in actual existence -- which lack these obnoxious effects upon social life and
which we would regard as blessed states, which we praise, rather than illness which we
deplore or pity.
11. Are judgments of medical goodness 'objectively' true or false?
A man who judges of some organ of his that it is diseased on the ground that he feels pain, is
not judging that he has pain. What he judges is that the pain he feels is caused by some defect
of the organ. In this he may, of course, be mistaken. The judgment is true or false. The fact
that pain is a state of consciousness of an individual subject does not detract from or void the
'objectivity' of the judgment.
Shall we say that a man, e.g. a doctor, who judges of some organ of another man that it is
diseased on the ground that this man complains of pain, is making two judgments -- one about
the pain of that other man and another about the cause of his pain? I don't think we should say
thus. The doctor, who judges about the state of health of the organ on the basis of signs of
pain, is presupposing or taking for granted the pain. This presupposition he may have reason
to question, say because an examination of the organ does not reveal any functional or
morphological anomaly, which were symptomatic of illness. His doubts he may express in a
judgment (conjecture) to the effect that the other man actually feels no pain. But the judgment
to the effect that an organ of that other man is
-60-
ill, is a judgment solely to the effect that a cause of a given discomfort is to be located to the
organ (including its functioning).
In a similar manner, when we judge organs or faculties bad on the ground that they are a
source of annoyance, frustration, or grief, we are not judging that a certain mental discomfort
or pain occurs, but that it is due to such and such causes. In these judgments, as will be
remembered, certain standards of normalcy are presupposed. These standards may be vague
and they may be difficult to apply -- for example because they cannot be applied uniformly to
all members of a certain zoological species, but must take account of age and sex and training
and, maybe, various external circumstances of life too. But this does not constitute a ground
for saying that judgments of goodness or badness, which employ such standards, were
'subjective'.
12. Organs and faculties, we have said, serve the good of a being. The organs of the body, one
could say without distorting language, serve immediately the good of the body. This good is
also called bodily health. The faculties of the mind serve immediatey the good of the mind,
our mental health. Remotely, organs and faculties serve the good of man.
The concept of health may be considered a model on a smaller scale of the more
comprehensive notion of the good of a being. That is: it may be suggested that one should try
to understand this good (welfare) in all its various aspects on the pattern of the notion of
health. On such a view, the good of man would be a medical notion by analogy, as are the
good of the body and of the mind literally.
The conception of the good of man on the basis of medical analogies is characteristic of the
ethics and political philosophy of Plato. The idea is profound and, I think, basically sound. It
is worth a much more thorough exploration than will be given to it in the present work.
We distinguished above (sect. 8) between a privative and a positive notion of health. The first,
we said, is the more basic notion. Health in the privative sense consists in the absence of
bodily pain and of pain-like states, which are consequent upon the frustration of needs and
wants of a normal life. Health in the positive sense consists in the presence of feelings of
fitness and strength and in similar pleasant (agreeable, joyful) states. In the enjoyment
-61-
of those states the healthy body and mind can be said to flourish.
In a similar manner, welfare may be said to present two aspects. The one, which answers to
the privative notion of health, is the basic aspect. It is conceptually allied to the needs and
wants of beings and to the notions of the beneficial and harmful. I am not, however, going to
suggest that it is a privative idea in the same strong sense in which the basic notion of health
seems to be this. The other aspect of welfare, which answers to the positive notion of health,
has a primary conceptual alliance with pleasure. Of the being, who enjoys this aspect of its
welfare, we say that it is happy. Happiness could also be called the flower of welfare.
-62-
IV
THE HEDONIC GOOD
1. OUR discussion in Chapter III brought us into touch with the notion of the good of a being.
This in its turn was found to be related to the notions of pain and pleasure, i.e. to a further
form of goodness, which is here called the hedonic good.
Our term 'hedonic goodness' is supposed to cover roughly the same ground as the word
'pleasure' in ordinary language. But, as we shall soon see, this ground is very heterogeneous
and the use of one word to cover it may produce an appearance of conceptual homogeneity,
by which we must not let ourselves become deluded.
To realize the heterogeneity of the conceptual field, in which we are moving in this chapter,
some observations on language may be helpful. In English, one is used to speaking of
pleasure and pain as a pair of contraries or opposites. In other languages, this contrast is not
so clearly marked. In German, for example, the nearest parallel to the pair 'pleasure-pain' in
ordinary parlance is 'LustUnlust'. But the German word for 'pain' is not 'Unlust'. It is
'Schmerz'. The German pair of substantives 'Lust-Unlust' answers in meaning more closely to
the English pair of adjectives 'pleasant-unpleasant' than to the substantive-pair 'pleasure-pain'.
But this correspondence too is not perfect. The words 'pleasant' and 'unpleasant' in English
would most naturally be translated by 'angenehm' and 'unangenehm' in German.
Considering the important rle which the concept of pleasure has played in ethics all through
the history of the subject, it is surprising how little this concept has been made the object of
special investigation. Neither Hume, nor the British utilitarians, nor Moore and the critics of
ethical naturalism in this century, seem to have been aware of the problematic character of
this key-notion
-63-
of their own writings.
1
Most writers in the past regard pleasure as either some kind of
sensation or as something between sensation and emotion. Moore, Broad, and the non-
naturalists in general take it for granted that pleasantness is a 'naturalistic' attribute of things
and states and not an axiological term.
2
This, I think, is a bad mistake. Some of the orthodox
views of pleasure were challenged by Professor Gilbert Ryle in an important essay a few
years ago.
3
Since then there is noticeable a new interest in the concept for its own sake -- and
not merely as an item in the ethicists' discussions of moral value.
Our discussion here of the concept of pleasure can claim neither to be deep-searching nor
even very systematic. My own feeling is that I am only scratching a surface, under which
important problems lie hidden.
I think it is useful, at least for purposes of a first approximation, to distinguish three main
forms -- as I shall call them -- of pleasure.
4

The first I call passive pleasure. It is the pleasure, or better: the pleasantness, which we
attribute primarily to sensations and other so-called states of consciousness and secondarily
also to their causes in the physical world. Pleasantness as an attribute of sensations can also
be spoken of as 'the pleasures of the senses' or as 'sensuous pleasure'. It seems to me that this
sub-form of passive pleasure is largely regarded as the prototype of all pleasure whatsoever,
and that this one-sided view has been much to the detriment of the philosophic discussion of
these topics.
The second form of pleasure I shall call, by contrast, active pleasure. It is the pleasure which
a man derives from doing things which he is keen on doing, enjoys doing, or likes to do.
Active pleasure can also be called 'the pleasures of an active life'. To the discussion of the
ethical relevance of pleasure, the pleasures of the active life seem to me to be at least as
important as the pleasures of the senses. Yet there are few moralists, apart from Aristotle, who
have paid much attention to active pleasure.
____________________
1
Hume calls pleasure and pain alternatingly 'impressions' and 'perceptions', sometimes also
'sensations'. Bentham calls them 'interesting perceptions'.
2
See e.g., the discussion by Broad in "Certain Features in Moore's Ethical Doctrines" in The
Philosopby of G. E. Moore ( 1942), pp. 57-67 and the reply by Moore, op. cit., especially
p. 587.
3
Gilbert Ryle, Dilemmas ( 1954), pp. 54-67.
4
Substantially the same distinction between three forms of pleasure is made by Broad in
Five Types of Ethical Theory ( 1930), pp. 187 and 191f.
-64-
In addition to passive and active pleasure there is that which I shall call the pleasure of
satisfaction or contentedness. It is the pleasure which we feel at getting that which we desire
or need or want -irrespective of whether the desired thing by itself gives us pleasure. The
pleasure of satisfaction has played, implicitly if not explicitly, a great rle in the formation
and discussion of the doctrine known as psychological hedonism.
2. As specimens of the use of 'good' to refer to the sub-form of passive pleasure, which we
call sensual, one may offer the phrases 'a good wine' or 'a good apple'. Let us here consider
the case of the good apple in some detail.
It should first be noted that there are many points of view, from which the goodness of apples
may become assessed. Apples are food. When we say that it is good to eat apples or that
apples are good for the children, we are probably thinking of the nourishing value and
wholesomeness of apples. This goodness of the fruit is of the form we have called the
beneficial. When the beneficial nature of apples is concerned, the attribution of goodness is
usually not of an individual apple but of apples as such or of some kind of apple.
When the cultivator or producer of apples judges of the goodness or badness of a kind of
apple, he may be thinking of such questions as whether this kind is easy to cultivate or -- in a
cold climate -is a hardy sort of apple. From the consumer's point of view, some apples are
particularly good for storing, others for making jam, others again for eating. When judged
from the producer's and the consumer's specific points of view, the goodness of apples is
often instrumental or utilitarian goodness for some purpose. When these forms of goodness in
apples are concerned, the judgment is usually about a kind of apple and not about individual
apples.
Calling an individual apple good is often another way of saying that it is not damaged or
decayed or diseased. The apple is then being treated as quasi a being, of whose good it makes
sense to talk. An apple can be 'healthy' as distinct from 'wholesome'. We need not here stop to
consider whether and when such talk is 'reducible' to talk of instrumental and utilitarian
goodness.
But calling an individual apple good can also be but another way of saying that we like its
taste, that it is good-tasting. Then the
-65-
goodness of the apple is hedonic. When it is hedonic it is, moreover, of the form we called
passive and the sub-form we called sensuous.
The hedonic judgment need not be about an individual apple. A person who says that apples
are good, may mean that he likes the taste of apples, and from this it would not follow that he
will like the taste of all individual apples. A person again, who says of an individual apple,
which he is not actually tasting, 'this is a good apple' would almost certainly be pronouncing
on a kind of apple, of which this individual is a specimen. He finds the taste of apples of this
sort good, but he perhaps dislikes the taste of some other kind of apple. I shall here disregard
such general or generalized hedonic judgments.
Consider the particular judgment expressed in the words 'this apple is good' or 'this is a good
apple', when 'good' is meant hedonically. The question may be raised: Of what is goodness
here really predicated? On the face of it, goodness is predicated of the apple. It could,
however, be suggested that the verbal formulation conceals a primary judgment, the overt
formulation of which would run 'the taste of this apple is good' or, alternatively, 'this apple is
good-tasting' or 'this apple has a good taste'. According to this suggestion, hedonic goodness
belongs primarily to the taste of the apple, and secondarily only to the apple itself. The taste is
a sensation, or bundle of sensations; the apple is a physical object. It is a causal or
dispositional characteristic of the physical object that it evokes or produces, under specific
circumstances, taste-sensations in a sensing subject. These sensations are the primary logical
subject of the hedonic value-judgment. The physical thing 'partakes', so to speak, in the
goodness of the sensations only by being their cause.
Against this idea of a primary and a secondary attribution of goodness the following objection
may be raised: To call a goodtasting apple good is both common and natural. It could hardly
be maintained that it were uncommon or unnatural to call its taste good too, this simply is one
of the uses of 'good'. But instead of calling the good-tasting apple's taste good, we could also
call it agreeable or pleasant, whereas the apple itself would not commonly and naturally be
called by those attributes. And now someone might wish to make a subtle distinction and say
that 'good' is primarily an attribute of the thing and secondarily of the sensations
-66-
it produces, 'pleasant' again primarily of the sensations and secondarily, if at all, of the
physical object.
There are other senses of 'good' which apply only to the object, and not to the sensations it
produces. For example: The apple can be good for storing or good for providing us with
Vitamin C, but its taste cannot possess such goodness. There are thus, it would seem, a great
many more senses in which the apple can be good or bad than in which its taste can be good
or bad. But to argue from this that there is no sense of 'good', which applies genuinely or
primarily to the taste of an apple (or to a sensation in general), seems to me to be quite wrong.
The lesson taught by the use of the word 'pleasant' in the context is not that sensations could
not be good in a primary sense, but that the word 'pleasant' is a synonym for the word 'good'
in one of the latter's primary uses.
I shall accept the view that hedonic goodness of the sub-form, which I have called the
pleasure of the senses, is primarily an attribute of sensations and secondarily of the objects
which produce those sensations. The sentence 'this taste is good' I shall say expresses a
primary hedonic value-judgment, and the sentence 'this apple is good', when 'good' is used in
the hedonic sense, a secondary hedonic value-judgment.
3. In the sensation which a thing, as we say, 'produces' when it affects our senses, we sense
one or several qualities of the thing. In tasting sugar, for example, we sense its taste-quality,
which is sweetness. In tasting an apple we sense many qualities: a certain juiciness, sourness,
maybe sweetness too. We could say that the taste-quality of the apple is a bundle or mixture
of several qualities. (Not all of these ingredient qualities, by the way, are what we would
ordinarily call taste-qualities. Some are olfactory qualities. Is juiciness a taste-quality?) In a
similar manner we could say that the taste-sensation which we have of the apple is a bundle
or mixture of several sensations. The several sensations themselves might be called
ingredients, or ingredient sensations, of the total sensation.
Assume that a sensation, which contains several ingredient sensations of different qualities, is
judged good or pleasant. Then it may happen that we can point to some of those ingredients
and say that we judge the sensation pleasant, because of the presence in it of those very
ingredients. The ingredients, thus pointed to, we could call good-making ingredients
(ingredient sensations); and
-67-
the qualities of the thing, thus sensed, we could call good-making qualities or properties of the
thing. For example: someone may wish to maintain that what makes him like the taste of a
certain apple, is the presence in it of a certain juiciness and sourness. Juiciness and sourness
would then be good-making qualities of the apple, and the sensations of juiciness and
sourness good-making ingredients of the total taste-sensation, which the person has of the
apple.
In a sensation judged pleasant several ingredient sensations may thus become distinguished,
i.e. several sense-qualities sensed. Now it may perhaps be thought that pleasantness itself is
one such quality, just as for example sweetness is.
1

The idea that pleasure or pleasantness were a sense-quality, i.e. something which we sense, is,
I think, a bad confusion. I shall briefly indicate why I think so.
In the sensations we sense qualities; sensations are of certain qualities, we also say. Thus, for
example, we may have a sensation of redness or sweetness. One sometimes calls a sensation
of redness a red sensation. One can do so for the sake of verbal convenience. But it is highly
misleading. For it suggests a view of sensations as a kind of thing, of the sensible qualities of
which: colour, shape, smell, etc., we can talk. (The talk of the fake-entities called sensedata
has, I am afraid, much encouraged this view among philosophers of an earlier generation.)
But a sensation of red is not the sort of entity which smells or has a colour or can be tasted.
One can sense a colour or a smell or a taste, but one cannot sense a sensation. What sort of
properties then do sensations have? A sensation can, for example, be intense or vivid or dim
or vague, it lasts for some time and then passes away, and it can be pleasant or unpleasant. A
pleasant sensation is, I believe, sometimes called a sensation of pleasure. Perhaps there is
some convenience in this mode of speech. But it should be remembered that 'a sensation of
____________________
1
Passow well-known Handwrterbuch der griechischen Sprache gives for the Greek word
++, of which 'hedonic' and 'hedonism' are linguistic off-springs, the German
equivalents sss ('sweet') and angenehm ('pleasant') and suggests that the word was
originally used for taste. Even if this suggestion be true, it does not follow, however, that
the word was used as a name of a sense-quality as disfinct from a value-attribute of
sensations of this quality. The two uses, when speaking of sweet-tasting things, probably
merged into one. These observations on the Greek ++ may be said to show how
deep-rooted is the tendency to view pleasure as quasi a sensible quality.
-68-
pleasure' is misleading in much the same way as 'a red sensation' is. That is: just as 'a red
sensation' may be regarded as a logically distorted form of 'a sensation of redness', similarly 'a
sensation of pleasure' may be regarded as a logical distortion of 'a pleasant sensation'.
But is not a sensation of sweetness a sweet sensation? I would answer: Properly speaking,
only when 'sweet' is used in an analogical sense to mean 'pleasant' or something near it.
'Sweet' in English has clear analogical uses as a value attribute. 'How sweet of you', 'How nice
of you', 'How good of you', say roughly the same. It is the existence of such analogical uses of
'sweet', I would suggest, which makes it appear more natural to speak of sweet sensations than
of red or round sensations. Of the other adjectives for taste-qualities, 'bitter' and 'sour' are also
used analogically as value attributes. But 'salt' is not quite in the same way and to the same
extent used analogically. This perhaps explains why -- at least in my ears -- it sounds less
natural to call a sensation of something salt a salt sensation than to call a sensation of
something sweet or bitter or sour a sweet or bitter or sour sensation.
4. As naming the opposite to the thing named by the substantive 'pleasure', language -- i.e. the
English language (cf. sect. 1) suggests the substantive 'pain'. 'Unpleasure' is not a word in
common use. 'Displeasure' again has to do with anger and trouble. It does not name an
opposite to that which we have here called passive pleasure, and of which the sensuous
pleasures are a sub-form. If it names an opposite of pleasure at all, then it would be of that
form which we have called the pleasure of satisfaction, or of some sub-form of it. The
displeased man feels dissatisfaction at or disapproves of something.
The adjective 'pleasant' may be said to have two linguistic opposites, 'unpleasant' and 'painful'.
The same arguments, which may be advanced for showing that pleasure and pleasantness are
not sense-qualities, also apply to the unpleasant. But which is the status of the painful in this
regard?
We speak of painful sensations. But the phrase 'a sensation of painfulness' sounds unnatural.
This would indicate that 'painful' too does not refer to a sense-quality.
On the other hand, we speak of a sensation of pain. We have pain in a tooth or pain in the
stomach. But we do not commonly
-69-
say that we have pleasure in the mouth, when eating an apple. Pain, as has often been
observed, is much more sensation-like than pleasure. The word 'pain' has analogical uses,
which resemble the use of 'pleasure' in that they make the word a value-attribute. But it seems
to me right to say that, in its primary sense, 'a pain' refers to a kind of sensation and that 'pain'
names a sense-quality, of which, however, there are many shades. In this respect 'pain' is on a
different logical level, both as compared with the substantive 'pleasure' and as compared with
the adjectives 'pleasant', 'unpleasant', and 'painful'.
Pains which are sensations are 'bodily pains'. So-called 'mental pains' are not sensations. They
are therefore 'pains' by analogy only.
That pleasure and pain are not contradictories is trivial. Not trivial, however, is that the two,
because of their logical 'asymmetry', are not even contraries in any of the senses of
'contraries', which logicians distinguish. 'Pleasant' and 'unpleasant' denote contraries, likewise
'pleasant' and 'painful' and, when used in the hedonic sense, 'good' and 'bad'. If, furthermore,
we regard the painful as a sub-form of the unpleasant, we could perhaps say that 'painfal'
names a stronger contrary to the pleasant than 'unpleasant'.
Neither 'painful' nor 'pleasant' are privative terms. Between the pleasant and the unpleasant
there is a zone of genuinely valueindifferent states, and not merely of states which are left
unclassified because of vagueness. This is a feature in which the hedonically good and bad
differ from the instrumentally good and poor and from the technically and medically good and
bad, but agrees with utilitarian goodness and badness.
Are pain-sensations ipso facto painful or, at least, unpleasant? I find the question difficult and
puzzling. Since sensations are 'naturalistic' things and unpleasantness is a value-attribute, an
affirmative answer would provide us with an interesting example of an intrinsic connexion
between a section of the world of facts and a section of the realm of values.
It seems to be a logical feature of the concept of pain that most pain is also painful or
unpleasant. This feature is probably logically connected with the facts that bodily injury
usually is painful and that severe pain in itself often is injurious, in the sense that it has
-70-
adverse effects on the possibilities of the individual concerned to satisfy his needs and wants.
Yet, though necessarily most pain is painful (unpleasant), all pain, it seems, is not. 'A pleasant
pain' is not a contradiction in terms, we have said before. (See Ch. III, sect. 9.) Some
painsensations, moreover, we actually like, judge pleasant. An example would be when a
father or mother pinches their child in a playful attitude of love or tenderness.
There can also be sensations which are both pleasant and painful to the same subject on the
same occasion. But they must, as far as I can see, be mixed sensations -- like a bitter
sweetness. Thus 'painful pleasures' are sensations or other experiences with pleasant and
painful ingredients. A 'pleasant pain' is different. It is not a sensation, which is both pleasant
and painful. It is not a painful sensation at all. It is a pain-sensation, which we happen to like,
judge pleasant, and not painful.
Of the painful pleasures it seems always true to say that their painful parts are something
which we would, as such, rather be without than suffer. We endure the pain because the
pleasure outbalances it, is greater than it. It, so to speak, pays to suffer the pain for the sake of
the pleasure. But when the child welcomes a pinch with a laugh, this is not because the pain,
though in itself disagreeable, were outbalanced by a greater pleasure. A pleasant pain is not a
price we pay for some greater pleasure, but is itself pleasant.
5. We distinguished (in sect. 2) between primary hedonic judgments, the logical subjects of
which are sensations or other states of consciousness, and secondary hedonic judgments, the
logical subjects of which are events or things in the physical world. The secondary judgments
are capable of analysis in terms of primary judgments and causal statements. The pattern of
analysis is as follows:
The secondary judgment 'this Xis good' has, roughly speaking, the same meaning as 'this X
produces or has a disposition or tendency to produce pleasant (agreeable) sensations of such
and such a kind'. An instantiation of the pattern would be: 'This apple is good' means 'this
apple produces pleasant gustatory sensations'. In ordinary life we should not express ourselves
thus, but instead of the last sentence say 'this apple has a good (pleasant) taste'. Let
-71-
us, however, not now mind the suggested piece of a philosopher's jargon. Let us also for the
moment forget about the fact that the analysans does not specify about whose sensations we
are talking, and thus is in an important respect an incomplete statement.
This analysis of secondary hedonic judgments shows a conspicuous resemblance to a well-
known attempt at analysing moral judgments. I am thinking of the theory, or a variant of it,
commonly known as the emotive theory of ethics.
1
According to this theory, broadly
speaking, the sentence 'this X is morally good' means the same as the sentence 'this X
produces or has a disposition or tendency to produce a feeling of moral approval'. The 'X'
usually stands for an arbitrary human act. Some authors speak simply of 'a feeling of
approval', omitting the adjective 'moral' from the analysans. This makes their theory simpler,
though hardly more plausible.
Our analysis of secondary hedonic judgments may be regarded as a simplified model of the
corresponding ethical theory. Some of the logical features of the ethical theory, which have
caused dispute, can, I think, be conveniently studied in the simplified model. One such
disputed point is whether moral judgments, on the emotivist analysis, are true or false, or
whether they merely are verbalized expressions of emotion and therefore lack truth-value. The
form of emotive theory, which holds the first opinion, could be called naturalistic
subjectivism. The form, which holds the second, we shall call non-cognitivist subjectivism.
Are hedonic judgments true or false? The secondary judgments involve a causal component.
It traces certain sensations back to a physical thing as their cause. This causal component we
shall here ignore. If we ignore it, our question reduces to the question whether primary
hedonic judgments are true or false. To get a firmer grasp of this second question it will be
necessary to distinguish two types of primary hedonic judgments. I shall call them, in a
technical sense, first person judgments and third person judgments.
In a first person hedonic judgment the subject is judging of a sensation, which he is himself
now experiencing or having, that it
____________________
1
As a prototype may be taken the theory of Edvard Westermarck, first presented in his
monumental work The Origin and Development of The Moral Ideas I-II ( 1906-08) and
further developed and defended in Ethical Relativity ( 1932).
-72-
is agreeable or pleasant, that he likes experiencing or having it. In a third person hedonic
judgment the subject is judging of the past, present, or future sensations of another subject
that this other subject found or finds or will find them pleasant. Also the case, when a subject
judges of the hedonic quality of his own past or future sensations, will here count as third
person judgments. The subject is then, as it were, speaking of himself from outside, in the
perspective of time.
Third person hedonic judgments obviously are true or false. That Mr. So-and-so likes or does
not like or dislikes the taste of an apple, which he is now eating, is true or false. So are the
statements that most people like the taste of this or that sort of apple or that they would, if
they tasted it, like this particular apple. The difficulties of coming to know the truth-value
may be considerable: when the apple is eaten, how can we know whether somebody, who
never tasted it, would have liked it? Perhaps the answer is that we cannot know this. But from
this does not follow that the statement is not true or false.
When ascertaining the truth of a third person hedonic judgment, we largely rely on first
person hedonic judgments. Does N. N. like the apple, which he is now eating? We ask him
and he replies 'Yes' or 'I like it'. His words express a first person hedonic judgment. It is used
for assessing the truth of the third person hedonic judgment that N. N. likes the taste of the
apple he is eating or, which means the same, that he finds the taste of the apple he is eating
good. We may regard the evidence provided by the first person judgment as being so strong
that all doubts about the truthvalue of the third person judgment are expelled. But sometimes
we do not attach much weight to the evidence, e.g. because N. N. is a very polite man and is
therefore likely to say of an apple, which we have offered him, that he likes it even when in
fact he does not. If this were the case, we should probably, in forming our opinion as to
whether he liked the fruit or not, rely more on N. N.'s facial expression, when eating the
apple, than on his words.
Do first person hedonic judgments have a truth-value? This is a very difficult question and
part of a much larger question pertaining to the logical status of first person present-tense
statements in general and first person statements about sensations in particular. One may
argue the view -- successfully, I think -- that in the first person hedonic judgments no
statements are made at all, and that
-73-
the judgments therefore cannot properly be called true or false. When the words 'the taste of
this apple is good' are used as a first person judgment, they express ('give vent to') my
pleasure at the taste and do not state that I am pleased or describe myself as a being, who
approves of the taste.
The same distinction between first person and third person judgments can, of course, also be
made for judgments about the occurrence of those more subtle phenomena called 'feelings of
moral approval or indignation'. In the case of such feelings, too, it is fairly obvious that the
third person judgments are objectively true or false, whereas it is at least arguable that the first
person judgments are not true or false statements about feelings, but neither true nor false
expressions of feeling.
One could try to do distributive justice to the claims both of naturalistic and non-cognitivist
subjectivism as theories of hedonic value-judgments by apportioning to each theory a due
share in the truth -- to the first because of the propositional character of the third person
judgments, and to the second because of the interjectional character of the first person
judgments. But by practising such impartiality one runs risk of obscuring an important point.
This point is that the third person judgments, just because of the feature of theirs which makes
them true or false, viz. that they are about the valuations of other subjects (or about the
judging subject's own valuations viewed in the perspective of time), are no genuine value-
judgments at all. They are no value-judgments, since they do not value, but report or
conjecture about human reactions, i.e. such reactions which we call valuations. The only
genuine value-judgments in the context are the first person judgments. In them the judging
subject values his sensations. They are not true or false, and therefore, in a sense of the word,
no 'judgments' even. For this reason it seems to me fair to say that non-cognitivist
subjectivism represents the correct view of hedonic valuejudgments, whereas naturalistic
subjectivism is not a theory of value-judgments at all.
In their relation to truth (primary) hedonic value-judgments, unless I am badly mistaken,
differ importantly from judgments of instrumental, technical, or utilitarian goodness. These
latter judgments are true or false; one can always be mistaken in them. The primary hedonic
value-judgments are neither true nor false, there is no room for mistake in them. They are, in
this peculiar
-74-
sense, 'subjective'. In their sphere one cannot distinguish between an apparent and a real
good; 'to be good' and 'to be judged (or considered or thought) good' are here one and the
same. But judgments in which we affirm or anticipate the occurrence of such and such
hedonic valuations in ourselves or in other subjects are, of course, true or false judgments --
though not value-judgments.
6. Hedonic goodness of the form which we called passive pleasure, is a value-attribute of
sensations and other states of consciousness. To call this form of goodness an attribute or
characteristic or property of sensations is useful when we want to explain why pleasure itself
is not a sensation or a sense-quality (of a thing, which is sensed in a sensation). But to call
hedonic goodness a property can also be misleading.
Sensations are tied to a subject. They are somebody's sensations. If the secondary hedonic
judgment expressed in the words 'this apple is good' is analysed in terms of primary
judgments about the goodness or pleasantness of some gustatory sensations, the question will
instantly arise: Of whose sensations are we here talking? The overt form of the sentence,
which expresses the secondary judgment, does not give us any guidance, since it does not
mention an apple-taster. Nor does the overt form of sentences, which express primary hedonic
judgments, always mention a subject. 'The taste of this apple is good' and 'this is a pleasant
taste' are complete sentences. But I think it is correct to say that the sense of the sentences is
incomplete, unless it is understood, from the context or otherwise, whose taste-sensations are
meant. If this is true of sentences expressing a primary hedonic judgment, the same will a
fortiori be true of any sentence expressing a secondary hedonic judgment, into the analysis of
which the primary judgment may enter.
How is mention of a sensing subject to be worked into the overt form of a sentence expressing
a hedonic judgment about the taste of an apple? Let us assume, for the sake of argument, that
the subject is the speaker himself. He might then, instead of 'the taste of this apple is good',
say 'I find the taste of this apple good'. Instead of 'find' he may also say 'think' or 'judge' or
'consider'. Perhaps he would use the form of words 'I think that the taste of this apple is good'.
But this sounds artificial, at least as an expression of his judgment, when he is actually tasting
the apple. (It sounds
-75-
more like a conjecture that he will like the taste.) It would also sound rather artificial to say
'the taste of this apple is good to me'. But to say 'I like the taste of this apple' -- thus not using
the word 'good' at all -- would sound perfectly natural.
It will strike us that a majority of these sentences, which mention a subject, are not of the
ordinary subject-predicate form, in which a property is predicated of a thing. It may be
questioned whether any of them really is of this form. It may also be questioned whether not
the overt subject-predicate form of sentences expressing hedonic judgments with suppressed
reference to a sensing subject is really spurious, viz. as a reflexion of the judgment's logical
form. Among the sentences which we gave as examples, the one which comes nearest to
being of the ordinary subjectpredicate form is 'the taste of this apple is good to me'. If 'good to
me' could be said to name a property, then this sentence could be safely said to be a subject-
predicate sentence. But 'good to me' rather suggests a relation between the apple-taster and the
appletaste than a property of the taste-sensation. If it be asked what this relation is, a plausible
answer would be that it is the liking-relation, or some relation closely akin to it. The sensing
subject likes, enjoys, approves of the taste of the apple. This is what could be reasonably
meant by the unnatural-sounding form of words 'the taste of this apple is good to me'.
The sentence 'I think that the taste of this apple is good' contains a subject-predicate sentence
as a part, within a 'that'-clause. It is not itself a subject-predicate sentence. As already noted, it
has a certain artificiality about itself. It suggests to us that a distinction could be made
between the taste's being good -- as it were 'in itself' -- and somebody's thinking or judging or
opining of the taste that it is good. I can, of course, think or judge or opine that I shall like the
taste of this apple, and perhaps find on tasting it that I was mistaken. But, as we had already
occasion to stress, when we were speaking of the subjectivity of hedonic valuejudgments, I
cannot judge the taste of the apple, which I am now tasting, good and be mistaken. Therefore,
to think that the taste is good is not to think that something is thus and thus, but might be
otherwise. The taste, which is the logical subject of the first person judgment, cannot, from
the point of view of its hedonic value, be anything else but what I think it is. The use of 'think
that' is here quite unlike the normal use of this phrase.
-76-
It was this slightly unnatural use of 'that' -- clauses in valuesentences which suggested to
Moore
1
the following refutation of subjectivism in value-theory: If 'this is good' means the
same as 'I think that this is good', then by raising the same question over again for the 'this is
good, which occurs inside the 'that'-clause, we get as in answer that 'this is good' means the
same as 'I think that I think that this is good', and so on ad infinitum. This is a clever point as
it stands, but not of much consequence for the task of refuting subjectivism. It only
establishes that, if the words 'I think that this is good' are offered as an equivalent of 'this is
good', then the 'I think that . . .'-form functions in a peculiar way, which should make its very
use here sound suspect to our logical ear. If the subjectivist says that 'this is good' means 'I
think this good' or 'I consider this good', omitting the word 'that', he is already better off, and
Moore's objection cannot be raised against him.
2

These considerations, in my opinion, tend to show that hedonic goodness is not a property
and that the subject-predicate form of sentences, which express hedonic judgments, is a
spurious reflexion of the logical form of such judgments, which is the relational form. As the
logically most satisfactory formulations of such judgments in language I should regard their
formulation with the aid of the verb 'to like', as for example in the sentence, 'I like the taste of
this apple'. But this, needless to say, does not make the subjectpredicate form either useless or
incorrect as a shorthand formulation in all those innumerous cases, when mention of a sensing
subject is omitted from the sentence. The sentences 'this is a good apple' or 'this is a good
taste' are all right as they stand. But they are apt to mislead the philosopher by concealing a
logical form.
7. To think of pleasure exclusively in terms of predications of pleasantness to various logical
subjects is thus philosophically
____________________
1
See Ethics ( 1912), p. 76, Moore, when giving the 'refutation', is immediately dealing with
the subjectivist suggestion that 'this is right' meant the same as 'somebody believes that this
is right'. But the same argument applies, he maintains, to the suggestion that the rightness
of an act meant that the act is thought to be right (p. 80) and also to the corresponding
suggestion about goodness (p. 98).
2
It is significant that Moore, when stating the subjectivist view, often uses the formulation
'thinks it right (good)', but when trying to refute it by means of the argument ad infinitum
switches to the formulation 'thinks (believes) that it is right (good)'.
-77-
misleading. To think of pleasure in terms of liking is philosophically enlightening and helpful.
Beside 'like', also 'approve' and 'enjoy' are relational verbs, which may be used for expressing
hedonic judgments. That of which we approve is normally said to please us, but is not
ordinarily said to give us pleasure. 'Enjoy' is more obviously hedonic than 'approve'; any
source of enjoyment can also be called a source of pleasure, and vice versa.
Among our likings an important position are held by things we like to do. One man likes to
watch cricket, another to play chess, a third likes to get up early in the morning.
The way in which liking to do is connected with pleasure is rather different in these three
examples. The pleasure of watching a game is mainly, I should think, of the form which we
have called passive pleasure. Watching a game means the acquisition of experiences which
the man, who likes watching the game, finds pleasant. Is his pleasure that of the senses? There
are reasons for saying that it is, since our man is enjoying a sight. There are perhaps even
stronger reasons for saying that it is not, since enjoying this sight normally requires both
knowledge of the rules of the game and some familiarity with the practice of playing it. The
border between sensuous pleasure and other forms of passive pleasure is very elastic.
The pleasure of playing a game has many aspects. Sometimes one plays a game just for
amusement, as a pastime. Then the pleasure which one derives from playing the game is
substantially passive pleasure, i.e. the pleasantness of certain experiences. But often we like
playing a game, not for the sake of amusement', but because we are interested in the game,
keen on the art of playing it. Then the delight we take in the game is of the form which I
called active pleasure. The same is the case with any pleasure derived from the practising of
any activity, of which it is true to say that we are keen on it or that we like it 'for its own sake'.
It will occur to us that there is a connexion between technical goodness and the form of
hedonic goodness called active pleasure. This connexion may be intrinsic. The two forms of
goodness are nevertheless distinct.
Consider next the man who likes to get up early in the morning. Must he find early rising
pleasant? Some men may rise early for 'hedonic' reasons, i.e. in order to enjoy the morning:
the freshness of the morning air, the beauty of the sunrise, etc. But rather few,
-78-
I think, of those who say they 'like' to get up early, would give such reasons for their liking.
Someone may like to get up early because he has so many things to do that, if he stays in bed
till late, he will not have time to do them at all, or his afternoon will be badly rushed or he
will have to work at night. But he may be completely indifferent to the peculiar pleasures of
the early morning hours. Should this man not rather then say that he wants to get up early than
say that he likes to get up early? Or perhaps the suggestion will be that our man should say
that he has to or must get up early, considering that this is not anything which he likes or
wants to do 'for its own sake', but something that is forced upon him by the 'practical
necessities' of life. (Cf. Ch. VIII.) The answer is that the uses of 'like to do' and 'want to do'
and 'have to do' shade into one another, and that we sometimes say that we like to do things,
the doing of which is a source neither of passive nor of active pleasure to us.
But is there not at least a remote connexion with pleasure also in this third case which we
have been discussing? The man who rises early may want to do so in order to avoid having to
rush his day's work, which is an unpleasant thing. Or he may be anxious to finish his set work
as early in the day as possible, so that he can relax and do in the afternoon what he 'really
likes', i.e. that which affords him (active or passive) pleasure. These possibilities are not
unrealistic. Now it may be argued that, unless our man has some such desire either to avoid
something unpleasant or to secure for himself some pleasure, then he could not say truly of
himself even that he wants to get up early in the morning. To argue thus about the man would
be to apply to his case a general philosophic thesis about the nature of man, viz. that all action
is, in the last resort, necessarily prompted by a desire to secure some pleasure or avoid
something unpleasant, This is the thesis, or a version of the thesis, known as psychoiogical
hedonism. We must here try to form some opinion of its truth.
8. Not everything which a man can be said to do is voluntary action. For example: getting fat
or sleeping. And not everything which a man does and which is voluntary action, can he also,
on any ordinary understanding of the words, be said to want to do. Most things which a man
does 'because this is the custom', e.g. taking off his hat when greeting a lady, or 'because this
is the rule
-79-
or law', e.g. driving to the left or halting in front of a major road ahead, are not things he
'wants to do', whenever he does them. But some of the things which a man voluntarily does,
he also wants to do, and some of the things he wants to do he also likes to do.
To maintain psychological hedonism for all voluntary action is hardly feasible. To maintain it
for those acts and activities only which a man enjoys doing or likes to do, may seem too
narrow to be of much interest. The exact scope of the claim of psychological hedonism is
seldom made clear. I shall here regard it as a thesis concerning at least all those things, which
a man can be properly said to want to do. What does this thesis say? That too is seldom made
sufficiently clear. I shall here understand it as saying something which can, for purposes of a
first approximation, be stated as follows: If, in an individual case, we raise the question why a
man wants that which he wants to do, the answer will, if not immediately then after a chain of
questions and answers, have to mention something which this man likes, finds pleasant or
dislikes, finds unpleasant or painful. This statement is admittedly vague, but I hope the
subsequent discussion of it will make its intended meaning clearer. (See also Ch. V, sect. 2.)
A man says he wants to get up early to-morrow morning. Why? Because he wants to see the
sunrise. Why? Because he enjoys the sight, likes it. The answer to the further question, why
he likes sunrises, is not that their sight gives him pleasure. For his liking of the sight of a
sunrise and the sight's giving him pleasure are one and the same.
The question may, of course, be raised, why the spectacle pleases him. It is not certain that it
can be answered. It would be answered by pointing to some fact in the man's life-history,
which tells us what made him like ('caused' him to like) sunrises. This question 'Why?' does
not ask for a reason or motive, but for a (kind of) causal explanation. The question 'Why?',
which is relevant to the thesis of psychological hedonism, is a question concerning reasons for
doing, i.e. ends in acting.
Our early riser can rightly be called a pleasure-seeker. His action is motivated by a desire to
secure for himself a certain pleasant experience. He wants to get up early, not because early
rising is, as such or in itself, pleasant to him, but because the act is conducive to pleasure, i.e.
is a necessary condition for his attaining a pleasant experience in the end.
-80-
Consider now the following case: A man wants to get up early some morning to see the
sunrise. He has never seen one before, but he has heard many people praise the beauty of the
sight. He does not, however, expect that he will particularly like it and he is not anxious to
secure for himself a new pleasure. He simply is curious to know what the sight, which so
many people praise, looks like. Inquisitiveness can be a thoroughly self-sufficient motive of
action. The chain of questions and answers could run as follows: Why does he want to get up
early? Because he wants to see the sunrise. Why does he want to see the sunrise? Because he
is curious about the sight. Here is no mention of liking or pleasure. Perhaps we can also
answer the question: What made him curious? and that the answer is: The fact that so many
poets, whom he reads, have written enthusiastically about sunrises. This question is causal.
The fact that its answer can be said to hint at the likings and pleasures of poets is here
irrelevant. Our man could equally well be curious to see something which people notoriously
dislike. His inquisitiveness, not his desire for some pleasant experience, prompts his action.
I think that the case of the curious man refutes hedonism, but the difficulty is to see that it
really does so decisively. For could one not argue as follows:
Our inquisitive man is anxious to satisfy his inquisitiveness. The satisfaction of this desire
gives him pleasure. If any case of doing that, which one wants to do, can be correctly
described as a case of satisfying some desire -- and let us not here query the correctness of
this view -- and if to have one's desires satisfied is a pleasure, is then not psychological
hedonism after all right?
Before we answer this question, we must say some words about the relation of pleasure to
satisfaction of desire.
Desire is sometimes called a dissatisfaction or discontentedness with a prevailing state of
things; it is an impulse or longing to change this state to another. The change, at which a
desire aims, could therefore also be characterized as a transition from a present 'unpleasure' to
a future pleasure. Here, however, great caution is needed. The phrase 'transition from a
present unpleasure to a future pleasure' can mean many things. It can mean that the state in
which we are now is judged unpleasant, and the one at which we aim is thought of as
pleasant. For example: we feel uncomfortable or bored and want to do something which will
make us feel
-81-
comfortable or which will amuse us, cheer us up. Here unpleasantness and pleasantness are
clearly hedonic features of the states. This is a common type of desire-situation. But it is not
the only type. The case, e.g., of the man who wants to satisfy his curiosity, is different. To
have an unsatisfied curiosity can be agitating, exciting, vexing. It can be unpleasant too. But it
need not be so. Someone may even think that curiosity is more of a pleasant than an
unpleasant feeling. Yet on the other hand: if attempts to satisfy our curiosity fail, we shall
probably feel annoyed or grieved or even outraged, and these states we should ordinarily
judge unpleasant. Something corresponding, it seems, holds good of desire in general:
frustration is unpleasant, sometimes to the degree of being painful.
I think it is correct to say that frustration of desire is intrinsically unpleasant. We should not
say that a desire had become 'frustrated', unless one or several unsuccessful attempts to gratify
the desire had not had some hedonically bad consequences, such as anger, annoyance, grief,
impatience, hurt vanity, or the like.
Similarly, it seems to me right to think that satisfaction of desire is intrinsically pleasant. But
the connexion between satisfaction and pleasure is more complicated than the connexion
between frustration and hedonic badness. This is due to the fact that the description of a desire
involves mention of an object of desire. The object is that which we desire. The attainment of
the object intrinsically, 'by definition', satisfies the desire. Thus attainment of the object is, in
a sense, intrinsically satisfying. But from this does not follow that the object itself were
intrinsically pleasant or otherwise hedonically good. The hedonically good consequences,
which are intrinsically connected with satisfaction of desire, consist in feelings of
contentedness or joy or power or relief or something similar. If they are not consequent upon
the attainment of the object, we should doubt whether there was any desire at all or whether it
had been correctly described as a desire for so-and-so.
It is on the evidence of such hedonically tinged consequences both of frustration and of
satisfaction that we often judge the strength of the desire. 'He cannot have wanted it very
eagerly, since he was not very glad when he got it'; 'He must have desired it strongly, since
failure to get it depressed him so much', we say.
For the pleasure, which is intrinsically connected with satisfaction of desire, we already
coined (in section 1) the name pleasureof satisfaction.
-82-
of satisfaction. It is the existence of an intrinsic tie between satisfaction and pleasure, I think,
which is above all responsible for the strong appearance of truth, which the thesis of
psychological hedonism undoubtedly possesses. The refutation of hedonism must therefore
not consist in an attempt to deny the existence of this connexion.
Wherein then does the refutation consist? As far as I can see, it consists solely in this:
The pleasure of satisfying a desire can never be an object of that same desire. For satisfaction
presupposes a desire, and a desire in its turn presupposes something which we desire, an
object. Therefore the object of desire must necessarily be different from the pleasure of
satisfying that desire. But the pleasure drawn from the satisfaction of a desire can itself
become the object of a new desire. This is perhaps not very common, but it is not an entirely
unrealistic possibility. Consider again the man who wants to see the sunrise, because he is
curious about the sight. It is conceivable that experience had taught him that satisfying a
curiosity is something very pleasant. He is of a peculiar inquisitive disposition or temper.
Each time he is curious about something he is extremely tense, and when his curiosity has
become satisfied he has an immensely exhilarating and joyful feeling of relief. This man
welcomes every opportunity for satisfying a curiosity, because of the pleasure this gives him.
This pleasure is the passive pleasure of some peculiar experience following upon the
satisfaction of a curiosity. There is nothing in the logic of things to prevent it from becoming
the object of a new desire, viz. the desire for the pleasure of having satisfied a curiosity. But
whether the original pleasure of satisfaction will or will not itself thus turn into an object of
desire is an entirely contingent matter. It is the contingent nature of this fact which, in my
opinion, constitutes the refutation of psychological hedonism. The error of hedonism is that it
mistakes the necessary connexion, which holds between the satisfaction of desire and
pleasure, for a necessary connexion between desire and pleasure as its object. This mistake is
easy to make, but not quite easy to expose. This is why I have spent so much time here on the
refutation of psychological hedonism, although this doctrine is said to have been refuted over
and over again in the past. I am not certain that any 'refutation' does full justice to the
complications of the theory, and the problems raised by it still continue to vex me.
-83-
9. Something can be an object of desire, although it is thought unpleasant. The question may
be raised whether anything can be an object of desire because it is considered unpleasant.
Can, in other words, the contrary of pleasure be an object of desire?
A pain can be an object of desire. A man can want to inflict pain upon himself as a
chastisement. But he can also want this simply because he likes the pain. This is perhaps
perverse, but it is not contrary to logic. It is not illogical, since -- as we have said before -a
pleasant pain is not a contradiction in terms.
A man can also want something for himself which he finds unpleasant -- for example to
undergo a surgical operation because he considers it necessary or good for his bodily health or
general well-being. A man can desire the unpleasant out of sheer curiosity. 'What will be my
reaction? Shall I faint or vomit?' One can be intensely curious about such things. A man could
take a perverse pleasure in vomiting, and for the sake of this pleasure want to eat something
extremely distasteful. In none of these cases, however, in which something unpleasant is
wanted, do we want the unpleasant for the sake of its being unpleasant, but for the sake of
something else -- some pleasure maybe -- to which it is conducive.
The fact that the contrary of pleasure, if I am right, thus necessarily is not an object of desire,
may have contributed to the illusion of psychological hedonism that pleasure necessarily is
the ultimate end, after which people aspire whenever they want something.
10. A few brief remarks will here be made on the doctrine known as Ethical Hedonism.
We shall distinguish between two principal forms, which this doctrine may assume:
Firstly, Ethical Hedonism can be a theory about the concept of goodness or the meaning of
the word 'good'. In its crudest form this theory maintains that any context where the word
'good' is used (not mentioned), either is one from which 'good' can be eliminated by simply
substituting for it the word 'pleasant' -- as in the phrase 'a good-smelling flower' -- or is one
from which 'good' can be eliminated by means of an analysis in terms of 'pleasure' -as, on our
suggested view of secondary hedonic judgments, it may become eliminated from the phrase 'a
good apple'.
Secondly, Ethical Hedonism can be an axiological theory about the character of good things.
In its crudest form this theory
-84-
defends some such view as that those and only those things are good, which either are judged
pleasant in themselves, or are (somehow) instrumental or useful for the production of
pleasure, i.e. causally responsible for the coming into being of pleasant things.
Moore, in his well-known criticism of hedonism, thought that the axiological theory could be
refuted on the ground that it conflicts with our value-intuitions and the conceptual theory on
the ground that it commits 'the Naturalistic Fallacy'. 'Good' means good, Moore says, and not
anything else, e.g. pleasant. About this I shall only say that it seems to me just as obvious that
'good' sometimes means 'pleasant' or can otherwise become translated into hedonic terms as it
seems to me obvious that 'good' does not always mean 'pleasant' or can become thus
translated. The pleasant, pleasure, we have called a form of the good or of goodness. It is
equally futile to try to reduce this form to one or several others as it is to try to reduce all
other forms to it. But there may exist logical connexions of a more complex and subtle nature
between the forms.
If one is aware of the multiform nature of goodness, one will realize that the general question
'Is pleasure good?' is unintelligible, unless the form of goodness is specified. In one sense of
'good' the question is just as empty of content or logically defect as the questions 'Is pleasure
pleasant?' or 'Is goodness good?'. When correctly stated, the question must mean something
along the following lines: Are the things, which are good hedonically, also good in some
other respect? And here this other respect must be specified. The question may be well worth
discussing. So may the converse question be: Are the things, which are in such-and-such
respect good, also good hedonically? In the case of neither question, however, would an
affirmative answer establish that pleasure is the 'sole and ultimate good' in any reasonable
sense of those unprecise words.
-85-
V
THE GOOD OF MAN
1. THE notion of the good of man, which will be discussed in this chapter, is the central
notion of our whole inquiry. The problems connected with it are of the utmost difficulty.
Many things which I say about them may well be wrong. Perhaps the best I can hope for is
that what I say will be interesting enough to be worth a refutation.
We have previously (Ch. III, sect. 6) discussed the question, what kind of being has a good.
We decided that it should make sense to talk of the good of everything, of the life of which it
is meaningful to speak. On this ruling there can be no doubt that man has a good.
Granted that man has a good -- what is it? The question can be understood in a multitude of
senses. It can, for example, be understood as a question of a name, a verbal equivalent of that
which we also call 'the good of man'.
We have already (Ch. I, sect. 5) had occasion to point out that the German equivalent of the
English substantive 'good', when this means the good of man or some other being, is das
Wohl. There is no substantive 'well' with this meaning in English. But there are two related
substantives, 'well-being' and 'welfare'.
A being who, so to speak, 'has' or 'enjoys' its good, is also said to be well and, sometimes, to
do well.
The notion of being well is related to the notion of health. Often 'to be well' means exactly the
same as 'to be in good, bodily and mental, health'. A man is said to be well when he is all
right, fit, in good shape generally. These various expressions may be said to refer to minimum
requirements of enjoying one's good.
Of the being who does well, we also say that it flourishes, thrives,
-86-
or prospers. And we call it happy. If health and well-being primarily connote something
privative, absence of illness and suffering; happiness and well-doing again primarily refer to
something positive, to an overflow or surplus of agreeable states and things.
From these observations on language three candidates for a name of the good of man may be
said to emerge. These are 'happiness', 'well-being', and 'welfare'.
The suggestion might be made that 'welfare' is a comprehensive term which covers the whole
of that which we also call 'the good of man' and of which happiness and well-being are
'aspects' or components' or 'parts'. It could further be suggested that there is a broad sense of
'happiness', and of 'well-being', to mean the same or roughly the same as 'welfare'. So that, on
one way of understanding them, the three terms could be regarded as rough synonyms and
alternative names of the good of man.
The suggestion that 'the good of man' and 'the welfare of man' are synonymous phrases I
accept without discussion. That is: I shall use and treat them as synonyms. (Cf. Ch. I, sect. 5;
also Ch. III, sect. 1.)
It is hardly to be doubted that 'happiness' is sometimes used as a rough synonym of 'welfare'.
More commonly, however, the two words are not used as synonyms. Happiness and welfare
may, in fact, become distinguished as two concepts of different logical category or type. We
shall here mention three features which may be used for differentiating the two concepts
logically.
First of all, the two concepts have a primary connexion with two different forms of the good.
One could say, though with caution, that happiness is a hedonic, welfare again a utilitarian
notion. Happiness is allied to pleasure, and therewith to such notions as those of enjoyment,
gladness, and liking. Happiness has no immediate logical connexion with the beneficial.
Welfare again is primarily a matter of things beneficial and harmful, i.e. good and bad, for the
being concerned. As happiness, through pleasure, is related to that which a man enjoys and
likes, in a similar manner welfare, through the beneficial, is connected with that which a man
wants and needs. (Cf. Ch. I, sect. 5.)
Further, happiness is more like a 'state' (state of affairs) than welfare is. A man can become
happy, be happy, and cease to be happy. He can be happy, and unhappy, more than once in
his life. Happiness, like an end, can be achieved and attained. Welfare has
-87-
not these same relationships to events, processes, and states in time.
Finally, a major logical difference between happiness and welfare is their relation to
causality. Considerations of welfare are essentially considerations of how the doing and
happening of various things will causally affect a being. One cannot pronounce on the
question whether something is good or bad for a man, without considering the causal
connexions in which this thing is or may become embedded. But one can pronounce on the
question whether a man is happy or not, without necessarily considering what were the causal
antecedents and what will be the consequences of his present situation.
The facts that happiness is primarily a hedonic and welfare primarily a utilitarian notion, and
that they have logically different relationships to time and to causality, mark the two concepts
as being of that which I have here called 'different logical category or type'. It does not follow,
however, that the two concepts are logically entirely unconnected. They are, on the contrary,
closely allied. What then is their mutual relation? This is a question, on which I have not been
able to form a clear view. Welfare (the good of a being) is, somehow, the broader and more
basic notion. (Cf. Ch. III, sect. 12.) It is also the notion which is of greater importance to
ethics and to a general study of the varieties of goodness. Calling happiness an 'aspect' or
'component' or 'part' of the good of man is a non-committal mode of speech which is not
meant to say more than this. Of happiness I could also say that it is the consummation or
crown or flower of welfare. But these are metaphorical terms and do not illuminate the logical
relationship between the two concepts.
2. By an end of action we shall understand anything, for the sake of which an action is
undertaken. If something, which we want to do, is not wanted for the sake of anything else,
the act or activity can be called an end in itself.
Ends can be intermediate or ultimate. Sometimes a man wants to attain an end for the sake of
some further end. Then the first end is intermediate. An end, which is not pursued for the sake
of any further end, is ultimate. We shall call a human act end-directed, if it is undertaken
either as an end in itself or for the sake of some end.
-88-
What is an ultimate end of action is settled by the last answer, which the agent himself can
give to the question, why he does or intends to do this or that. It is then understood that the
question 'Why?' asks for a reason and not for a causal explanation of his behaviour. (Cf. Ch.
IV, sect. 8.)
In the terms which have here been introduced, we could redefine Psychological Hedonism as
the doctrine that every enddirected human act is undertaken, ultimately, for the sake either of
attaining some pleasure or avoiding something unpleasant. The doctrine again that every end-
directed human act is undertaken, ultimately, for the sake of the acting agent's happiness we
shall call Psychological Eudaimonism. A doctrine to the effect that every end-directed act is
ultimately undertaken for the sake of the acting agent's welfare (good) has, to the best of my
knowledge, never been defended. We need not here invent a name for it.
Aristotle sometimes talks
1
as though he had subscribed to the doctrine of psychological
eudaimonism. If this was his view, he was certainly mistaken and, moreover, contradicting
himself. It would be sheer nonsense to maintain that every chain of (noncausal) questions
'Why did you do this?' and answers to them must terminate in a reference to happiness. The
view that man, in everything he does, is aiming at happiness (and the avoidance of misery) is
even more absurd than the doctrine that he, in everything he does, is aiming at pleasure (and
the avoidance of pain).
I said that, if Aristotle maintained psychological eudaimonism, he was contradicting himself.
(And for this reason I doubt that Aristotle wanted to maintain it, though some of his
formulations would indicate that he did.) For Aristotle also admits that there are ends, other
than happiness, which we pursue for their own sake. He mentions pleasure and honour among
them.
2
Even 'if nothing resulted from them, we should still choose each of them', he says.
3

On the other hand, those other final ends are sometimes desired, not for their own sake, but
for the sake of something else. Whereas happiness, Aristotle thinks, is never desired for the
sake of anything else.
4
Pleasure, e.g. pleasant amusement, can be desired for
____________________
1
e.g., Ethica Nicomachea (EN), 1094a 18-21, 1095a 14-20, and 1176b 30-31.
2
EN, 1097b 1-2. See also 1172b 20-23.
3
EN, 1097b 2-3.
4
EN, 1097b 1 and again 1097b 5-6.
-89-
relaxation, and relaxation for the sake of continued activity.
1
Then pleasure is not a final end.
I would understand Aristotle's so-called eudaimonism in the following light: among possible
ends of human action, eudaimonia holds a unique position. This unique position is not that
eudaimonia is the final end of all action. It is that eudaimonia is the only end that is never
anything except final. It is of the nature of eudaimonia that it cannot be desired for the sake of
anything else. This is, so Aristotle seems to think, why eudaimonia is the highest good for
man.
2

It is plausible to think that a man can pursue, i.e. do things for the sake of promoting or
safeguarding, his own happiness only as an ultimate end of his action. A man can also do
things for the sake of promoting or safeguarding the happiness of some other being. It may be
thought that he can do this only as an intermediate end of his action. The idea has an apparent
plausibility, but is nevertheless a mistake. The truth seems to be that a man can pursue the
happiness of others either as intermediate or as ultimate end.
The delight of a king can be the happiness of his subjects. He gives all his energies and work
to the promotion of this end. Maybe he sacrifices his so-called 'personal happiness' for the
good of those over whom he is set to rule. Yet, if this is what he likes to do, it is also that in
which his happiness consists. To say this is not so distort facts logically. But to say that the
king sacrifices himself for the sake of becoming happy and not for the sake of making others
happy, would be a distortion. It would be a distortion similar to that of which psychological
hedonism is guilty, when it maintains that everything is done for the sake of pleasure, on the
ground that all satisfaction of desire may be thought intrinsically pleasant.
Can a man's welfare be an end of his own action? The question is equivalent to asking,
whether a man can ever be truly said to do things for the sake of promoting or protecting his
own good. It is not quite clear which is the correct answer.
____________________
1
Cf. EN, 1176b 34-35.
2
There is no phrase in Aristotle's ethics which corresponds to our phrase 'the good of man'.
Eudaimonia (happiness, well-being) Aristotle also calls the best or the highest good. The
notion of a summum bonum, however, is not identical with the notion of the good of man
as we use it here. But the two notions may be related.
-90-
On the view which is here taken of the good of a being, to do something for the sake of
promoting one's own good, means to do something because one considers doing it good for
oneself. And to do something for the sake of protecting one's own good means to do
something because one considers neglecting it bad for oneself.
For all I can see, men sometimes do things for the reasons just mentioned. This would show
that a man's welfare can be an end of his own action.
Yet the good of a being as an end of action is a very peculiar sort of 'end'. Normally, an end of
action is a state of affairs, something which 'is there', when the end has been attained. But
welfare is not a state of affairs. (Cf. the discussion in section 1.) For this reason I shall say that
welfare, the good of a being, can only in an oblique sense be called an end of action.
Obviously, the reason why a man does something, which he considers good for himself, is not
always and necessarily that he considers doing it good for himself. Similarly, the reason why
a man does something, which he considers bad for him to neglect, is not always and
necessarily that he considers neglecting it bad for himself. This shows that a man's own
welfare is not always an ultimate end of his action. It also shows that a man's own welfare is
not always an end of his action at all. It does not show, however, that a man's own welfare is
sometimes an intermediate end of his action. Whether it can be an intermediate end, I shall
Thus, on our definitions, the answer to the question whether a certain thing is good or bad for
a man, is independent of the following two factors: First, it is independent of whether he (or
anybody else) judges or does not judge of the value of this thing for him. Secondly, it is
independent of what he (and everybody else) happens to know or not to know about the causal
connecxions of this thing. Yet, in spite of this independence of judgment and knowledge, the
notions of the beneficial and the harmful are in an important sense subjective. Their
subjectivity consists in their dependence upon the preferences (wants) of the subject
concerned.
Considering what has just been said, it is clear that we must distinguish between that which is
good or bad for a man and that which appears,i.e. is judged or considered or thought (by
himself or by others), to be good of bad for him.
Any judgment to the effect that something is good or bad for a man is based on such
knowledge of the relevant causal connexions which the judging subject happens to possess.
Since this knowledge may be imperfect, the judgment which he actually passes may be
different from the judgment which he would pass, if he had perfect knowledge of the causal
connexions. When there is this discrepancy between the actual and the potential judgment, we
shall say that a man's apparent good is being mistaken for his real good.
Of certain things it is easier to judge correctly whether they are good or bad for us, than of
certain other things. This means: the risks of mistaking our apparent good for our real good
are sometimes greater, sometimes less. It is, on the whole, easier to judge correctly in matters
relating to a person's health than in matters relating to his future career. For example: the
judgment that it will do a man good to take regular exercise is, on the whole, safer than the
judgment that it will be better for him to go into business than study medicine. Sometimes the
difficulties to judge correctly are so great that it will be altogether idle and useless to try to
form a judgment.
Sometimes we know for certain that a choice, which we are facing, is of great importance to
us in the sense that it will make considerable difference to our future life, whether we choose
the one or the other of two alternatives. An example could be a choice between getting
married or remaining single or between accepting employment in a foreign country or
continuing life at home. But
-109-
certainty that the choice will make a great difference is fully compatible with uncertainty as to
whether the difference will be for good or for bad. The feeling that our welfare may become
radically affected by the choice, can make the choice very agonizing for us.
Also of many things in our past, which we did not deliberately choose, we may know for
certain that they have been of great importance to us in the sense that our lives would have
been very different, had these things not existed. This could be manifestly true, for example,
of the influence which some powerful personality has had on our education or on the
formation of our opinions. We may wonder whether it was not bad for us that we should have
been so strongly under this influence. Yet, if we know only that our life would have been very
different but cannot at all imagine how it would have been different, we may also be quite
incompetent to form a judgment of the beneficial and harmful nature of this factor in our past
history.
It is a deeply impressive fact about the condition of man that it should be difficult, or even
humanly impossible, to judge confidently of many things which are known to affect our lives
importantly, whether they are good or bad for us. I think that becoming overwhelmed by this
fact is one of the things which can incline a man towards taking a religious view of life. 'Only
God knows what is good or bad for us.' One could say thus -- and yet accept that a man's
welfare is a subjective notion in the sense that it is determined by what he wants and shuns.
13. Are judgments of the beneficial or harmful nature of things objectively true or false?
When we try to answer this question, we must again observe the distinction between a first
person judgment and a third person judgment. (Cf. Ch. IV, sect. 5 and this chapter, sections 6
and 7.)
When somebody judges of something that it is (was, will be) good or bad for somebody else,
the judgment is a third person judgment. It depends for its truth-value on two things. The one
is whether certain causal connexions are as the judging subject thinks that they are. The other
is whether certain valuations (preferences, wants) of another subject are as the judging subject
thinks that they are. Both to judge of causal connexions and to judge of the valuations of other
subjects is to judge of empirical
-110-
matters of fact. The judgment is 'objectively' true or false. It is, properly speaking, not a
value-judgment, since the 'axiological' component involved in it is not a valuation but a
judgment about (the existence or occurrence of) valuations.
The case of the first person judgment is more complicated. Its causal component is a
judgment of matters of fact. In this respect the first person judgment is on a level with the
third person judgment. Its axiological component, however, is a valuation and not a judgment
about valuations. With regard to this component the judgment cannot be true or false. There is
no 'room' for mistake concerning its truth-value. In this respect the first person judgment of
the beneficial and the harmful is like the first person hedonic or eudaimonic judgment.
Although the first person judgment cannot be false in its axiological component, it can be
insincere. The problem of sincerity of judgments concerning that which is good or bad for a
man is most complicated. It is intimately connected with the problems relating to the notions
of regret and of weakness of will. A few words will be said about them later.
A subject can also make a statement about his own valuations in the past or a conjecture about
his valuations in the future. Such a statement or conjecture is, logically, a third person
judgment. It is true or false both in its causal and in its axiological component.
Whether a judgment is, logically, a first person judgment, cannot be seen from the person and
tense of its grammatical form alone. A man says 'This will do me good'. In saying this he
could be anticipating certain consequences and expressing his valuation of them. But he could
also be anticipating certain consequences and anticipating his valuation of them. In the first
case, the judgment he makes is of the kind which I here call a first person judgment of the
beneficial or harmful nature of things. In the second case, the judgment is (logically) a third
person judgment. The subject is speaking about himself, i.e. about his future valuations.
Sometimes a judgment of the beneficial or the harmful is clearly anticipative both of
consequences and of valuations. Sometimes it is clearly anticipative with regard to
consequences and expressive with regard to valuations. But very often, it seems, the status of
the judgment is not clear even to the judging subject himself. The judgment may contain both
anticipations and expressions of valuations. Perhaps it is true to say that men's judgments of
what is
-111-
good of bad for themselves tend on the whole to be anticipative rather than expressive with
regard to valuations.
The distinction between the apparent and the real good, it should be observed, can be upheld
both for third person and for first person judgments of the beneficial and the harmful. In this
respect judgments of the beneficial and the harmful differ from hedonic and eudaimonic
judgments. (For the two last kinds of judgment the distinction vanishes in the first person
case, i.e. in the genuine value-judgments.) Because of the presence of the causal component in
the judgment, a subject can always be mistaken concerning the beneficial or the harmful
nature of a thing -- even when there is no 'room' for mistake with regard to valuation.
14. A man's answer to the question whether a certain good is worth its price or whether a
thing is beneficial or harmful, may undergo alterations in the course of time. Such alterations
in his judgments can be due either to changes in his knowledge of the relevant causal
connexions or to changes in his valuations. For example: a man attains an end, which he
considers worth while to have pursued, until years afterwards he comes to realize that he had
to pay for it with the ruin of his health. Then he revises his judgment and regrets.
There are two types of regret-situation relating to choices of ends and goods in general.
Sometimes the choice can, in principle if not in practice, be repeated. To profess regret is then
to say that one would not choose the same thing again next time, when there is an
opportunity. But sometimes the choice is not repeatable. The reason for this could be that the
consequences, of which one is aware and which are the ground for one's regret, continue to
operate throughout one's whole life. There is no opportunity of making good one's folly in the
past by acting more wisely in the future. Then to express regret is to pass judgment on one's
life. It is like saying: If I were to live my life over again, I would, when arrived at the fatal
station, act differently.
The value-judgments of regret and no-regret, like hedonic judgments and judgments of
happiness, are neither true not false. But they may be sincere or insincere. A person can say
that he regrets, when in fact he does not, and he can stubbornly refuse to admit regret which
he 'feels'. How is such insincerity unmasked? For example in this way: If a man, after having
suffered the
-112-
consequences, says he regrets his action, but on a new occasion repeats his previous choice,
then we may doubt whether his remorse was not pretence only. He was perhaps annoyed at
having had to pay so much for the coveted thing and therefore said it was not worth it, but at
the bottom of his heart he was pleased at having got it. These are familiar phenomena.
Yet to think that a repetition of the professed folly were a sure sign of insincere regret, would
be to ignore the complications of the practical problems relating to the good of man. A good,
if strongly desired in itself and near at hand, may be a temptation to which a man succumbs,
when the evil consequences are far ahead and the recollection of having suffered them in the
past is perhaps already fading. There is no logical absurdity in the idea that a man sincerely
regrets something as having been a mistake, a bad choice with a view to his welfare, i.e. with
a view to what he 'really' wants for himself, and yet wilfully commits the same mistake over
again, whenever there is an opportunity.
When a man succumbs to temptation and chooses a lesser immediate good, i.e. thing wanted
in itself, rather than escapes a greater future bad, i.e. thing unwanted in itself, then he is acting
wilfully against the interests of his own good. It is in such situations that those features of
character which we call virtues, are needed to safeguard a man's welfare. We shall talk about
them later (in Chapter VII).
That a man can do evil to himself through ignorance of the consequences of his acts or
through negligence is obvious. That he can also harm himself through akrasia or weakness of
will has a certain appearance of paradox. He then, as it were, both wants and does not want,
welcomes and shuns, one and the same thing. When viewed in the short perspective,
'prerequisites and consequences apart', he wants it; when viewed in the prolonged perspective
of the appropriate causal setting, he shuns it. One could say that, if he lets himself be carried
away by the short perspective, then he was not capable of viewing clearly his situation in the
long perspective. Or one could say that, if a man has an articulated grasp of what he wants,
he can never harm himself through weakness of will. But saying this must not encourage an
undue optimism about man's possibilities of acting in accordance with cool reasoning.
-113-
VI
GOOD AND ACTION
1. IN this chapter I shall discuss three typical attributions of goodness: the good act, the good
intention of will, and -- quite briefly -- the good man.
It would seem that 'good' in the phrase 'a good act', unlike 'good' in such phrases as 'a good
knife' or "a good general', does not signify a goodness of its kind. (Cf. Ch. II, sect. 1.) There is
no excellence, which is typical of acts as acts, as there is an excellence of knives as knives
and of generals as generals. But acts can participate in several forms of goodness.
There is first to be noted a connexion between acts and technical goodness. This last, as we
know, is primarily an attribute of agents who are good at some activity. It is, in the last resort,
in the skilful performance of some acts that technical goodness reveals itself. Acts are
sometimes judged good or bad on the basis of the perfection of their performance. 'Well
done', we say of the good performance, and a well done so-and-so we sometimes call a good
so-and-so, e.g. when speaking of a good ski-jump or a good race.
The goodness, which we attribute to an act on the basis of the excellence of its performance, I
shall call the technical goodness of the act. It is an attribute of an act as a member of a kind of
act. It is typically a goodness of its kind. The good ski-jump is as skijump good. If there
existed standards of excellence of performance for all kinds of acts, then one could in a
secondary sense call the goodness of an act's performance the goodness of the act qua act, as a
good golfer might also be called a good sportsman.
It should be noted that technical goodness as an attribute of acts is necessarily an attribute of
act-individuals, i.e. individual per-
-114-
formances of acts of some kind, and never an attribute of actcategories or kinds of act.
2. One of the most common and familiar uses of 'good' is its use in the phrase 'a good (bad,
better, less good) way of doing something'. Here 'doing something' stands for an arbitrary
human act. An act is the bringing about or production at will of a change in the world, e.g. the
unlocking of a door. The change brought about or effected we call the result of the act, e.g.
the fact that a certain door, which was locked, is now open. The 'way of doing' again is some
act or activity which 'leads up to' the result of an act, e.g. the turning of a key and pulling of a
handle, which opens the door. The tie between 'way of doing' and 'thing done' is an intrinsic
connexion between a kind of act or activity and some generic state of affairs. Certain turnings
of a key is a way of unlocking a door, only provided it results in the door's being unlocked.
Value-judgments, which are passed on ways of doing things, are usually comparative
judgments. That one way of doing something is better than another way of doing the same
thing can mean that it is easier or quicker or less expensive or more pleasant or more tidy or is
not connected with certain unwanted side-effects, etc. What 'better' means in the individual
case, depends upon the aim of the doer, whether he wants to do the thing as cheaply as
possible or as quickly or as tidily or what not. This relativity of the goodness of a way of
doing to aims of the agent does not, however, mean that "good' here is an attribute of the
individual performance. The very phrase 'way of doing' indicates generality.
It should be observed that also a bad way of doing a thing is a way of doing it, just as a poor
or a bad knife is still a knife. This means that there is a sense of 'good', in which even a bad
way of doing something is good, viz. 'good' as opposed to 'no good' for the purpose of
effecting a certain result. Ways of doing, which are no good in a certain acting situation are
(usually) not ways of doing the wanted thing at all -- and not bad ways of doing something. In
all these respects the use of 'good' and 'bad' and 'better' as attributes of ways of doing
resembles the use of 'good' and 'poor' (sometimes also 'bad') and 'better' for rating instruments.
3. The attribution of instrumental goodness to ways of doing things must not be confused with
the attribution of that which we
-115-
have here called utilitarian goodness, i.e. roughly speaking usefulness, to acts as means to
certain ends.
Means-end relationships, which are relevant to judgments of utilitarian goodness, are causal
or extrinsic and not logical or intrinsic relationships. (Cf. above Ch. III, sect. 1 and below Ch.
VIII, sect. 5.) An act is good or useful for a certain purpose or with a view to a certain end, if
the doing of this act promotes or favours this purpose or the attainment of this end. It is bad or
harmful if it hinders or counteracts the attainment of the end. For example: to disobey the
doctor's orders may be a hindrance to quick recovery from an illness. It is a bad thing to do
with a view to one's recovery. But we do not, except ironically, call it 'a bad way of
recovering'.
'Good ways of doing things' and 'bad ways of doing things', I would say, are opposed to one
another as contradictories rather than as contraries. But utilitarian goodness and badness as
attributes of acts denote contraries and not contradictories. The good act is useful by
favouring, the bad act detrimental or harmful by counteracting the attainment of the end.
These, clearly, are contrarily related alternatives. Between them fall the acts, which neither
favour nor counteract the attainment of the end.
It is useful to distinguish between the result and the consequences of an act. The result of an
act is that state of affairs which must obtain, if we are to say truly that the act has been done.
For example: the result of the act of opening a window is that a certain window is open (at
least for a short time). The consequences of an act are states of affairs which, by virtue of
causal necessity, come about when the act has been done. (Ch. V, sect. 8; see also the
discussion in Ch. VIII, sect. 5.) For example: A consequence of the act of opening a window
may be that the temperature in the room goes down. The relation between an act and its result
is intrinsic; the relation between an act and its consequences again is extrinsic.
It follows from what has been said that whether an act is good or bad in a utilitarian sense,
depends upon its consequences. Which will be the consequences of an act again largely
depends upon the circumstances under which the act was done. The circumstances may vary
from one individual performance to another of an act of a certain category or kind. For this
reason, attributions of utilitarian goodness to acts will often be restricted to actindividuals.
But they are not necessarily thus restricted. Sometimes the consequences of individual
performances of an act are nearly
-116-
always the same, independently of variations in circumstances. Then utilitarian goodness may
become attributed to the actcategory. For example: medicines and drugs have nearly uniform
effects on the human body. Hence such general judgments become possible as, say, that
taking Vitamin C is good (useful) for preventing colds.
Means are not clearly distinguished in language from ways of doing things. Sometimes, it
would seem, the phrase 'a good (bad) means' means exactly the same as 'a good (bad) way of
doing'. But sometimes 'good (bad) means' refers to some act which is causally connected with
some end. In the first case 'good' in 'good means' connoted instrumental, in the second case
utilitarian goodness.
It is of some interest to note that, whereas technical goodness can be attributed only to
individual acts, viz. as performances, and instrumental goodness only to kinds of act or
activity, viz. as ways of doing, utilitarian goodness may become attributed either to an act-
individual or, sometimes, to an act-category.
The utilitarian goodness of an individual act or of an actcategory depends solely upon a causal
relation to some end of action. It is completely independent of any goodness, which the end
may possess, in addition to being 'a good' by virtue of being an end. There can be good means
to bad ends and bad means to good ends, just as there can be good ways of doing bad things
and bad ways of doing good things.
4. Are acts and activities rated as good or bad according to whether they are wanted
(welcomed) or unwanted (shunned) in themselves?
Action can be rated as wanted or unwanted in itself by the agent who performs it, but also by
various subjects who 'suffer' it, i.e. whose lives are affected by it. The ratings by different
subjects of one and the same individual act or of one and the same category of act can, of
course, be different.
That an act or activity is in itself wanted by the agent means that it is something which he, its
causal requirements and consequences apart, would want to do for its own sake. If an act is in
itself wanted, then its result is a thing wanted in itself. If again an activity is in itself wanted,
then practising it is a thing wanted in itself.
-117-
Practising an activity for its own sake is frequently called by the agent a nice thing or a fine
thing or a lovely thing to do. Sometimes it is also called by him a good thing to do. But in
calling it good he would usually, I think, be implying that he considers this thing good for
him, i.e. beneficial.
That which an agent does is not very often in itself, i.e. its consequences apart, a wanted thing
for another subject. But it is not infrequently in itself unwanted by others. They find it, say,
annoying to hear or disagreeable to watch. Action on the part of other subjects, which we
consider in itself unwanted, is commonly said to be a nuisance (to us). But we would hesitate
to call it bad, unless we consider it positively harmful, obnoxious (for somebody).
By the intrinsic value of acts and activities we may understand their character of being
wanted or unwanted or indifferent in themselves. Accepting this terminology, we seem
entitled to say that judging action good or bad on the basis of its intrinsic value is neither a
common nor a very important kind of valuation.
Common, however, is the rating of acts and activities as good or bad according to whether
they are (thought) beneficial or harmful. The agent of a beneficial act is said to do good to the
subject or subjects for whom the act is beneficial. Similarly, the agent of a harmful act is said
to do harm to the subject(s), for whom the act is harmful. Also the acts themselves are
sometimes said to do good and harm respectively to the subjects. More commonly, however,
they are said to be good for and bad for those whom they affect.
One and the same (individual) act can be beneficial for one but harmful for another subject.
Such acts are sometimes said to be both good and bad. There is no contradiction in this, as
long as we remember the relativity of the beneficial and the harmful to ('suffering') subjects.
An agent can be mistaken in thinking that an act of his is beneficial (harmful) for some other
subject. This he can be, because he is ignorant of the (causal prerequisites and) consequences
of his act. But he can also be mistaken because he is ignorant of the valuation of his act by the
other subject. Perhaps he did something which he was sure his neighbour would welcome.
Perhaps he did this even for the sake of promoting his neighbour's good. Yet, if his neighbour
sincerely resents what was done to him, including its consequences, the act was not
beneficial.
Similarly, a subject can be mistaken in judging of something,
-118-
which has been done 'to him', that it is good (bad) for him. He can be mistaken, because he is
ignorant of the consequences, but also because he fails to anticipate correctly his own
valuation of the causal whole, of which the act in question is the 'nucleus'.
Judgments of the goodness or badness of human acts according to how they affect the good of
various beings, thus share in the precariousness and uncertainties of judgments of things
beneficial and harmful in general.
5. Human acts are perhaps the most important category of things, which are judged good or
bad 'in a moral sense' or 'from a moral point of view'.
Is there a special moral sense of the word 'good'? Is moral goodness a special form of the
good -- on a par with hedonic, technical, utilitarian, etc. goodness?
Moral philosophers often discuss the good and the bad (evil), as though the answer to the
above questions were affirmative. Whether they are prepared explicitly to defend such a view
is usually not clear from the very scanty treatment which those philosophers give to the other
forms of goodness beside the moral. There is a tendency to dismiss the other forms, or some
of them, as ethically irrelevant.
My own view of the matter is roughly as follows: Moral goodness is not a form of the good
on a level with the other forms, which we have distinguished. If it be called a form of
goodness at all, it is this in a secondary sense. By this I mean that an account of the
conceptual nature of moral goodness has to be given in terms of some other form of the good.
(Cf. Ch. I, sect. 8.)
I shall here attempt to give such an account. The form of goodness, in the terms of which I
propose to explicate the notion of the morally good, is the sub-form of utilitarian goodness
which we have called the beneficial. To put my main idea very crudely: Whether an act is
morally good or bad depends upon its character of being beneficial or harmful, i.e. depends
upon the way in which it affects the good of various beings.
There is a prima facie objection to an account of moral goodness in terms of the beneficial. It
can be framed as follows:
The beneficial and the harmful are relative to subjects. If an act is called good on the ground
that it is beneficial, the judgment is incomplete unless we are told for whom it is good
(beneficial).
-119-
Similarly, if an act is called bad on the ground that it does some harm, the statement is elliptic
unless it is added for whom the act is bad. 'Good' when it means 'beneficial' is always 'good
for somebody', and 'bad' when it means 'harmful' is always 'bad for somebody'.
The morally good and bad is not in this sense relative to subjects. Phrases such as 'morally
good for me' of 'morally bad for him' must be dismissed as nonsensical. The fact that an act
does harm to somebody may be relevantly connected with the moral badness of the act. But if
this act is morally bad, then it is bad simpliciter-- and not for some subject, as opposed to
others. And similarly for moral goodness.
There is thus one sense, in which moral goodness is 'absolute' and 'objective' and in which the
beneficial is 'relative' and 'subjective'. This marks an important logical distinction between the
morally good and bad on the one hand and the beneficial and the harmful on the other hand.
Does not this difference between the two 'forms' of goodness doom to failure any attempt to
define moral goodness in terms of the beneficial? I hope to be able to show that this is not so.
6. One and the same individual act, we said (p. 118 ), can affect the good of different beings
differently-be beneficial for some and harmful for other beings. The question may be raised
whether, taking into consideration all such effects, one could form a resultant judgment of the
value of the act. If the answer is affirmative, the further question could be raised whether this
resultant value of the act, its 'overall character' of beneficial or harmful, could be identified
with its moral worth.
The question whether a value-resultant can be formed on the basis of the character of an act of
being beneficial for some and harmful for other beings, is related to the following more
general problem: Can the welfare of one being be 'balanced' against the welfare of another?
Does it, for example, make sense to say that the good which an act did to some person was
greater than the harm it did to another? of to say that the total amount of moderate good which
an act does to several beings, 'outweighs' the great amount of harm which this same act does
to one or a few beings?
It seems to me obvious that, if a value-resultant of the ways in which acts affect the welfare of
various beings can be formed at all,
-120-
this can only be done by judging the beneficial or harmful nature of the act from the point of
view of the welfare of a collectivity, of which the various beings concerned are the individual
members. Such judging of acts is, as a matter of fact, often attempted. It is attempted, e.g.,
when we argue that it is better that one man's interests are sacrificed than that all the members
of a community shall suffer. Or when we say that it is better that one man is thrown overboard
than that the whole crew shall perish. We are then, as it were, saying: this is how the
community would prefer to have it. We compare the community to a man who ponders
whether he should undergo a painful operation or do something he dislikes for the sake of his
subsequent health or well-being.
Are such arguments logically legitimate? The answer is tied to the problem whether the
notion of the good (welfare) of a collectivity or community of men is logically legitimate. I
am not suggesting that the answers must be negative. But I am sure that the conditions of
estimations of value from the point of view of the good of a community are extremely
complicated, and also that the appeal to the welfare of a collecdvity over and against the
welfare of some of its members is often misused in practice. We are here touching upon a
major problem in the ethics of politics. I shall not discuss this problem further in the present
work. I leave it open.
In the subsequent discussion I shall disregard the possibility of forming a judgment as to the
'overall' beneficial or harmful character of an act. I shall discuss the notion of moral goodness
from what could also be called the point of view of the human individual, as distinct from the
point of view of the human community. On the logical complications connected with the
second point of view I shall not here touch, nor on the question whether the two points of
view can be brought into harmony with one another.
7. We next put forward the following suggestions of how the moral value of a human act may
be considered a 'function' of the way in which this act affects the good of various beings
favourably or adversely:
an act is morally good, if and only if it does good to at least one being and does not do bad
(harm) to any being; and
an act is morally bad, if and only if it does bad (harm) to at least one being.
-121-
The suggestions are open to a number of objections.
First, it may be objected that an act could not be called morally good on the ground that it is
beneficial for the agent himself-even if it does not hurt anybody else. Similarly, it is at least
doubtful, whether harming oneself could be considered morally bad.
If one accepts one or both of these objections as valid, one takes the view that moral action is
essentially 'social'. On this view, that part of a man's action which affects solely his own good
is morally irrelevant.
There is a certain inclination, it seems to me, to say that harming oneself is morally bad,
though doing good to oneself is not morally good. It may be suggested, however, that the
foundation of this inclination is the fact that by harming himself a person can hardly fail to
become a nuisance to or a burden on his fellow humans. According to this suggestion, the
moral badness of doing bad to oneself consists in the bad which the agent (indirectly) causes
to others.
We shall not here take a stand on the issue, whether action which solely affects the acting
agent's own good is morally relevant or not.
Another objection to the suggested definition of moral goodness runs as follows: An act
which does good to some and bad to no being need not have been done for the sake of doing
good. The agent may not even be aware of the beneficial nature of what he did. Would it not
in either case be absurd to label this act morally good?
A similar objection may be raised against our proposed definition of moral badness. If the
agent is not aware of the harmful nature of what he did, is his act then morally bad?
These objections mean that our proposed definition of moral goodness was too 'lenient' and
our proposed definition of moral badness was too 'severe'. The first was too lenient because it
was compatible with the possibility that action can be morally good, though no good is
intended or even foreseen by the agent. The second again was too severe because according to
it action can be morally bad even when no bad is intended or foreseen.
There is, however, also an objection to our proposed definition of moral goodness on the
ground that it is too severe, and to the definition of moral badness on the ground that it is too
lenient.
For, under certain circumstances, is not an intention to do some good enough to warrant the
moral goodness of the act, even if
-122-
no good actually results to anybody? And is not similarly an intention to do some bad enough
to label the act as morally bad, even if no harm actually results?
Accepting these objections thus means that our proposed definitions of morally good and bad
acts were at once too lenient and too severe.
I think we must accept these objections--or some of them at least. The proposed definitions
cannot be regarded as successful attempts to catch hold of the 'essence' of moral goodness and
badness. The chief reason why they fail is that they make the moral quality of an act
independent of intentions in acting and of the foreseeing of good and harm to other beings.
Before we revise our proposals for defining moral goodness and badness, some words must
be said on the concept of intending.
8. How is intention in acting related to the foreseeing of consequences?
In order to answer the question, we shall have to observe that intention is primarily connected
with results of action--and not with consequences. An intention is an intention to do
something. That which is intended, the object of intention, is the result of an act.
Suppose that I open a window with the intention, as we say, of cooling the room. The cooling
of the room is a consequence of the opening of the window. Is not here the object of intention
the consequence and not the result of my act? I propose to answer the question as follows:
If I open a window with the intention of cooling the room and the temperature in the room
goes down as a consequence of my manipulations with the window, then the question 'What
did I do?' can also be answered by 'I cooled the room'. Cooling the room is something I can
do. There is an act of cooling the room. Its result is that the temperature in the room is now
lower than it was before. It is a different act from the act of opening the window, to which it
has a causal, and not an intrinsic, relation. The act of cooling the room would be different
from the act of opening the window, even if opening the window were the only means to
cooling the room. They are different because the result of the one is a consequence of the
result of the other. But the activity which I display in performing the two acts, i.e. the
manipulations with the
-123-
window, is the same in both. (In this sense the two acts could be said to 'look' the same.)
Suppose a person intends to cool the room and, with a view to this, opens the window. And
suppose that he succeeds in making his intention effective. Shall we then say that he also
intended to open the window? We would certainly call his act of opening the window
intentional. We may wish to call it intended, but we may also wish to reserve this term for
acts, the results of which have the character of ultimate ends in acting. Usage seems to be
somewhat vacillating at this point, and we need not force ourselves to a decision. But we shall
decide to call any act, which is done for the sake of or in order to attain some end, intentional
or intended.
Be it observed in passing that 'intended' is used both of the act and of the state of affairs which
is its result, whereas 'intentional' is normally used only of the act. One suggestion could be
that the results both of intended and intentional acts be called 'intended'. I shall adopt this
suggestion.
The man who opens the window in order to cool the room, can rightly be said to intend two
things 'at once' or to have two intentions -- one to have the window open and another to have
the temperature lower.
Although everything, which is done for the sake of an end, is also intended (as well as the
end), it is certainly not the case that everything, which is a consequence of action, is also
intended. If the agent, at the time when he is acting, does not foresee the consequence (or at
least realizes the 'serious possibility' that it will happen), then he can, in a sense, not even be
rightly said to have done the consequent thing. 'Look what you have done? we sometimes say
of such cases -- particularly when we consider the consequences undesirable. This is said in
order to draw the agent's attention to a causal connexion of which he was not aware. Once he
is aware of the connexion he can not, on a new occasion, do the first thing without also doing
the second.
It is here appropriate to make a distinction between foreseen and rightly foreseen
consequences of action. Something is a foreseen consequence, if the agent at the time of
acting knows or believes that this thing will happen as a consequence of his action. A
foreseen consequence will be called rightly foreseen if it actually happens (as a consequence
of action). Agents sometimes foresee consequences of their action which never come true.
-124-
I shall say that everything which is a rightly foreseen consequence of one's action, is also a
thing done. But I shall not say that everything which is a foreseen consequence of one's
action, is also a thing intended. Suppose our man opens a window and that, as a consequence,
the temperature in the room goes down. When asked why he opened the window, he answers
that he did. it in order to hear the birds sing. Let us assume that he foresaw that the
temperature was going to sink--or at least was aware of the possibility. Then he could not say
that he lowered the temperature unintentionally. But to say that he did not lower the
temperature unintentionally and to say that he intentionally lowered it, is not--on the ordinary
understanding of the words -- to say the same. Of everything which is a rightly foreseen
consequence of my action, I can say truly that I did this and that I did not do it
unintentionally. But only of such consequences of my action, which were also ends of action--
intermediate or ultimate--can I say truly that I intended them or did them intentionally. That
consequences are ends entails that they are foreseen--but not that they are rightly foreseen.
This much will have to suffice about the notions of intention and foreseen consequences of
action, and the related notions of the intentional, the unintentional, and the not unintentional.
9. Can intentions possess an excellence of their kind, i.e. as intentions? Intentions can be firm
or vague, strong or weak. To say of a man that he had 'good intentions' or 'the best of
intentions' to do a certain thing is, it seems, sometimes another way of saying that he was
firmly determined to do this. We should not ordinarily, however, call a vague or weak
intention to do something a 'bad intention'.
When 'good' is used as an attribute of intentions to indicate firmness of determination or
strength of will, it can perhaps be said to connote a goodness of its kind. This is one use of the
word 'good'. But it is certainly not that use of the word which we have in mind, when we
speak of intention as the bearer of moral value or as a component in the moral valuation of
acts.
'Good' and 'bad' as moral attributes of intentions do not connote goodness or badness of its
kind, i.e. of intentions as intentions. It would seem, moreover, that 'good' and 'bad' as moral
attributes of intentions are in an important sense secondary. By this I mean
-125-
that the primary attribution of 'good' and 'bad' here is not of the intention as such, but of the
objects of intention, the intended results of action. Basically, 'good intention' is intention to do
(some) good, and 'bad (evil) intention' is intention to do (some) bad or evil. The problem is
how to give a satisfactory formulation to the dependence of the value of intention on the value
of the intended.
Complications are caused here by the fact that an intention may have several objects, and by
the fact that everything which is done in order to attain an intended end of action is also
intended. Evidently, these complications cannot be ignored in an attempt to assess the moral
worth of intentions. Be it said in passing that Kant's doctrine of the good will seems to me to
suffer badly from Kant's ignoring the 'multiplicity of intention', which exists thanks to the fact
that human acts have causal relationships of which the agent, in acting, is seldom totally
unaware.
By an intended good I shall understand an end of action which is judged by the agent to be
beneficial for some being. By an intended bad I shall understand an end of action which is
judged by the agent to be harmful for some being. By a foreseen bad I shall understand a
result or foreseen consequence of action which is judged by the agent to be harmful for some
being. Every intended bad is also a foreseen bad, but not conversely.
An intended good is thus an intended state of affairs, the production of which is thought to
affect the good of some being favourably. That this state of affairs is produced and that it
affects the good of a being favourably are two logically different things. The first may depend
upon the agent alone. The second also depends upon the valuation of the being(s) affected by
the agent's action. By calling the state of affairs an intended good we thus attribute two
objects of intention to the agent. The first is the state of affairs itself. I shall call it his factual
object of intention. The second is that some being(s) will value this state of affairs as good for
him (them). I shall call this the agent's axiological object of intention.
An intended good or bad can fail to materialize. This can happen for two chief reasons. The
intention may fail in regard to its factual object. Or it may fail in regard to its axiological
object.
When the intention fails in regard to its factual object, we sometimes contrast the 'good
intention' with the 'poor performance'.
-126-
A discrepancy between a professed intention and the actual performance may make us doubt
whether the intention was there at all or how firm or strong or serious it was. But if no such
doubts occur, mere failure with regard to its factual object would not, it seems, be considered
relevant to the question of the moral worth of the intention. One of the reasons for thinking
that the moral worth of action resides in the intention alone, is probably the idea that this
value must not become affected by adversities in acting or by mistakes in foreseeing
consequences. This in no way contradicts the view, which I think we must in any case accept,
namely that the moral value of the intention depends upon the value of the thing intended.
There is, however, also the case to be considered when an intention to do good reaches its
factual object but fails in regard to its axiological object. We did something for the sake of
promoting our neighbour's good, for example, and everything went exactly according to our
plans -- except that our neighbour strongly resented what we did, thoroughly disliked it,
perhaps even badly suffered from it. Should such discrepancy between apparent and real
value, as we could call it, influence our judgment of the goodness of intention?
If by good and bad things we understand things beneficial or harmful, as defined in this
inquiry, then to mistake the apparent value of something which has been done for its real
value is to make a false judgment or conjecture either about the consequences of the achieved
result of action or about the valuation of this result and its consequences by some subject(s)
or about both these things. In a first person judgment 'I like this', 'I would rather have this than
be without it', 'No, thank you', the subject is valuing something. Then there is no possibility of
mistake and no room for a discrepancy between apparent and real value either. In a third
person judgment 'This will do him good', or 'This is bad for him', or 'He will like it', we are
not valuing but saying what we think that the valuations of subjects are or will be. As far as
the value of his acts for others is concerned, the agent's judgment will necessarily be a third
person judgment and thus nota(genuine)value-judgment. A mistake on his part concerning
this value will therefore be a mistake concerning empirical matters of fact.
If we want the moral value of intention to be independent of intellectual mistakes in judging,
then we must in the name of
-127-
consistency admit that mistakes as to the value of the factual objects of our intentions do not
'maculate', i.e. spoil the goodness of our intentions. 'He meant it well', we often say when a
mistake as to value (valuation) has occurred, and since he meant it thus it may seem illogical
to blame him morally for what he did.
But even if, on the ground of the goodness of their intentions, we morally excuse persons for
some evil they have done, we nevertheless blame them for ignorance--either of consequences
of action, or of the beneficial or harmful nature of things, or of both. 'You should have known
that this is not the way a man wants to be helped; by what you did you only managed to hurt
his feelings', we say of many a case of intended beneficiality. Because of ignorance, much bad
is done for good motives. This is an aspect of the matter to which moralists in a Kantian spirit,
as far as I can see, have habitually paid but little attention. Yet it is an important aspect of the
moral life -- or at least of action affecting the good of man. Paying due attention to this aspect
need not, however, influence our view as to what constitutes the goodness of intention. But it
must lead us to realize clearly that intention to do good is, by itself, of rather limited
utilitarian value for the promotion of the welfare of man and of human collectivities.
10. An agent, who intends to do good to some being, can do so either for the sake of
promoting the good of that being or for some other reason. When doing good to somebody is
intended solely for the sake of promoting the good of that being, I shall say that good is
intended for its own sake.
I now propose the following definition of morally good and bad intention in acting:
the intention in acting is morally good, if and only if, good for somebody is intended for its
own sake and harm is not foreseen to follow for anybody from the act; and
the intention in acting is morally bad, if and only if, harm is foreseen to follow for somebody
from the act.
I shall not here further discuss the question whether we may include or must exclude the case
that the subject, whose good the act is foreseen to affect, is the agent himself. (Cf. sect. 7.)
That the intended good must be intended for its own sake, if the intention in acting is to be
called morally good, seems fairly obvious. A master who takes good care of his servants, in
order
-128-
that they shall be fitter to work hard for him, cannot be said to have a morally good intention
in his treatment of his servants. But, if from an attitude of gratitude or love he is anxious to
see his servants flourish or thrive, then we may attribute moral goodness to his intentions.
It seems obvious, too, that the intention in acting is deserving of moral blame, not only on the
ground that some harm is intended, but already on the ground that some harm is foreseen in
acting. Suppose a person opens a window in order to hear the birds sing. Thereby he cools the
room and causes another person, who is present, to catch a cold. He is aware of the fact that
opening the window will cool the room and also aware, let us assume, of the serious
possibility that the other person will catch a cold. If 'with a clean conscience' he can declare
that he opened the window in order to hear the birds and that this was his only end, the only
thing he intended beside the opening itself of the window, then the cooling of the room and
the cold which the other person caught were foreseen conseqnences of his action but not
things he can be said to have intended to achieve. Yet it would be reasonable to blame him for
having opened the window, since he realized at the time of doing it that he was exposing
another person to danger. What he intended to do was morally blameworthy, because of the
foreseen harmful consequences, and I see no reason why we should not therefore also say that
his intention in acting was morally bad. But we could not rightly say that his intention in
acting was malicious. This we can do only when the foreseen harm of his intended action was
also intended and not only foreseen.
Some ethicists may be of the opinion that goodness or badness of intention is the only thing
that is morally relevant in action, and that there is no need to distinguish between the moral
value of the intention and the moral value of the act. It seems to me, however, that a
distinction can be sensibly made.
If the intention in acting is morally good but the good, which is for its own sake intended,
fails to materialize--either because the intention fails in regard to its factual or in regard to its
axiological object--then the act is not morally good (but morally neutral). If, on the other
hand, the intended good materializes, then the act too may be called morally good.
Similarly, if the intention in acting is morally bad but the harm, which is foreseen, does not
come about, then the act is not morally
-129-
bad (but morally neutral). If, on the other hand, the foreseen harm comes about, then the act
too is morally bad.
It is a noteworthy asymmetry between moral goodness and moral badness that the first
presupposes that some good should be intended and, moreover, intended for its own sake,
whereas the second only requires that some bad should be foreseen to follow from the act.
It is characteristic of the logical complications of the concepts, which we have been
discussing, that the notion of a morally good or bad act is secondary to the notion of a morally
good or bad intention (will) in acting, which in its turn is secondary to the notion of a good or
bad, i.e. beneficial or harmful, thing.
11. In our proposed definition of a morally bad act is tacitly presupposed that in every
situation there is a course of action open to the agent, from which he foresees no harm to
anybody. The presupposition, in other words, is that the agent is always 'free' to choose a
course of action which is not morally bad.
(This presupposition, be it observed, is not that there always is a course of action from which
no harm actually follows. Harm which follows, but which could not have been foreseen at the
time of acting, is not relevant to the question of the moral quality of the act.)
However, the presupposition which we just mentioned is not always fulfilled. Cases may
occur in which the agent foresees harm to some being from whatever course of action he
chooses. Such cases are not common. But they happen. Their rarity does not nullify the
importance which their gravity gives to them.
When harm is (rightly) foreseen to follow from whatever an agent does, I shall say that harm
is unavoidable. This 'absolute' notion of unavoidable harm must not be confused with the
'relative' notion of unavoidable harm. By the second I mean harm which cannot be avoided, if
some particular end of action has to be reached.
It is a logical characteristic of the type of acting-situation, which we are now considering, that
there should be some act open to the agent which is such that bad is foreseen to follow both
from doing and from forbearing this act. Forbearing to act, it should be observed, is also
action. Forbearing to act can have consequences just as well as acting can have. The
consequences of forbearance is the happening of those things which action would have
prevented.
-130-
What one would wish to say from a moral point of view of the type of case under discussion
is, I think, the following: In a situation when the agent foresees harm both from doing and
from forbearing the very same thing, at least one of the two, acting or forbearing, must be
morally excusable. If the harm, which is foreseen to follow from doing, is less than the harm
which is foreseen to follow from forbearing, then acting is morally excusable, i.e.is not the
doing of a morally bad act. If again the harm which is foreseen to follow from forbearing is
less, then forbearing is morally excusable. If, finally, the foreseen harm from acting equals the
foreseen harm from forbearing, either course of action is morally excusable.
We could call this the rule of minimizing unavoidable harm. It determines under which
conditions the causing of bad to some being can be morally excused. The harm, which it
exculpates, is unavoidable in the 'absolute' sense. Harm, which is in the 'relative' sense
unavoidable, can, on my view, never be morally excused. To argue that it could be excused, is
a form of arguing that 'the end justifies the means. This seems to me the very prototype of
immoral argument.
The rule of minimizing unavoidable harm has an obvious plausibility. Yet it takes for granted
that a difficult problem can be solved when the rule has to be applied. This is the problem of
determining ('measuring') the relative magnitude of harm to various beings. It is not certain
that this problem has a significant solution in every case. Let us consider the problem in the
light of an example.
A man x can save either the man y or the man z but not both from an impending disaster. y
and z are, say, wounded and x can carry one of them at a time but not both of them at once to
safety. x foresees that, if he carries away one of the men, the disaster will in the meantime
reach the place and consume the remaining man.
We can describe x's possible action in the terms of two choices. His first choice is this: Shall I
save one of the two or shall I leave both to perish? If he chooses the second, then by his
chosen course of action he, in a sense, becomes responsible for the death or disaster of one
man. (x's action is then forbearance.) Or, strictly speaking: x becomes responsible for the
disaster of one more man than would have perished, had x acted differently.
Assume that x chooses to do his best and save one of the two. We would all agree, I think,
that this choice of his is morally right,
-131-
and that a choice to abstain from action would have been morally wrong. To let both perish is
to cause more bad than to leave one to perish. The general principle for comparing the relative
magnitudes of harm can here be formulated as follows: That all of a number of men suffer
some harm is worse (a greater bad) than that some only of these men suffer the same harm. (It
is essential that the smaller group of people, who suffer harm, should be a subclass of the
greater group of people, and that the harm which each one of them suffers as a member of the
first group is the very same harm which each one of them suffers as a member of the second
group.)
So far the case seems clear. But, having chosen to save one of the two rather than to leave
both to suffer, x is faced with a second choice: Shall he save y or z? If he chooses the first,
then as a consequence of his chosen course of action z will suffer. If he chooses the second, y
will suffer. x, we assume, foresees this quite clearly. x takes y and carries him to safety.
Disaster befalls z.
Can this second choice of x's be justified with reference to the rule of minimizing unavoidable
harm? There are two ways in which the justification may be attempted. One is to argue that
the harm, which follows from leaving z, is neither greater nor less than the harm which
follows from leaving y, and that hence x is 'morally free' to choose between saving y or saving
z. The other is to argue that more harm follows from leaving y than from leaving z and that x
hence is 'morally bound' to choose to save y. Both ways of arguing have to cope with similar
logical difficulties. Two examples will be given to show what these difficulties could be and
how one might try to cope with them:
y is the commander of a group of men, of which x and z are members. x argues: it is worse for
the group to lose y than to lose Z.xis then contemplating the consequences of his action from
the point of view of the good (welfare) of a group (community) of men. His argument is that
it is preferable (better) for the group to have ywithout Z than to have z without y.
y is head of a family, z is single. x argues: If y is left to suffer or die, a number of other people
will suffer heavily too. If z is left, others will not be seriously affected. Assuming that the bad,
which y and z are facing, is an equal bad for both, then it is a greater bad that beside y also
others should suffer harm than that no one in addition to z himself should suffer.
-132-
The first way of arguing proceeds on the assumption that an appeal to the good of a
community of men over and against the good of its individual members is possible. The
second argument assumes that it is possible to 'balance' the good of two or more beings
against one another. Both assumptions were briefly discussed in section 6. Of the second we
said that it, too, is based on the possibility of appealing to the good of a community over and
against the good of its individual members. The logical nature and moral character of this
appeal, however, will not be further discussed in the present work. What has been said in this
section illustrates the urgency of such discussion.
12. We conclude this chapter with some brief remarks about the use of the word 'good' as an
attribute of a man.
If there existed a purpose which were essentially associated with man as a kind, then one
could by the phrase 'a good man' understand a human individual, who serves this purpose of
the kind well. Some may think that there is such an essential purpose of man, e.g. that men
exist to serve the purposes of their Creator. We shall not here discuss this view.
Men often become accidentally associated with purposes, i.e. they are needed and used to
serve the purposes of other men or of human institutions. The phrase 'he is a good man' is
very commonly used to mean that a man fits or serves some such purpose or task well. The
man, who is thus judged instrumentally good, is usually also in some respect technically
good, viz. as a member of a professional class. Such instrumental or technical goodness of a
man, however, does not mean that he were, as a man, good, even if it is natural to call him a
good man.
The idea that there is an activity essentially associated with man as man, may be considered
inherently more plausible than the idea that there existed some purpose thus associated with
him. The function proper to man, Aristotle thought
1
is activity in accordance with a rational
principle or life according to reason. The better a man performs this proper function of his, the
better he is, as a man. This is a clear use of 'good' and 'better' to attribute to a man a goodness
of its kind. I shall not here discuss Aristotle's idea. It seems to me that, even if it were true to
say that there is an activity which is essential to man, e.g. reasoning, it would be
____________________
1
See especially the discussion in EN, Bk. I, Ch. 7.
-133-
doubtful whether this activity is of the sort at which a being can be said to be good or bad. To
say that a man lives or does not live in accordance with reason is vague, but we understand
roughly what it means. But to say that a man is 'good at' living in accordance with reason,
would seem to require a special interpretation of 'good at', which conceptually distinguishes
this case from other cases of that which I have here called technical goodness.
When the phrase 'a good man' is used, not in the sense of instrumental or technical goodness
but with a moral tinge, it is related to the notions of doing good and of having good intentions.
But it has no clear and distinct relationship to these notions.
A man who is intent on doing good is often called benevolent. A man can be a 'benefactor'
without being benevolent, e.g. if he does good to others mainly for the sake of promoting his
own social prestige. A true benefactor must be a man who does good (to others), but he need
not be that which is ordinarily called a good man.
I am not suggesting that there is a common and important use of the phrase 'a good man' to
mean the same as 'a benevolent man'. But I think it is true to say that when the phrase 'a good
man' is used with a so-called moral meaning, it is related to our idea of a benevolent man. It is
of some interest in this connexion to notice that the opposite to a benevolent man, i.e. a
malevolent man or a man who is intent on doing evil or mischief, is quite commonly and
naturally also called, in a moral sense, 'a bad man'.
One affinity between the morally relevant notion of a good man and the notion of a
benevolent man is in any case that both notions have to do with features of human character.
A man may do some bad acts and even entertain some evil intentions--and yet be a good man.
But the bad he does or intends must count as an occasional aberration. Or it must have some
special excuse. If we were asked how much evil the good man can be 'allowed', we could, of
course, not answer by giving an exact measure--'hereunto and no further'. But we could give
an inexact and yet significant measure by saying that the bad he does or intends must not
affect our judgment of his character. A good man may do some mischief, or revenge a wrong
which he has suffered from another man, or tell a lie. But he cannot be mischievous or
revengeful or untruthful.
-134-
With this last remark we are also touching upon one of the differences between goodness and
benevolence as attributes of men. A benevolent man is not necessarily a virtuous man, and he
may be lacking in a sense of justice. Virtue and justice are two prominent features in our
picture of moral excellence. Until we have discussed them, our notion of the good man will
remain insufficiently clarified.
-135-
VII
VIRTUE
1. VIRTUE is a neglected topic in modern ethics. The only fullscale modern treatment of it,
known to me, is by Nicolai Hartmann. When one compares the place accorded to virtue in
modern moral philosophy with that accorded to it in traditional moral philosophy, one may
get the impression that virtue as a topic of philosophic discussion has become obsolete,
outmoded. This impression may gain additional strength from the fact that traditional
discussion has -- with rather few notable exceptions -- followed the footsteps of Aristotle
without much variation or innovation or controversy. Kant's famous dictum that formal logic
had made no appreciable progress since Aristotle, could be paraphrased and applied -- with at
least equally good justification -- to the ethics of virtue.
Kant thought that the reason why there had been so little progress in logic (as he saw it), was
that Aristotle had accomplished most of what there was to be done in the field. As we know
now, Kant was badly mistaken. It would be unwise to prophesy a renaissance for the ethics of
virtue, comparable to the renaissance which we have in this century witnessed in logic. But I
think the time has come when the impression which the discussion of virtue in traditional
moral philosophy conveys to the modern spectator, should no longer be that of something
accomplished or obsolete, but rather that of a subject awaiting fresh developments.
The relative neglect of the discussion of virtue is certainly connected with the predominance,
for a long period, of that which could be called the (purely) axiological and deontological
aspects of moral philosophy. Good and bad and evil are value-terms. Right and wrong and
duty are normative terms. But courage, temperance, and truthfulness we would not ordinarily
call value-concepts nor
-136-
normative concepts. Some people would call them psychological concepts, but this too is not
a very fitting name. (Cf. Ch. I,sect.4.) sect. 4.)
The discipline known as General Theory of Value, now much in fashion, tries to cater among
other things for those aspects of traditional moral theory which are concerned with value --
the axiological aspects. Since virtue is not prominent among them, the ethics of virtue has
stayed aside from these developments towards greater generality. There is another discipline
of recent origin which also has a wider scope than traditional ethics and within which a
discussion of virtue could claim for itself a natural place. This is the discipline sometimes
called Philosophical Anthropology. Yet it, too, as far as I know, has not so far paid much
attention to the topic of virtue.
2. The Latin word virtus, of which 'virtue' is a derivation, has a rather more restricted
connotation than the English virtue. Its original meaning is perhaps best rendered into English
by words such as 'manlihood' or 'prowess' or 'valour'. The Greek arete again, which it has
become customary to translate by 'virtue', has a much wider connotation than the English
word. Its primary and original meaning is the excellence or goodness of any thing whatsoever
according to its kind or for its proper purpose. The word 'virtue' too is often used with this
meaning. We tend, however, to regard this as a secondary or analogical use. We have an idea
of what could be meant by the virtues of a good knife. We easily say such things as that so-
and-so had all the virtues of a great general -without wishing to call him 'virtuous' in the
common understanding of that word. Then we are using virtue in a secondary sense -- as the
Greeks would have used arete in a primary sense.
When, however, we call courage, temperance, generosity, or justice virtues, we are using the
word 'virtue' very differently from that meaning of arete, which refers to an excellence of its
kind. To see this clearly is, I think, of some importance. Aristotle, I would suggest, did not see
quite clearly at this point. He was misled by peculiarities of the Greek language into thinking
that those features of human character, which are called virtues, are more closely similar than
they actually are to abilities and skills, in which a man can possess that which I have
previously called technical goodness.
-137-
I shall not maintain, however, that our word 'virtue', when used with a primary meaning,
stands for one sharply bounded concept. For this it obviously does not do.
There is first to be noted a feature of our use of the word 'virtue', which comes near to being
an ambiguity.
1
There is one meaning of 'virtue', which admits of a plural, 'virtues'. This
meaning is in question, for example, when we call courage a virtue. There is another meaning
of 'virtue', which lacks the plural. This is (usually) in question, when virtue is contrasted with
vice, or when -- as is sometimes done -- to do one's duty is said to be virtue. Virtue in this
second sense comes near to being an axiological or a normative attribute -- or a mixture of
both.
2
It has definitely a moral tinge. It is related to goodness and rightness and to that which
in the Bible is called righteousness. With this meaning of 'virtue', be it observed, we are not at
all concerned here. One could also say that we are here not concerned with the meaning of
'virtue', but with the meaning of 'a virtue'. Therefore we are here also dealing with that
meaning of 'virtuous', which is the display or practising of virtues, and not (directly) with that
which is virtuous as opposed to vicious conduct or character.
Although our notion of a virtue obviously possesses a greater 'logical homogeneity' than the
notion of an arete with the Ancients, not even it can be said to be sharply bounded. Courage,
temperance, and industry, for example, seem to me to be virtues in a rather different sense
from, say, piety or obedience or justice. I am not at all certain that those features which, as I
view it, mark the first three as virtues also hold good for the second three.
Considering both the unstable usage of the word and the unsatisfactory state of the subject,
the task before us could be described as one of moulding or giving shape to a concept of a
virtue. We cannot claim that everything which is commonly and naturally called a virtue falls
under the concept as shaped by us. But, unless I am badly mistaken, some of the most obvious
and uncontroversial examples of virtues do fall under it. It is therefore perhaps not vain to
hope that our shaping process will contribute
____________________
1
I am indebted to Mr. F. Kemp for having drawn my attention to this distinction.
2
Kant uses Tugend mainly to mean virtue in this sense. Kant also calls it Sittlichkeit or
moralische Gesinnung. The best English translation is perhaps 'morality'.
-138-
to a better understanding at least of one important aspect of the question, what a virtue is.
3. As a first step towards shaping a concept of virtue I shall say that a virtue is neither an
acquired nor an innate skill in any particular activity. 'To be courageous' or 'to show courage'
do not name an activity in the same sense in which 'to breathe' or 'to walk' or 'to chop wood'
name activities. If I ask a person who is engaged in some activity, 'What are you doing?' and
he answers, 'I am courageous; this is very dangerous,' he may be speaking the truth, but he is
not telling me what he is doing.
The lack of an essential tie between a specific virtue and a specific activity distinguishes
virtue from that which we have called technical goodness. We attribute technical goodness or
excellence to a man on the ground that he is good at some activity. But there is no specific
activity at which, say, the courageous man must be good -- as the skilled chess-player must be
good at playing chess and the skilful teacher must be good at teaching. There is no art of
'couraging', in which the brave man excels.
Nor is a virtue a goodness of the sort which we attribute to faculties (or organs) and which, as
we have said before, is related to technical goodness. The virtues are not to be classified along
with good sight or memory or ratiocination.
One difference between virtues and faculties is that virtues are acquired rather than innate. In
this respect virtues resemble technical excellence. Both of the, in some respect, virtuous man
and of the man who is skilled in an art, it makes sense to ask, 'What has he learnt?' The
significance of the question is not minimized by the fact that men are by nature more or less
talented for various arts and also more or less disposed towards various virtues. Sometimes a
man can be truly said to be skilled or virtuous without any or much previous education and
training. But this is the exception rather than the rule. Contrarywise, a man of good sight has
not learnt to see well and a man of good memory is not commonly said to have learnt to
remember well. This must not be interpreted as meaning that the faculties of man were not to
some extent capable of being improved by training.
Yet there is another regard in which virtues are more akin to faculties than to skill in
activities. This similarity is their relation to the good of man.
-139-
To do that on which one is keen of which one does well is a source of active pleasure and
may, on that ground, be a 'positive constituent' of a man's good. This is one reason why it is
important that a man should, if possible, be trained in a profession which answers to his
natural gifts and interests. But none of the various professions which a man may choose, and
the various skills in which he may be trained and may come to excel, is, by itself, needed for
the good of man.
With the faculties it is different. (Cf. Ch. III, sect. 7.) Sight and hearing, memory and
ratiocination are part of a normal individual's equipment for a normal life. Loss of one of the
faculties can be disastrous, weakness of some or several of them is usually to some extent
detrimental to the well-being of a man. This, of course, does not exclude that one can make up
for the loss or weakness of some faculty. And some men are more dependent for their well-
being upon one faculty than upon another. For example: some men need good eyes more than
other men, and some men need the use of their eyes and ears more than the use of their brains.
Such individual differences depend upon differences in profession and upon other
contingencies of life.
Virtues, like faculties, are needed in the service of the good of man. This usefulness of theirs
is their meaning and purpose, I would say. How virtues fulfil their natural purpose I shall try
to show presently. Be it here only remarked that, just as the contingencies of life can make a
man more dependent upon some of his faculties than upon others, in a similar manner can
contingencies make the acquisition of one virtue be of the utmost importance to a man and the
possession of another of relatively little value. Sometimes the factors, which determine the
relative utility of the various virtues, have the character of contingencies in the history of
society or of mankind rather than in the life of individual men. In a warlike society, such as
perhaps were ancient Sparta or Rome of the Republic, courage is more important for the
individual man than, say, chastity or modesty. There is then more demand, so to speak, for
brave than for modest men. The different ratings, which moralists of different ages and
societies have given to the various virtues reflect, sometimes at least, such contingencies in
the conditions under which these moralists themselves have lived. (Cf. sect. 9.)
-140-
4. It is important to distinguish between acts and activities. Acts are named after that which I
have called the results of action, i.e. states of affairs brought about or produced by the agent
in performing the acts. Acts leave an 'imprint', as it were, on the world; when activities cease,
no 'traces' of them need remain. Lighting a cigarette, e.g., is an act; it results in a cigarette
being lit. Smoking is an activity.
Acts named after the same generic state of affairs are said to form an act-category. The
individual performances of acts again we have called act-individuals. The lighting of a
cigarette by a certain agent on a certain occasion is an individual act of the category labelled
'lighting a cigarette'.
As a first step towards shaping the notion of a virtue we said that to the specific virtues there
do not answer specific activities. As a further step we shall now say that to the specific virtues
there do not answer specific act-categories either.
Of the man of whom a certain virtue is characteristic, acts of a certain kind are also
characteristic. But these acts do not constitute an act-category in the sense here defined. They
are named after the virtue from which they spring, and not after the states in which they
result. There is nothing wrong about saying that a courageous man often does courageous
acts. But it is not very illuminating. It becomes misleading if it makes us think that we could
define the virtues in terms of certain achievements in acting.
There are at least two reasons why the question, which acts are courageous, cannot be
answered by pointing to results achieved in the successful performance of courageous acts.
The first is that the results of all courageously performed acts need not have any 'outward'
feature in common. Killing a tiger and jumping into cold water can both be acts of courage,
though 'outwardly' most dissimilar. No list of achievements could possibly exhaust the range
of results in courageous action. The second reason why brave acts do not form an act-category
is that the result of any courageous act could also have been achieved through action which
was not courageous. Not even to have killed a tiger is a sure proof that a man is courageous.
What is true of courage in the said respect is also true of the other virtues. Virtuous acts
cannot be characterized in terms of their results, and therefore virtues not in terms of
achievements. We can express this insight in many ways. We could say that the
-141-
notion of a brave, generous, temperate, etc. act is secondary to the notion of a brave,
generous, temperate, etc. man. Or that virtuous action is secondary to virtue. Or that a virtue
is an 'inner' quality of an agent and of his acts, and not an 'outer' feature of his conduct. But all
these modes of expression are also in various ways misleading, and they should therefore
either not be used at all or with great caution only.
One of the things which is most commonly said about virtues, particularly in modern books
on ethics, is that the virtues are dispositions. Here a warning is in place. It has to do with the
fact that to the specific virtues do not answer specific act-categories.
What is a disposition? The term has become something of a catch-word. As catch-words
generally, it can mean almost anything -- and therefore often means nothing.
When is a man, in ordinary parlance, said to have a disposition towards something? One
typical case is when we are talking of matters relating to health. A man can have a disposition
to catch colds, for example. Or he can have a hereditary disposition for headaches. So-called
allergic diseases are typically dispositional. A man is, e.g., sensitive to the scent of horses --
whenever he comes near a horse, he begins to sneeze or to breathe heavily.
The word 'disposition' is also commonly used in connexion with so-called states of temper. If
a man easily gets angry or upset or moved to tears or sad, we may speak of him as having a
certain disposition.
Dispositions both of health and of temper can be called latent traits which, under specific
circumstances, manifest themselves in characteristic signs -- such as sneezing or shedding
tears. The appearance, with some regularity, of these signs in the appropriate circumstances
decides whether there is a disposition of not. Dispositions are typically 'inward' things with
'outward' criteria. They are that, which the virtues would be, if there existed act-categories or
specified activities answering to virtues.
I do not insist upon a'common meaning of the term 'disposition' for all the cases in which this
word is ordinarily used. But I would maintain that there is no current sense of the word
'disposition', in which the various virtues could be said to be dispositions. The philosopher
who calls them dispositions is therefore giving to the term 'disposition' a novel use. This he is
entitled to do. But then he must explain what this novel use is. This I have never seen done.
-142-
The nearest equivalent in matters of conduct to 'dispositions' in matters relating to health and
temper are habits, I would say. A habit may be defined as a certain acquired regularity of
acting. A habit manifests itself in the doing of a characteristic act or in the performing of a
characteristic activity under recurrent conditions. A habit can be to take a nap after lunch or a
whisky before bed-time. Habits, like virtues, are greatly relevant to questions of good and
evil. They differ from virtues in that to them always corresponds either a specific activity or a
specific act. This is of the essence of habits, and is reflected in the fact that habits are nearly
always named after acts or activities, which can be performed independently of the existence
of the habit.
To regard virtues as habits would be to misunderstand the nature of virtues completely. One
may even go as far as to saying that, if virtuous conduct assumes the aspect of habitual
performance, this is a sign that virtue is absent. But if somebody were to say that the
acquisition or learning of a virtue is, partly at least, a matter of habituation, i.e. of getting used
to something, then he would probably be hinting at some important truth.
5. We have so far mainly said negative things about the virtues. Virtues are not associated
with specific activities; therefore virtue is different from technical goodness and the genus of
the virtues is not the genus skill. The virtues are not associated with categories of acts;
therefore their genus is neither that of disposition nor that of habit. The question may be
raised: Can the virtues become specified within a genus at all? If so, what is the genus of the
virtues?
The master philosopher in the field in which we are now moving, did not think that the virtues
were all of one genus. His opinion, no doubt, was partly influenced by the obvious logical
inhomogeneity of the meaning of arete in Greek. It has also to do with his division of the
virtues of man into the two groups, which are usually called in English by the names moral
and intellectual virtues respectively.
The intellectual virtues with Aristotle are a very mixed bunch. None of them is what we
would without hesitation call a virtue. This negative trait seems to be nearly the only thing
that is common to them all. Among the intellectual virtues Aristotle counts 'art' (in Greek
techne) or 'knowledge of how to make things'. Excellence
____________________
1
EN, Bk. VI, Ch. 4.
-143-
in such virtue is not identical with, but related to, that which we have called technical
goodness. Among the intellectual virtue he further counts demonstrative knowledge and
intuitive grasp of first principles. Excellence in such virtue again is more like the goodness of
a faculty. Yet it would not be right to think of them as intellectual endowments or gifts. For
intellectual virtue, according to Aristotle, is acquired. It 'owes both its birth and its growth to
teaching', he says.
1

Finally, Aristotle's list of intellectual virtues mentions practical wisdom (phronesis) and the
related 'minor intellectual virtues' of deliberation, understanding, and judgment. Practical
wisdom is knowledge of how to secure the ends of human life, or 'a reasoned and true state of
capacity to act with regard to human goods'.
2
.
2
. It is not, however, knowledge of how to
secure ends in general. Rather it is capacity to act with regard to that which is, in our
terminology, beneficial or harmful, i.e. good or bad for us.
3
This establishes a link between
practical wisdom and the moral virtues. (See below, sect. 7.)
To call the second group of virtues with Aristotle 'moral', as has become the custom, is rather
misleading. The Greek word is ethikos.What it points to in the context is not so much that
which we regard as the ethical or moral flavour of the virtues concerned, as another trait
which, on Aristotle's view, distinguishes them from the intellectual virtues. This trait is that
they are acquired, not as the intellectual virtues through teaching alone, but through teaching
in combination with habituation or the practising of virtue.
Aristotle thought that his second group of virtues, the 'ethical' or 'moral' ones, fall under a
genus. This genus of theirs is state or trait of character. This, I think, hits the nail on the head.
Not skills, not dispositions, not habits, not features of temperament, but traits of character is
what the virtues are.
I shall not here discuss the concept of character. It is one of the obviously most important but
at the same time most strangely neglected concepts of moral philosophy. It can be said with
some, though hardly undue, exaggeration that only two philosophers
____________________
1
EN, 1103
a
15.
2
EN, 1140
b
20-21
2
EN, 1140
b
20-21
3
In EN, 1140
b
4-5 Aristotle calls practical wisdom 'a true and reasoned state of capacity to
act with regard to the things that are good and bad for man'
-144-
have dug deep into the nature of this concept. One is Aristotle, who thought that character is
acquired rather than inborn, and that it is capable of becoming moulded and developed
according to human design and not only mutable according to the ordinances of natural
necessity. The other is Schopenhauer. To him character seemed inborn and immutable. On
Schopenhauer's view there is no moulding of character according to human design. Yet, on a
transcendental plane, man is responsible for the choice of his character -- a strange opinion,
but not entirely unlike the views of another profound searcher of human nature, Plato.
6. 'Virtue', i.e. moral virtue, Aristotle goes on to say, 'is a state of character concerned with
choice, lying in a mean . . ., this being determined by a rational principle.'
1

The idea that the path of virtue is a via media aurea between two extremes, I shall not discuss.
It is a fine conceptual observation, and has nothing to do with philistine mediocrity, as has
sometimes been maintained. It may, however, be doubted whether it has the general validity
which Aristotle asserted for it. Aristotle himself seems to have had doubts about this.
The idea that virtue is concerned with choice seems to me to hint at something more essential
than the idea that virtue lies in the mean. I propose to make the following use of it:
Virtues are essentially connected with action. This connexion, however, is with act-
individuals and not with act-categories. In Aristotelian phraseology, virtues have an essential
and peculiar connexion with particulars. It is here that choice enters the picture. One could
also put it as follows: Because of the lack of an essential tie between a virtue and an act-
category, the path of virtue is never laid out in advance. It is for the man of virtue to
determine where it goes in the particular case. This determination can be called a choice, but
we must not necessarily think of it as a choice between alternatives. ( Aristotle's notion of
prohairesis is therefore, perhaps, not altogether adequate to the logical nature of the case.)
The choice connected with a virtue could also be termed a choice of right course of action.
But then the question will arise: 'right' in which respect or 'right' with a view to what? To
answer: right with a view to meeting the demands of virtue, is no help. We must look out for a
better answer.
____________________
1
EN, 1106
b
36.
-145-
By no means every choice-situation is one in which there is room for practising some virtue.
Normally, when I choose a dish from the menu, virtue is not called upon to be displayed. But
when I deliberate whether to have another helping of a dish of which I have already had three
helpings, virtue may be needed for choosing rightly.
What then is characteristic of a choice-situation, in which a virtue becomes relevant? Briefly
speaking: That the case should be one in which the good of some being -- either the choosing
agent's own good or the good of some other being or beings -- is at stake, i.e. is likely to
become affected by the choice. To have another helping of a delicious dish is tempting, but
may cause indigestion. Here temperance is needed for choosing rightly. Or, if I provide
myself with a third helping, some other person at the table may be deprived of the possibility
of having a second helping. Then consideration is required.
We can now answer the question, with regard to what the choice, for which a virtue is needed,
is right. It is right, we could say, with a view to the good of some being involved.
To further problems connected with the distinction between the case, when the good at stake
is the choosing agent's own good and when it is somebody else's, we shall return presently.
First, we must try to get a clearer view of the rle of a virtue, as such, in a choice.
7. In the sentence, which we quoted from Aristotle, mentioning state of character, choice, and
the mean as characteristics of virtue, there is also mention of a 'rational principle' determining
the choice. This mention of a rational principle refers to the rle of the intellectual virtue,
which Aristotle calls practical wisdom, in the exercise or practising of any moral virtue.
We shall not here discuss Aristotle's view of the relation of knowledge to virtuous action. (
Aristotle's theory of phronesis is one of the most difficult and obscure, perhaps also one of the
most profound, chapters of his ethics.) I think we must accept the idea that in action in
accordance with virtue (or perhaps rather: in the right choice behind such action) a kind of
knowledge is involved. This, as I see it, is essentially knowledge relating to the beneficial and
the harmful, i.e. to that which is good or bad for a being. 'Practical knowledge' is not ill-suited
as a name for such insight. It could also be called 'knowledge of good and evil'.
-146-
In the sentence which we quoted, there is no mention of one essential feature which is still
missing from our logical picture of a virtue. (I am not suggesting that Aristotle himself did
not, elsewhere, pay due attention to it.) It can be called an emotion or feeling or passion. This
feeling 'contends' in the choice of the right course of action with our rational insight into good
and evil. It tends to eclipse or obscure our judgment both as regards consequences and as
regards wants (valuations). The rle of a virtue, to put it briefly, is to counteract, eliminate,
rule out the obscuring effects which emotion may have on our practical judgment, i.e.
judgment relating to the beneficial and harmful nature of a chosen course of action.
Action in accordance with virtue may thus be said to be the outcome of a contest between
'reason' and 'passion'. If we raise the question: What has the man of virtue learnt? -- and this
question, we have said, always makes sense -- the general form of the answer is: He has learnt
to conquer the obscuring effects of passion upon his judgments of good and evil, i.e. of the
beneficial and the harmful, in situations when he is acting.
In the case of every specific virtue there is some specific passion which the man of that virtue
has learnt to master. In the case of courage, for example, the passion is fear in the face of
danger. In the case of temperance it is lust for pleasure.
Consider, e.g., courage. In courage the good of man is involved via the notion of danger.
Danger may be defined as an impending bad or evil, i.e. as something which threatens a being
with harm.
If we ask what the courageous man has learnt, the answer is not primarily that he has learnt to
estimate and to cope with danger, but that he has learnt to conquer or control or master or
subdue his fear, when facing danger.
What does it mean that the courageous man has learnt to conquer his fear? It does not mean
that he no longer feels fear when facing danger. The brave man is not necessarily 'fearless' in
the sense that he knows no fear. Some courageous men may even feel fear intensely.
Considering this, what then does the brave man's conquest of fear amount to? Here we have to
note the fact that men's conduct, when facing danger, is often influenced by fear. Fear can
paralyse a man so that he becomes unable to do anything to meet the danger.
-147-
Of it makes him run away panic-stricken. Fear may thus be a bad thing due to its influence on
a man's conduct. He who has conquered fear has learnt not to let fear, should he feel it, do him
harm. He has learnt not to let fear paralyse him, not to get panic-stricken, not to lose his head
because of fear, but to act coolly when facing danger. In short: he has learnt not to let fear
obscure his judgment as to what is the right course of action for him. When he has learnt this,
he has learnt courage.
It should be observed that the course of action, which is the virtuous man's choice in the
particular case, is not necessarily that which we call a virtuous act or an act of virtue. A man
of courage, for example, may sometimes rightly choose to retreat from danger rather than to
fight it. Similarly, a temperate man may sometimes rightly choose to have for himself 'another
helping of pleasure'; and an industrious man may rightly choose to 'take a day off from work'.
Such choices, however, do not terminate in acts called after the virtues. To retreat from
danger is never an act of courage, to enjoy pleasure never an act of temperance, and to rest
never a display of industry. This is, I think, an observation of some importance. It shows a
new sense, in which a virtue is an 'inward' trait of character rather than an 'outward' feature of
conduct. The right choice in a situation, when a virtue is involved, need not be the choice of a
so-called virtuous act.
8. In this place it is pertinent to say a few words about a problem which much occupied Plato
and to a lesser extent Aristotle, too, in their thinking about the virtues. It could be called the
problem of the unity of virtue. Are not all virtues substantially the same frame or state of
character -- and their diversity due only to the diversity of passions which the virtuous man
has to master, or perhaps to typical differences in the acting-situations in which virtue is
displayed?
It seems to me that there is some foundation for the statement that there is, fundamentally, but
one virtue. What would then be a suitable name for it, if we do not simply call it 'virtue'? The
question of a name, it would seem, is of some importance here. For if we cannot call this
master virtue by another name but 'virtue', then to say that all virtues are but forms of one
virtue is an uninteresting tautology. It would also lead to confusion with that
-148-
sense of the word 'virtue', which is the opposite of vice, and which does not admit of a plural.
(See above, sect. 2.)
On the view which we have taken here of the various virtues, the name of the master virtue
could not be 'justice'. It is, for one thing, doubtful whether justice fits the conceptual pattern of
a virtue, which I have here been outlining, and thus also doubtful whether justice, on our
definition, is to be counted as one of the virtues at all.
But another name comes to mind: self-control. The various virtues, it may be said, are so
many forms of self-control. For what is self-control but the feature of character which helps a
man never to lose his head, be it for fear of pain or for lust after pleasure, and always let his
action be guided by a dispassionate judgment as to that which is the right thing for him to do.
The untranslatable Greek word sophrosyne, which was sometimes called the master virtue or
harmony and unity of all virtues, may have connoted something similar to this all-embracing
virtue of selfcontrol.
It is inviting to compare by analogy self-control to justice -independently of the question
whether justice is a 'virtue' in the same sense as self-control or not. I.e.: it is inviting to
compare the man who rules his passions by self-control, to the state in which justice reigns.
This analogy between 'the state within us' and 'the state without', as is well known, is
fundamental to Plato's political philosophy.
9. The man of virtue, we have said, has learnt to conquer some passion.
The conquest of passion presupposes that one has been susceptible to its influence -- at least
to some extent. If this condition is not fulfilled, one's conduct in the relevant situations may be
exactly similar to the virtuous man's conduct. Yet one could not be said to possess virtue --
except in a purely 'external' regard. A man who is totally insensitive to the temptations of
pleasure could not be temperate, and a man with no amorous passions could not be chaste,
although 'outwardly' his conduct could be the very paragon of temperance and chastity. It may
seem more difficult to admit that a man who never experienced fear and thus was literally
fearless, could not, strictly speaking, be brave. But this is probably because fear is such a
fundamental passion that its total absence in
-149-
a man comes near to being a mental defect. Halfwits, who do not grasp danger as normal men
do, can show the most astonishing fearlessness, e.g. in battle. But there need not be any false
resentment behind the hesitation we naturally feel to call such men brave.
How does one achieve the conquest of passion, which is a necessary condition of acquiring a
virtue? This question is certainly not important only to educationists and psychologists. Its
conceptual aspects are interesting too. I shall here touch upon them very lightly only.
According to the explanation of a virtue, which we have given, the conquest of passion means
that the 'obscuring' influence of passion on the practical judgment in particular acting-
situations has become eliminated. To call the influence 'obscuring' is to say that it induces us
to make wrong choices, i.e. choices which we later have reason to regret and of which we can
subsequently say 'had I surveyed the situation and its implications clearly, I should have acted
differently'. (Not always, of course, when we can speak thus, is it because some passion has
obscured our insight; the mistake we made can, e.g., have been a 'purely intellectual' mistake
about the consequences of our actions.)
The conquest of passion, which is the road to a virtue, is thus also a gain with a view to our
welfare. To subdue passion, in that sense of 'subdue' and 'conquer' of which we are now
speaking, is a useful thing. Awareness of this usefulness, it seems to me, must be an important
factor in the education to virtues.
How does one become aware of the usefulness of conquering passion? It is by no means
obvious that passions must have detrimental influences. To argue that passions, too, basically
serve the good of man does not seem unplausible. Could one not say, e.g., that the natural and
proper function of fear is to warn of impending danger; fear holds us back and makes us take
precautions, where otherwise we should easily run into disaster? The question is a little
bewildering, and from the fact that some passions may have an obvious usefulness it does not
follow that they all have. I shall not here try to argue the general question of the usefulness of
the passions. It suffices for present purposes to note that there is nothing obviously bad about
them and that our natural inclination certainly is to follow their impetus and not to go against
it. To realize the usefulness of conquering passion may therefore be connected with
considerable difficulties.
-150-
A man may come to realize the usefulness of temperance from having suffered the pains
consequent upon overeating, or the usefulness of industry from having witnessed the miseries
of destitution. This is not to say that only through 'vice' can one learn 'virtue'. Only in
exceptional cases does the vicious man turn virtuous. But some foretaste of the life of the
wicked may be effective and even necessary as a means in the education to a virtue. This
foretaste is often provided in the form of discouraging or deterrent examples. A whole genre
of literature is concerned with the fictitious setting up of such anti-models of virtues. Its
educational or edifying value is often disputed. I shall not enter the disputes of educationists. I
shall only say that I think it of some importance for educationists of a certain bent of mind to
remember that virtues are no ends in themselves but instruments in the service of the good of
man, and for educationists of a certain other bent of mind not to lose sight of the fact that it is
only by being aware of harmful consequences of yielding to passion that man has a rational
ground for aspiring after virtues. For, be it observed in this connexion, we do not commonly
and naturally call the virtues 'beneficial'. This is significant. The virtues are needed; absence
of virtue is a bad thing for us. The goodness of the virtues is that they protect us from harm
and not that they supply us with some good. (Cf. Ch. III, sect. 1 and Ch. V, sect. 10.) This,
incidentally, is why pride of possessing various virtues is stupid conceit and exhibitionism in
them a counterfeit of the good life.
There has been much dispute among philosophers, whether such and such is a virtue or not.
That courage is a virtue has, as far as I know, never been disputed. But whether charity or
chastity or humility are virtues has been put in question.
Disputes of this kind sometimes concern the conceptual status of the objects of dispute. For
example: whether they can rightly be called traits of character or whether they are relevantly
concerned with choice or with the conquest of passion. Such dispute may be of considerable
interest. It may serve to sharpen our idea of a virtue or lead us to distinguish between different
kinds of virtues or maybe even between different concepts of a virtue.
But the question, whether something is a virtue or not, can also have an entirely different
meaning. It can ask whether something which undoubtedly is a state of character concerned
with choice
-151-
guided by a rational principle in contest with passion, really has the usefulness of a virtue, i.e.
is needed for protecting our welfare. To question whether something is, in this sense, a virtue
or not may be a thoroughly sensible thing to do. This is so, becauseas was already noted -- it
is by no means obvious that the influence of passion on the practical judgment must be
obscuring, i.e. induce us to make choices which later we have reason to regret. If no such
harm is to be expected, conquest of passion or thwarting of natural impulse becomes a
pseudo-virtue. To encourage pseudovirtues in oneself or in others is moralistic perversion.
But to know whether something is or is not a pseudo-virtue, may require great psychological
insight into the conditions of man.
It is here good to remember that man's needs and wants and wishes, his fears and hopes, are
not immutable facts of his natural history. They are conditioned by a number of factors which
are themselves susceptible to change. We need only think of the important rle which
religious beliefs have played in the formation of man's views as to what is good or bad for
him, and therefore also as to what is a virtue or not. Changes in the religious outlook of an age
need not affect its conception of what a virtue is. But they may influence the estimation of the
importance of various features of character to the good life and therewith also its conception
of what is a virtue.
10. There are various ways of dividing the virtues into groups on the basis of characteristic
differences and similarities.
Thus, for example, courage and temperance may be regarded as specimens of two essentially
different types of virtues. Courage normally manifests itself in virtuous acts, temperance in
virtuous abstentions or forbearances. This conceptual asymmetry between the two virtues is a
consequence of their different relationships to pleasure and pain, i.e. to the hedonic good and
bad. The courageous man, when acting bravely, agrees to suffer or undertakes to do
something in itself unpleasant, or maybe even painful, in order to avoid a future greater evil.
The temperate man, when showing temperance, forsakes an immediate pleasure for a similar
reason.
One could suggest the name ascetic virtues for those virtues which, because of some
conceptual peculiarity of theirs, normally manifest themselves in forbearances. Temperance is
thus an example of an ascetic virtue. Chastity is another.
-152-
Virtues are often also divided into self-regarding and otherregarding virtues. One way of
marking the distinction between them is to say that self-regarding virtues essentially serve the
welfare of the agent himself, who possesses and practises them, whereas other-regarding
virtues essentially serve the good of other beings. Courage, temperance, and industry are self-
regarding; consideration, helpfulness, and honesty other-regarding.
The sharpness of the distinction is not obliterated by the obvious fact that virtues, which are
essentially self-regarding, may also be accidentally other-regarding, and vice versa. Take
courage, for example. To be courageous is necessarily a good thing for the brave man himself,
when he is facing danger -- although, of course, what courage does to help him may become
counteracted by other things which work against him. Courage can also be of the greatest
importance to action which is done for the sake of others, e.g. in battle or when saving our
neighbour from disaster. But whether a man's courage is or is not useful in the service of the
good of others, depends upon the attitude which he happens to take to this good, or is
compelled to take by external circumstances, such as threat of punishment if he flies. The
courage which burglars or robbers display, can be much to the detriment of their neighbour's
welfare. The value of courage from a 'social' point of view is therefore accidental; it depends
upon the attitude, which the brave man happens to take to his fellow-humans.
The virtues we call other-regarding are essentially useful from the neighbour's point of view.
The considerate man, for example, has learnt to conquer the influence of selfish impulses on
his judgment as to how his action will interfere with the good of others. That he has learnt this
much is already a welcome thing for his neighbour. It means that he can show consideration,
if he cares for his neighbour's good. He has acquired the necessary mental discipline. And we
would, of course, not call him considerate, unless he also cares. But there is a 'logical gap'
between a man's practical wisdom and his virtue, in the case of the other-regarding virtues,
which is not there in the case of the self-regarding virtues. We can explain the difference as
follows:
There is a sense in which a man necessarily will, in the particular case, practise as much self-
regarding virtue as he happens to possess. This follows from the relation, as we have defined
it (. 10), of a man's good to that which he wants and
-153-
shuns, and the relation of virtue to knowledge of the right course of action with a view to a
man's good.
Knowledge of the right course of action in a particular situation with a view to our
neighbour's good involves, however, no such necessity. For whether action is in accordance
with this knowledge or not, depends upon whether one wants to do that which one's neighbour
would welcome one to do, and whether one wants this or not is an entirely contingent matter.
To say that the practising of self-regarding virtue presupposes, beside an unobscured
judgment, also an interest in one's own welfare is not to state a presupposition at all. For the
sense in which interest in one's own welfare is presupposed is the sense in which this interest
is necessarily there, viz. as defined in terms of that which the agent wants (welcomes) and
shuns.
To say, however, that the practising of other-regarding virtues presupposes, in addition to an
unobscured practical judgment, also an interest in the good of some other being is to state a
presupposition. It can contingently be fulfilled or not.
To let the other-regarding virtues be, in the sense here explained, dependent upon
contingencies, may seem to some unsatisfactory. Is there no sense then, in which practising
other-regarding virtues is incumbent on man? To think that it is incumbent is to think that it is
a man's duty to aspire after and observe virtues independently of what his contingent interests
happen to be. With this we have arrived at the big question, how considerations of duty and
norm are related to considerations of good and evil. The three last chapters of this work will
be devoted to a discussion of some aspects of this problem.
-154-
VIII
'GOOD' AND 'MUST'
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
-155-
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
-156-
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
-157-
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.
-158-
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.
____________________
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.'
-159-
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'.
1

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
____________________
1
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 wellbeing or happiness (Glckseligkeit) 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.
-160-
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.
-161-
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.
-162-
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.
-163-
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
-164-
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 U=223Cq. 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
-165-
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 had 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
-166-
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
-167-
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 fiance. 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.
1
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
____________________
1
De Motu Animalium 701
a
10-40. Cf. also EN 1147
a
25-31. It should be remembered,
however, that Aristotle's concept of a practical syllogism has a wider scope than ours.
-168-
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 himseff 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 oneseff to do something.
When the act is completed, we can ask in retrospect, 'What made him 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.
-169-
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 verbalizing 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 rle
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).
1

____________________
1
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,
Grundlegung 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
betrifft, analytisch.' It is interesting to note that Kant regarded the proposition as analytical.
-170-
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 fiance, 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.
-171-
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 out 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
-172-
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.
-173-
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
-174-
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
-175-
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.
-176-
Also the acts of obeying the norms can become thus value-directed.
It will perhaps be objected to out 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.
-177-
IX
DUTY
1. IN the preceding chapter we introduced the notion of a wellgrounded norm. A
heteronomous command we call well-grounded, when the act of issuing it is a practical
necessity with a view to the norm-authority's ends. Autonomous norms are ipso facto
wellgrounded. (Ch. VIII, sect. 9.)
We shall in this chapter study a special case of well-grounded norms. This is the case, when
the ultimate end, relative to which the norm is well-grounded, is the good of some being. Of
this case I shall say that it imposes a duty on the norm-subject.
The term 'duty' is used with a multitude of meanings in ordinary language. Here it is used as a
technical term. Not everything which is called a 'duty', is a duty in our sense. Legal duties,
e.g., need not be. Whether so-called moral duties are or not, will depend upon the view we
take of the nature of morality. But I think it is true to say that everything which is a duty in
our sense, i.e. is a practical necessity with a view to (promoting or respecting) the good of
some being, is in common speech quite naturally called a duty. Therefore our use of 'duty' as
technical term is in good agreement with one of the uses of this word in ordinary language.
Duties in our sense can be suitably divided into certain main categories.
A first division is into self-regarding and other-regarding duties. The names and also the
sense, which we give to them here, are familiar from traditional ethics. When a duty is self-
regarding, the good, which the agent is supposed to serve by his dutiful action, is the agent's
own welfare. When a duty is other-regarding, it is the welfare of some being other than the
agent himself.
A second division is into autonomous and heteronomous duties.
-178-
A duty is autonomous when dutiful action is incumbent on the agent itself as a practical
necessity -- independently of whether the good, which it serves, is the agent's own or some
other being's. A duty is heteronomous when dutiful action is heteronomously prescribed to the
agent.
When we combine these two grounds of division, we get in all four basic types of duty, viz.
autonomous self-regarding, autonomous other-regarding, heteronomous self-regarding, and
heteronomous other-regarding duties.
There is, however, also a third way of dividing duties, which must be noted. It is their division
into that which I propose to call positive and negative duties. The first are duties to do, the
second duties to forbear something. The two kinds of duty thus answer to two sub-kinds of
that which we have called positive and negative commands. (See Ch. VIII, sect. 3.) Of the
negative commands we said that they are also commonly called prohibitions.
From the dutybound agent's point of view positive duties are (mainly) duties to promote,
negative duties are duties to respect the good of beings. From. the point of view of the norm-
authority again the purpose of negative duties can be said to be (chiefly) to protect the good
of some being.
Negative other-regarding duties are related to one of the many different concepts of a right.
This notion of a right is a normative idea with a characteristic dual aspect. From the point of
view of the right-holder, a right in this sense is a freedom or permission to act in a certain
way. From the point of view of the dutybound agents again, the right is a prohibition to
interfere with the right-holder's action, should he choose to avail himself of his right.
2. An autonomous self-regarding duty, according to the definitions we have given, is an
autonomous necessitation of the will to do something for the sake of promoting or protecting
the acting agent's own good.
Autonomous self-regarding duties must not be confused with autonomous practical
necessities in general. Consider once again the example of the man, who runs to the station.
(See Ch. VIII, sect. 7.) He wants to be in time for the train. Unless he runs, he will be late.
Therefore he has to run. This is autonomous practical necessitation. Why is it not autonomous
self-regarding duty?
-179-
Wanting to be at the station in time is usually not an ultimate end of action. A man may want
this because he wants to meet someone, or because he has promised to meet someone and is
anxious to fulfil his promise. These could be ultimate ends. Some such ends may actually
make his action other-regarding duty. They would not make it self-regarding, unless the agent
could say truly of himself that he wants to be at the station in time, because his welfare
demands this. I am not suggesting that a man could never say this truly of himself. But it
would certainly be an uncommon case.
There are two main types of case, when a man can be said to care for his own good and on
that account to have autonomous duties towards himself.
The first case of caring for one's good gives rise to negative duties only. There are certain
things which a man shuns or regards as 'in themselves' unwanted. He may sometimes be
willing to suffer those things for the sake of some end which he wants to attain -- as, e.g.,
when he decides to have a tooth pulled out. But unless he has some such end or some other
wanted consequences in view, he will consider those things bad for him and try to avoid them.
Now assume that he realizes that, if he does p, then some such in itself unwanted thing q will
befall him. He finds, for example, that a particular kind of food, which he likes in itself,
upsets him very violently. He could then adopt as a maxim or rule never to take food of that
sort. He, as it were, enforces a certain prohibition on himself for the sake of protecting
himself (his 'good') against something unwanted. Such self-protective selfprohibitions can be
called autonomous self-regarding duties.
The second case of caring for one's own good occurs, when a man deliberates about his ends.
But does a man ever deliberate about his ends? Can one deliberate about ends? The question
was discussed with much ingenuity by Aristotle.
1
On Aristotle's view deliberation is about
means, not about ends. So-called deliberation about ends is about intermediate ends and
therefore really about means to some ulterior ends.
It is probably right to say that the word 'to deliberate' is most commonly used in situations
when we raise such questions as, 'How shall I achieve this?' or 'Which way shall I choose to do
this?, When we raise such questions, we are sometimes deliberating
____________________
1
See EN, Bk. III, Ch. 3.
-180-
about means to given ends, and sometimes about that which we have called 'ways of doing
things'. ( Aristotle did not distinguish the two.) But we are not deliberating about ends.
Sometimes, however, a man stops to consider questions like this: 'What consequences is my
pursuit of this end going to have -beside my possibly attaining the end?' He is perhaps
working hard to get a promotion in his job. 'Am I ruining my health? Is it good for me never
to afford time for relaxation? And when I get my promotion, what will then happen? I shall
get a higher pay, but there will also be heavier work and more responsibility and even less
time than now for relaxation.'
When a man in pursuit of ends stops to reflect upon the consequences of his pursuing and/or
of his attaining his ends, then he can truly be said to be deliberating about ends. This he can
also be said to do when, without a view to existing ends, he asks himself the question, 'What
shall I do?'
In such situations, however, a man can also be said to be deliberating about his own welfare.
He asks what is good and bad for him. As a consequence of such self-searching activity he
may come to aim at new or revised ends.
A man who deliberates about ends may come to see that there are certain things which he
ought to do and, perhaps more often, must not do, lest he shall regret it later in life. He may
come to the conclusion that he must take some physical exercise or else he will neglect his
health or that he must afford some time for reading novels or listening to music or else his
soul will dry up completely or, if he tends to be a spendthrift, that he must think of saving for
his old age. In reaching such conclusions a man may be said to impose upon himself
autonomous self-regarding duties. He, as it were, forces upon himself a certain course of
action or way of living with a view to what he considers necessary for his welfare.
There cannot exist an autonomous self-regarding duty to care for one's own good (welfare) in
either of the two chief ways, in which such care takes place. Self-care is the foundation and
source of all autonomous self-regarding duties. To say that a man imposes upon himself the
duty to care for himself is to say that his care for his own good makes it necessary for him to
care for his own good. This is empty talk, unless it means that it imposes upon him specific
self-regarding duties, e.g. a duty to care that he gets enough sleep or exercise.
-181-
3. The practical syllogism, which embodies the prototype of an autonomous other-regarding
duty, is this:
x wants to promote or respect the good of y for its own sake.
Unless x does (forbears) p, he will not be promoting or
respecting the good of y.

x must do (forbear) p.
By respecting the good of another being one may understand the forbearance of any act
which, if done, would be bad for this being. On this definition, the duty to respect is
necessarily a duty to forbear. By promoting the good of another being, again one may
understand the doing of something which will be good for this being. On this definition, the
duty to promote a being's good is necessarily a duty to act in a certain way.
It may, however, happen that a man by forbearing to act becomes responsible for damage to
another man, e.g. because a third man does some harm to him. Then a duty to act (to interfere)
can be called a duty to protect the neighbour's good, and this duty can be distinguished both
from the positive duty to promote and from the negative duty to respect this good.
It may also happen that a man by forbearing to act will promote the welfare of another man,
e.g. because this gives the other man an opportunity, which he would otherwise not have had,
of doing something he likes. Then a duty to forbear (to stay aside) can be called a duty to
forsake, and this negative duty can be distinguished from the other negative duty to respect
another man's good.
There are thus in all four types of duty, which fall under the case we are now discussing: the
positive duties to promote and protect another man's good and the negative duties to respect
and to forsake. But forsaking can also be regarded as a negative form of promoting a good,
and protecting as a positive form of respecting it. I shall therefore include forsaking under
promoting and protecting under respecting. And I shall say that, basically, promoting is a
positive and respecting a negative duty.
Autonomous other-regarding duties to promote the welfare of other beings have to do with the
feeling or sentiment we call love in a broad sense of this word.
Other-regarding duties, both autonomous and heteronomous, which require us to respect the
good of other beings, are felt to have a peculiar connexion with morality, to be the moral
duties
-182-
par excellence. We shall in the last chapter examine this opinion more in detail.
We distinguished in a previous chapter between self-and otherregarding virtues. To try to
acquire the other-regarding virtues: consideration, helpfulness, honesty, etc. can be part of the
autonomous other-regarding duties of a man. To try to acquire and observe them is certainly
part of the heteronomous other-regarding duties which we wish to implant in others. Of the
other-regarding virtues most, it would seem, have to do with respecting our neighbour's good;
but some, such as helpfulness, are obviously relevant to the promotion of this good.
When speaking of self-and other-regarding virtues (see Ch. VII, sect. 10), we noted that
practising the latter presupposes an interest in the welfare of other people. This is the interest
to which the first premiss of our practical syllogism above refers. Whether a man takes such
an interest in another man, or in other men generally, is contingent. For this reason there can
exist no such thing as an autonomous duty to love our neighbour or to want to respect his
good for its own sake. Love of our neighbour is the foundation and source of any autonomous
other-regarding duty we may have, i.e. impose upon ourselves, to promote our neighbour's
good. Similarly, the will to respect his good for its own sake is the foundation and source of
every autonomous other-regarding duty to respect it in this or that particular regard. Interest in
another man's good for its own sake cannot be autonomous other-regarding duty, since it is
presupposed in every such duty.
But is such interest even possible, albeit not duty? Here we are touching upon a major
problem of ethics, viz. the problem of egoism and altruism.
4. By interest in or regard for somebody's good I shall here understand a desire to promote or
respect this good.
That promoting his neighbour's welfare can be an intermediate end of a man's action presents
no problem. This case simply amounts to that regard for his neighbour's welfare is a means to
some ulterior end of the agent's. If the means is necessary, regard for his neighbour is forced
upon the agent by autonomous practical necessity.
There are innumerable ways in which regard for the good of others can become a means to
self-interested ends. The master
-183-
wants his servant to work effectively for him. This the servant will not do, unless the master
cares for his welfare -- not to speak of the necessity of not maltreating the servant. Ergo will
the master take heed to respect and promote the welfare of his servant to the extent he thinks
necessary. This is not what I here call autonomous otherregarding duty, since the ultimate
end, which necessitates action, is not an interest in the welfare of another being.
By the thesis of Psychological Egoism one could understand the idea that promoting and
respecting his neighbour's good is never anything but (at most) an intermediate end of a man's
action. Regard for the good of others for its own sake we could call pursuit of an altruistic
end. Egoism, as here defined, is thus a doctrine which denies the existence of altruism, i.e. of
altruistic ends of action.
The nature of this negative thesis must not be misunderstood. Egoism, as here defined, does
of course not deny that promoting and respecting another man's good can be an end of action.
It can even have the appearance of being an ultimate end. But egoism maintains that, if such
an end is not overtly intermediate, it can become unmasked as an intermediate end.
The thesis of egoism, however, admits of several interpretations -- which supporters of it have
not always kept well apart. The denial of altruism can be understood as the denial of a logical
possibility or as the denial that something is a fact. It is egoism as a denial of the logical
possibility of altruistic ends of action, which here interests us first of all.
It is easy to see, I think, what is the nerve of the idea that action cannot be genuinely
altruistic. It is a logical fact that any end of action whatsoever is some agent's end, i.e. is
something which the agent, whose ends we are considering, wants to attain. Therefore, if
promoting my neighbour's welfare is an end of my action, it is yet an end of my action,
something I want, my interest, for me a good. There is a sense in which aims in acting are
helplessly selfcentred.
The thesis of egoism, which we are now discussing, rests on a misunderstanding of the
significance of the logical facts, which we have just mentioned, about ends of action, wanting,
and the good of a being. Egoism misconstrues the necessary connexion, which there exists
between my neighbour's welfare as an end of my action and my welfare, as an impossibility of
pursuing the first
-184-
except as a means to safeguarding or promoting the second. This is the very same mistake as
the one which psychological eudaimonism commits. (Cf. Ch. V, sect. 2, the discussion of the
example on p. 90.) It is related to the self-refuting mistake, which psychological hedonism
commits, when it misinterprets the necessary connexion between satisfaction of desire and
pleasure as meaning that pleasure is necessarily the ultimate object of every desire.
Psychological hedonism, egoism, and eudaimonism are closely related philosophical views of
human nature. They are all false. They all commit very much the same mistake. Their mistake
consists in a misinterpretation of an existing necessary connexion. These necessities 'behind'
the falsehoods give to the three doctrines the strong appearance of truth which they possess. A
refutation of hedonism which denies the necessary connexion between desire and pleasure,
and a refutation of egoism or of eudaimonism which denies the necessary self-centredness of
all ends of action, is therefore doomed to fail of its object. The aim of the refutation is not to
deny these connexions but to put right their significance.
When the logical mistake of egoism has been corrected, the logical possibility of altruistic
(ends of) action has become established.
The admission of the logical possibility of altruism, however, is fully compatible with the
view that, as a matter of fact, regard for our neighbour is never an ultimate moving force of
human conduct. Apparent altruism, on this view, is always egoism in disguise.
We need not here discuss this doctrine in detail. It contains a good portion of truth. Its claim
to the whole truth is founded, I think, partly on an exaggerated pessimism about human nature
and partly on the logical mistake, which we just mentioned, relating to the self-centredness of
ends of action. In the gloom of pessimism it becomes tempting to give to one's insight into the
vileness of human nature an absolute character, which can be given to it only at the expense
of committing a logical mistake.
Beside the logical question, whether a man can take an ultimate interest in the good of his
neighbour, and the psychological question whether a man ever is interested in his neighbour's
good except as a means to some end, there is also the genetic question, whether every interest
in another man's good as an end was not in origin an interest in this good as a means.
-185-
Those who take the view that altruism must have a genetic foundation in self-interest are,
somehow, astonished at the fact that there should exist such a thing as pure unselfishness, i.e.
action which springs from affection, friendship, love, sympathy, or respect and which is
completely untinged by calculations about its utilitarian value for the agent. That man acts
self-interestedly seems not to require any explanation. But that he acts unselfishly may strike
us as something of a 'mystery'.
Is unselfishness a 'mystery', once we have seen clearly through the logical mistake which
egoism commits? I do not know what is the right answer to the question. The answer, of
course, partly depends upon what we mean by a 'mystery' here. But I tend to think that the
appearance of 'mystery' or of something standing in need of 'explanation', which pure
unselfish action can be said to exhibit, is really nothing but a conceptual confusion nourished
by the same mistake as the one underlying hedonism and eudaimonism. Therefore I am also
inclined to think that action inspired by affection, friendship, and love is just as typical of
human nature as action in a spirit of self-interest, and that the view of man as essentially a
self-interested creature is a logical misconception. But I may be mistaken.
5. Behind heteronomous self-regarding duties there is a practical syllogism of the following
type:
x wants the good of y to become promoted for its own sake.
Unless y does p, his good will not be promoted.

x must make y do p.
The distinction between promoting and respecting and also the distinction between doing and
forbearing is not, it seems, of much importance to the case which we are now considering. I
shall therefore here use the verb 'promote' to cover 'promote or respect' and the verb 'do' to
cover 'do or forbear'.
The conclusion of the above syllogism is an autonomous practical necessity incumbent upon
x.x ought to make y do p. There are many ways in which x can try to achieve this. One way is
by commanding y. This is the only way which interests us here. When it is resorted to, then x
imposes upon y a duty to act in a manner which x considers good for y. The duty is
heteronomous.
-186-
y or the norm-subject is not moved to action by his own wants and insights into necessities. It
is x or the norm-authority, who is thus moved.
Thus when there is a heteronomous self-regarding duty, the norm-subject's good is the norm-
authority's end. The authority imposes upon the subject a duty for the sake of (in the name of)
the subject's welfare. But the subject, if he obeys, does not necessarily do the ordered thing
for the sake of his good. He may do it simply and solely for the sake of escaping punishment.
Or he may do it because he wants to please the authority, or out of respect for the authority.
6. When does commanding actually take place 'in the name of' the good of the commanded?
There are some typical cases. One is when parents order their children to do certain things,
because it is good for the child, or to forbear something because it would be bad for the child
to do it. Also in the relations of teachers to pupils orders of a similar nature are sometimes
given. Finally, people who are feebleminded or for some other reason incapable of looking
after themselves may have to be commanded by others, who have a better understanding of
their welfare.
It will be noted that the agent who issues commands to others in the name of their good,
usually enjoys some recognized position of 'authority' over those whom he commands -- such
as parent or teacher or guardian. It will also be noted that two of the three typical cases which
we mentioned, have to do with that which we call education.
I shall call -- partly for want of a better name -- commands, which are issued in the name of
the good of the commanded, moral commands. Of the educational purposes, which such
commanding may serve, I shall speak as moral education. And the authority who gives such
orders, I shall call a moral authority.
This use of the adjective 'moral', incidentally, is not unnatural or at great variance with
ordinary usage -- though maybe not very familiar in moral philosophy. It has no immediate
connexion with so-called moral goodness nor with so-called moral duty. I am not suggesting
that the duties imposed by what I call moral commands are those which we ordinarily call
moral duties. The activity, however, which in common speech is called moralizing, largely
consists in advising and recommending to other people
-187-
things in the name of their good, and also in drawing their attention to various ways in which
people may have neglected their own welfare. Moralizing is not co-extensive with that which
I here call moral commanding. But that which I here call moral commanding falls under that
which is ordinarily called moralizing.
We shall distinguish between the reasons and the justification for commanding others in the
name of their good.
The immediate reason for such commanding is that the normauthority takes an interest in the
welfare of the norm-subject. The immediate reason need not, however, be an ultimate reason.
Sometimes a man takes an interest in the good of another man for the sake of some ulterior
end of his own. This is the case of the master who cares for the good of his servants because
he expects a better return of services. Sometimes a man is under a legal obligation to care for
the good of others in a manner which involves commanding them. Parents and guardians have
such obligations. It is an important, though not the sole, aspect of parental love that this love
should be the ultimate reason why parents oblige their children to acts and forbearances in the
name of the children's good.
Only when the ultimate reason for commanding others is an interest in their welfare, shall we
say that a heteronomous selfregarding duty is being imposed.
Commanding others in the name of their good can be said to entail a claim on the part of the
norm-authority to a better insight into the requirements of the welfare of the norm-subject
than the norm-subject can claim for himself. The justification of this claim is part of the
justification of heteronomous self-regarding duties.
It is a fact that some men have a much less developed conception than other men of that
which is good or bad for themselves. This can be due to various reasons, e.g. to lack of
experience of that which may befall a man as a consequence of his actions. Small infants can
be said to have as yet no conception at all of that which is good or bad for them. Children,
who already have some views in the matter, may nevertheless be completely mistaken in their
views, e.g. strongly want to do things which they will regret, thinking that they will not. It is
in such situations, when a being's conception of his own welfare either is lacking or
undeveloped or distorted,
-188-
that men are called upon to exercise moral authority over others.
Part of the justification of heteronomous other-regarding duties is thus the superior wisdom of
some men as to that which is good or bad for other men. The moral authority of parents,
guardians, and teachers can normally be expected to possess this justification for its exercise.
If men necessarily pursued their real good, any claim to superior wisdom on the part of others
would necessarily lack justification -- and therewith also any attempt to impose heteronomous
selfregarding duties. As a matter of fact, however, men often pursue things which 'on second
thoughts' they consider unwanted. The test which shows that something was a constituent of a
man's apparent as opposed to his real good, is that he regrets it. (See Ch. V, sect. 14.) To
know the requirements of another man's welfare better than he does it himself is essentially
the superior capacity of seeing what a man will later regret having done or neglected. The
man who exercises moral authority over another, is justified in doing so only to the extent that
he can say truthfully to the other man: 'There will be a day when you will have reason to be
grateful for that which I did to guide you.' The important point is not whether the other man
acknowledges his gratitude. Some men are by nature little inclined to be grateful even when
there is a reason. The important point is that there should be reason for thinking that, had the
other man not obeyed the commands, he would have been worse off in the long run, or for
thinking that, had he obeyed, he would be better off than he is now. But what is 'better off'
and 'worse off' is, in the last resort, for the subject himself to decide.
When the claim to superior knowledge of good and evil lacks justification, the exercise of
moral authority of one man over another deteriorates into something which may be called
moral tyranny. The subject is then forced to do things 'in the name of his good', when actually
he knows better himself what is good or bad for him. To know whether the claim to superior
knowledge on the part of the norm-authority is justified or not, can be most difficult. The
struggle for moral freedom,i.e. for the right to act in accordance with that which one considers
good or bad for oneself, against alleged moral tyranny, can greatly upset the relations between
parents and children in the period before the latter come of age.
-189-
It is an important aspect of that which we call a child's 'coming of age' that the child reaches
that which we shall here call moral maturity. An individual can be said to have reached moral
maturity when there is no obvious justification for a claim, which others may put forward, to
superior insight into the requirements of that individual's good.
With the reaching of moral maturity the moral education of men -- in the form of
heteronomous self-regarding duties being imposed on them -- comes to an end. It is the aim of
moral education not only to promote and protect the good of others during their moral
immaturity, but also to facilitate the process of maturation by teaching them to care for their
own welfare. But here again it is important to remember that the criterion of the present
superior wisdom of the teacher is set by the future verdict of the pupil on the fruits of his
teaching.
7. We now proceed to heteronomous other-regarding duties.
A man x commands another man y to do something. The command, let us assume, is well-
grounded. This means that x must issue it in order to secure some end of his. If this end is the
good of y and if it is an ultimate end, then the command imposes a heteronomous self-
regarding duty on y. If the end is the good of some being other than y and is an ultimate end,
then the command imposes a heteronomous other-regarding duty on y in relation to that other
being.
The case when one man commands a second man in the name of the welfare of a third man, is
a thoroughly possible and by no means unrealistic case. Nevertheless it seems to be a case of
rather subordinate interest. We shall therefore here disregard it.
With this exclusion, the remaining cases of heteronomous otherregarding duties are those in
which one man for the sake of his own welfare commands another man. Such duties must be
distinguished from heteronomous commands for self-interested ends generally. We do not
wish to say that we are imposing a heteronomous other-regarding duty on a man, if we order
him to open a window because we want to have the room ventilated. But we may wish to say
that an order not to disturb us, when we are working, aims at imposing an other-regarding
duty.
As observed in sect. 2, one way in which a man can care for his own welfare is by taking
various self-protective steps in order to
-190-
escape things, which he shuns as being unwanted either in themselves or due to their
consequences. It would seem that it is this way of caring for one's own good that is
predominantly relevant to heteronomous other-regarding duties. Such duties, when imposed
by one man on another in the interest of the first, are mainly negative duties, i.e. prohibitions
to do things which the first man considers harmful for himself. From the norm-authority's
point of view heteronomous other-regarding duties are chiefly selfprotective prohibitions.
8. We have so far been talking as though, as a matter of course, any man could, if he wanted
to, command any other man. The recipient of an order may disobey the order, nevertheless he
has been commanded.
As a matter of fact, men do not often command other men. When this happens, it usually
takes place in circumstances under which commanding has become somehow
'institutionalized'. Parents command their children, but not the children of other parents.
Teachers sometimes command their pupils; for pupils to command their teachers is out of
place. Officers of superior rank command officers of inferior rank; commoners do not often
command one another. When I ask my friend to shut the door I do not command him; not to
speak of the case when I ask a stranger for a service. When I order a book through my
bookseller, this is not ordering in the sense of commanding.
Why is it that we do not command our friends or strangers? Is it because we are too polite
thus to intrude upon them? Can we command our friends? The question is worth asking.
The fact that commanding usually takes place under some 'institutionalized' circumstances
may suggest to us the idea that in order to be able to command one must possess something
which could be called a right to command, enjoy some recognized position of authority.
Parents have a right to command their children, children no right to command their parents.
Has the highwayman, who commands 'hands up', a right to command this? Certainly not. Is
he then not commanding 'Hands up'? He certainly is.
We have said before (Ch. VIII, sect. 3) that to a command is essentially tied a threat of
punishment in case of disobedience. This threat must not be empty or a mock threat. It must
be an effective threat. That it is effective does not mean that it necessarily
-191-
secures obedience. Nor does it mean that disobedience is without exception punished. I shall
not here discuss in detail, what constitutes the efficacy of a threat of punishment.
It thus follows that the possibility of commanding is essentially tied to the possibility of
effectively threatening the recipient of the order with punishment in case of disobedience.
This second possibility again depends upon that which I shall call the relative strength of the
agents concerned. Effective threatening is, on the whole, possible only when the individual
who threatens, is stronger than the individual who is being threatened. Thus the possibility of
commanding is founded upon the superior strength of the commander over the commanded.
This concept of strength stands in need of elucidation. It is not the same as that which is
ordinarily understood by 'physical strength', though it is related to it. A big, strong man may
not be able to command a small, weak one, because the second can run away and put himself
out of reach of the punishing arm of the first. But then it must be remembered that running,
too, is part of a man's physical powers.
Our strong man can, of course, continue to shout out commands to the small man and even
accompany his orders by the most fearful threats. Is he then not 'issuing commands'? If he is
not convinced that his words will not impress the small man, he can rightly be said to be
trying to command him. If he knows that his shouting and threatening is all in vain, we could
say that he is not commanding but only giving vent to his wishes and his anger. We could
also, if we wanted, distinguish between commanding, which would consist merely in the use
of words combined with a certain intention, and effectively commanding, which would
combine intention and the use of words with an effective threat of punishment. Then we could
say that it is only effective commanding, which constitutes normative relationships of the kind
we are here studying, between men.
The notion of strength, which is essential to the possibility and existence of normative
relationships among men, I shall define as capacity of affecting through one's action the good
of other beings favourably or adversely. By saying that two agents are of roughly equal
strength we here mean that they have roughly the same capacities (powers) of influencing the
welfare of one another. By calling one stronger than another we mean that the first can do
-192-
more for promoting but also more for injuring the good of the second than the second can do
with regard to the good of the first.
It is obvious that capacity of favouring and capacity of injuring someone's welfare need not be
proportionate. The fact that one man can do more good to another than this other man can do
to him does not exclude the possibility that the second man can do more harm to the first than
the first can do to the second.
To the capacity of commanding, the capacity of visiting another man with evil is of more
importance than the capacity of doing good to him. This is so, because commanding is
essentially connected with punishing disobedience but not with rewarding obedience. The
superior strength, necessary for commanding, is therefore essentially a superior power of
causing suffering to others. That such should be the logical foundation of heteronomous law
and heteronomous duties may seem brutish, but must, I think, be accepted as fact.
One man may accidentally enjoy superior power or strength, in the sense defined, in relation
to another man. Such accidental superiority can then be used for commanding. Blackmail is a
species of commanding, which is based upon accidental superiority of strength.
Consider two men 'by themselves', i.e. in abstraction from such relationships which they may
have to other men and which may be relevant to their capacity of doing good or harm to one
another. For example: the one man is king and can command his bodyguard to seize the other
man, if he does not obey the king's orders. Then abstract from this and think of the king
without his bodyguard. Abstract also from such accidental superiority, which one man may
have over another. For example: the one man has a rifle. Abstract from this, or endow in
thought the other man with a rifle too. The balance of strength, which remains after these
abstractions, I shall call the natural relative strength of the two men.
Such logical fictions as this about the powers of men in a state of nature may strike one as
old-fashioned and unrealistic. They were much in favour with writers of the 17th and the 18th
centuries. They may be grossly unrealistic as hypotheses concerning the conditions which
existed before the conventions, customs, and norms of society had 'perverted' the 'natural'
ways of life. As
-193-
abstractions for the purposes of studying concepts, these fictions may be both legitimate and
useful. It is to be regretted that an exaggerated respect for the 'naturalistic' study of man by
anthropologists, historians, and psychologists has made philosophers generally abandon these
fictions, of which the supreme product is the Contract Theory of the State.
It is a fact of fundamental importance alike to moral, legal, and political philosophy that men
are, by nature, roughly equally strong, i.e. endowed with roughly the same natural powers of
affecting the welfare of one another favourably and adversely. I shall express this fact by
saying that men are, by nature, approximate equals.
There are, of course, noteworthy natural inequalities too. Some of these inequalities are
founded in physical, others in intellectual superiority, some in both.
There is, first of all, the physical and intellectual superiority of adult people over children
before the latters' coming of age. On it the normative relationship between parents and
children is ultimately based. It should be noted that the natural limits to the normative
authority of a father or mother over children other than their own are not set by the children of
other parents, but by the parents of other children.
There is further a certain inequality, at least in physical strength, between the sexes. Whether
it is relevant to normative relationships is arguable. It certainly has been considered relevant
in some quarters.
Here must also be noted the fact that physical strength, up to a point, declines with increasing
age. This decline, however, is, up to a point, compensated for by an increase in the mental
power which we call 'wisdom' -- maybe not in matters purely intellectual, but certainly in
matters pertaining to the good of man.
Of the normative relationships between parents and children it is characteristic, not only that
it exists 'naturally', i.e. on a basis of natural superiority in strength, but also that it dissolves
'naturally', i.e. as a consequence of the children's growing up to become the approximate
equals of their parents. It is an interesting and deeply significant fact about man that children
become their parents' equals in strength roughly at the time when they reach that which I
called moral maturity, i.e. become capable of caring
-194-
for their own welfare. The basis of parental normative authority thus vanishes roughly at the
time when its exercise loses its justification.
At least as important as the fact that men are approximate equals by nature is the fact that men
can add to their strength by joining forces with other men. The ability to co-operate for ends
is not a uniquely human ability. But the rle which this ability plays in the creation of
normative relationships among men, has no counterpart elsewhere in the animal kingdom. For
it is on this possibility of joining forces that much of social life on the human level depends
and above all the great fabric of commanding power we call the state.
The significance for normative relationships of man's capacity to join forces with other men is
twofold. Firstly it can overrule all natural (and accidental) inequalities, which may exist
among individual men. If there are two men, x and y, on a desert island and x is the stronger,
there may also develop between them a relationship which we could call that of master to
servant. x commands and y, on the whole, obeys. If there are three men x and y and z and x is
strongest, x may succeed in making himself master of both y and z. His superior strength may
partly consist in his skill to kindle unfriendly relations between y and z and thus prevent them
from joining forces against him. But if y and z join forces, it is most likely that they shall be
able to withstand any attempt on the part of x to tyrannize over them.
Here we find the second respect in which man's ability to cooperate is important. It can create
'artificial' inequalities on a basis of 'natural' equalities, It is on such created inequalities that
institutionalized normative power of some men over other men mainly rests. The 'ruler', in the
widest sense of the word, is not necessarily stronger than each of his subjects. He is almost
certainly weaker than even two of them jointly. Yet he can command each of his subjects
because, should one of them be recalcitrant and refuse to obey orders, there will be others
who join hands with the ruler to punish the recalcitrant.
Thanks to the fact that men are approximate equals by nature the possibilities of the human
individual to command and therewith also to impose heteronomous duties on other
individuals are narrowly restricted. But thanks to the power of co-operation
-195-
heteronomous normative relationships yet play an all-pervading rle in the life of human
communities.
Inequalities of strength based on men's capacity to co-operate are the foundation of legal
duties and the reign of law. These inequalities, however, may also serve the establishment of
moral duties and the reign of justice. How this happens I shall try to show in the next chapter.
-196-
X
JUSTICE
1. IN the preceding chapter we stressed the importance of the two facts that men are
approximate equals by nature and that they can join hands to co-operate. Thanks to the first,
the possibilities of individual men to exercise normative authority over other men are
narrowly limited. The significance of the second fact, we said, was twofold. Thanks to it men
can withstand and overrule any 'tyranny', which one man may attempt to exercise over his
equals by commanding them. But thanks to it men can also subjugate their natural equals and
exercise joint normative authority over them. Both consequences of men's capacity to join
strength are simultaneously at work in society. The significance of both can be summed up in
the words 'society is stronger than the individual'.
Let us raise the question: What makes men join hands in cooperative efforts?
One thing which makes men co-operate is the practical necessity, autonomously incumbent
upon each of the members of a group of men, of working together for the attainment of a
common end. That the end is common shall mean that each of the men in the group wants a
certain state of affairs p to exist. That working together is necessary shall mean that, if any
one of the men stays away from work, the end will not be attained.
This case of co-operation under autonomous necessity is an extreme case. But it is an
important prototype.
If ends are goods attainable through action, an end which is common to several agents can be
called a common good. Not everything, however, which is a common good, i.e. a good for
each member of a group of men, is also a common end. That something is a good for a man
means that it is something which is, in itself
-197-
welcome to him. But not everything a man welcomes, if it befalls him, is an end of his action.
Hence, though common ends are also common goods, all common goods are not also
common ends.
Co-operation is for some end. This end can be a common good of all the co-operants. We
shall call this the first case. Or it is a common good of some of them. We call this the second
case. Or it is a good for none of them. This is the third case.
In the first case there are two possibilities. Either the men have to work for the common good
under autonomous practical necessitation, or some can stay aside and let the rest work, or
relax and let the rest work harder.
The first of these two possibilities answers to the prototype case which we have already
described.
The second possibility means that a group of men have a common end, but that it is not
necessary that they all work for it. We can in every such case distinguish a minimum number
of necessary co-operants and a maximum number of possible bystanders. This division gives
rise to a number of problems, which we cannot discuss here. We give only one example:
Three men have a common end of their action. The co-operation of two of them is necessary
in order that the end shall be attained. But one man can stay aside and look on, when the other
two work. Assume that x and y work and that z stays aside. To make z participate in the work
for the common good of the three men may then become another common interest of x and y.
Since they are two and z is one, they will probably be able to order z to co-operate with them.
If z remonstrates, they may jointly proceed to punish him. The cooperation of x and y in
commanding z is co-operation in the sense of our prototype case.
We proceed to the second case. Then the end of co-operation is a common good of some but
not of all of the co-operants. For example: Three men co-operate for a result p. Two of them
are anxious that p shall be. For the third, however, p is not a good, Perhaps he co-operates
because he has been commanded by the other two and is anxious to escape punishment. If this
is the case, then making the third man work is another common interest of the first two men --
beside the attainment of p. The co-operation of the two for this common end again is co-
operation of the prototype case, i.e. of the case when men are under autonomous
-198-
necessitation to co-operate for a common good. If, however, the group of men in pursuit of
the first common good is great, then co-operation of a few of them may be sufficient for
commanding some men from outside the group to labour for their end. Then the problem of
co-operation for commanding outsiders becomes a case of the problem of co-operation in a
group, which may be divided into necessary co-operants and possible bystanders. This
problem we already discussed.
Consider finally the third case. Here the end of co-operation is not an end of any man in the
group of co-operants. If this is the case, their co-operation serves as a means to some end of a
being or of some beings outside the group. Use of this means may have to be secured by
commanding the men to work. Then the problem of co-operation for a common end will arise
for the men, for whose ends the co-operants are assumed to work. This problem falls under
our first case.
By these considerations I have wanted to indicate, how that which I called a prototype case of
co-operation for a common good holds a crucial position among a variety of cases of co-
operation. I would maintain that all cases, which are not in themselves of this prototype and in
which co-operation has to be secured by commanding, rest -- in the last resort -- upon
autonomously necessitated co-operation for a common good.
There are, however, forms of co-operation, which do not necessarily serve the attainment of a
common good. We now turn to a study of some such cases.
2. Let it be assumed that x and y join hands to attain an end of x's, which is not also an end of
y's. x could not attain his end alone. He needs the assistance of y. Let us rule out the
possibility that x can command y to serve him. Since men are approximate equals, the
assumption that the one cannot command the other is thoroughly realistic. Let us further rule
out the possibility that y wants to help x, because he wants to promote the good of x for its
own sake. Since men, even if they are not unlimitedly egoistic, are not overwhelmingly
altruistic either, this assumption is thoroughly realistic too. Is there then any motive left which
could make y co-operate?
There is at least one possible motive left: hope of reciprocal service in return. y could argue: if
I do not help x to attain p now,
-199-
x will probably not help me to attain q on another occasion. If I help x now, there is at least a
good chance that he will help me then, considering that he may need my assistance on some
other occasion. Hence it is in my interest to help him.
This form of co-operation may be said to hinge logically upon three facts. The first is that
men are approximate equals and cannot normally exercise commanding authority over one
another. The second is that men far from always pursue the good of their neighbours as an
ultimate end of their own. The third is that men are to some extent dependent for their welfare
upon the assistance, i.e. co-operation, of others. Co-operation, which is not for a common
good, can yet be to mutual advantage.
Co-operation, which is to mutual advantage but not for the attainment of a common good, is
exchange of services.
Exchange of services, be it observed, need not take the form of co-operation. y does
something alone, say p, which is wanted or welcomed by x. p is thus a good for x. x, in return,
does something, say q, which is wanted by y. The exchange of services which has taken place
can be called an exchange of goods.
Exchange of services and of goods plays an important rle in the conceptual genealogy of the
normative ideas known as agreement, contract, and promise. I shall not, however, discuss
them here.
When exchange of services or of goods takes place, people can be said to work in turns for
the promotion of one another's good.
Hope of reciprocal service or of a good in return can thus be a motive for co-operation and for
promoting another man's good generally. But it can also be a motive for respecting it.
A man may gain an advantage for himself by harming his neighbour, i.e. by doing something
which his neighbour thinks unwanted. A man cannot command another man of roughly equal
strength to respect his good. But if the other man harms him, he may be able to 'pay back',
revenge himself and thus return evil by evil. Revenge may be regarded, logically, as a preform
of punishment, and threat of revenge as a preform -- on the level of rough equality of strength
among agents -- of normative relationships. One could say that revenge and retaliation are
what punishment 'logically' deteriorates into, when approximate equals try to command one
another.
To return evil by evil is also an exchange of services or of goods
-200-
of its kind. One could call it an exchange of 'negative' services or goods. Such exchange is not
to the mutual advantage, but to the mutual disadvantage of the agents, who engage in this sort
of commerce. But the reciprocity of disadvantage is at the same time a ground for a
reciprocity of advantage, viz. the mutual advantage gained from not doing evil to one another.
Thus respect for the neighbour's good can be to the mutual advantage of agents, who are
approximate equals by nature and therefore cannot effectively exercise normative authority
over one another.
The principle called the Golden Rule is sometimes given the following formulation: Do to
others what you want them to do to you, and don't do to others what you do not want them to
do to you.
When the principle is formulated in this way, it is presupposed that one man regards as
wanted and unwanted roughly the same things as those which any other man regards as
wanted and unwanted respectively. This presupposition is largely fulfilled. But it is no logical
necessity. There are exceptions to it. Something which I shun, another man may like. Then it
is not to be seen why the fact that I do not want it should be a motive for not acting in such a
way that he gets this thing. Similarly, it is not to be seen why I should do him something he
does not want, although I may want him to do this for me. But if we both want and shun
similar things, then we can exchange services to our mutual advantage by doing to one
another what we welcome and not doing to one another what we shun.
If the Golden Rule is formulated in a way which is independent of the presupposition of
similar wants, it would run as follows: Do to others what they want you to do to them and
don't do to others what they do not want you to do to them. The utilitarian defence of this
principle would then be that its adoption by all agents is to their mutual advantage.
The rule consists of two parts. We could call them the positive and the negative parts. The
positive part urges us to promote, the negative part to respect the neighbour's good.
There is a noteworthy asymmetry between the parts. That the adoption of the negative part of
the rule by all agents would be to their mutual advantage is a priori clear. Everybody would
then gain the advantage of never being harmed by anybody else. That the adoption of the
positive part of the rule by all agents must be
-201-
to the mutual advantage of all is not in a similar way clear. Some men are more demanding on
their neighbour's services than other men. If one man unlimitedly works for his neighbour,
there is simply no time for his neighbour to return the service. If promoting the good of one
another shall be to mutual advantage, there must be some 'check' on the demands which men
have on their neighbours' good services.
I shall not here discuss at length how this check is provided. I only suggest that it may be
provided by the negative part of the rule. This would mean that a man must never demand of
his neighbour greater services than those which the neighbour can render without detriment to
his own good. By demanding more he may become responsible for harm to the neighbour and
thus sin against the negative part of the Golden Rule.
These last considerations support the view that the negative part of the principle holds a more
fundamental position than the positive part. I shall henceforth in this chapter discuss only the
negative part, i.e. the rule which urges us to respect our neighbour's good.
The big problem before us is this: How can respect for our neighbour's good become duty? In
order to see how this is possible, we must try to link considerations relating to mutual
advantage with considerations relating to co-operation for a common good.
3. Consider a number of men 'in a state of nature'. They occasionally do harm to one another;
not often merely for the sake of harming, more often perhaps as an act of revenge for some
evil they have suffered from others, usually however because pursuit of some end of theirs is
causally incompatible with respect of the neighbour's good. Since some of them do harm,
some of them also suffer harm. Some of them may never do and often suffer, others again
often do and never suffer; but even those who escape harm from their neighbours may suffer
under the inconveniences of having to take various self-protective measures. The more evil a
man does, the more reason he has to fear evil himself.
Suppose the men come to agree that each of them has to gain more from never being harmed
by anybody else than from sometimes harming somebody else. I shall call this a basic
inequality of goods.
The men who agree or subscribe to the basic inequality, I shall speak of as constituting a
community. Observance of the rule never
-202-
to harm anybody (in the community) I shall speak of as the adoption of a Practice. We can
call it 'the practice of not-harming'.
A member of the community who adopts the practice, can be said to make a 'personal
sacrifice'. He gives up any such advantage which he may gain from doing harm to others. It is
not certain that, by giving up this advantage, he will gain the greater advantage of never being
harmed by anybody. But it is certain that, if every member of the community relinquishes the
smaller advantage of sometimes harming somebody else, then they will all gain the greater
advantage of never being harmed by anybody (in the community). Universal adoption of the
practice is thus a necessary and sufficient condition for universal attainment of the greater
good of the basic inequality. For being to the mutual advantage of all its members, universal
adoption of the practice is also a common good of the community.
Escaping harm from others I shall call each member's share in the community's greater good.
Universal adoption of the practice of not-harming can be called the price, which the
community has to pay for the greater good. Individual adoption of the practice, i.e. not doing
harm to others, can be called each member's due or due share of the community's price for
the greater good.
In order that the greater good of the community shall exist, it is (logically) necessary that all
members pay their due. But in order that an individual member shall get his share in the
greater good of the community, it is not (logically) necessary that he pays his due. To this end
it is in fact not necessary that anybody pays. The only thing which is necessary, is that nobody
does harm to him. If, however, everybody else pays his due, it is (logically) necessary that he
will get his share. For, that all the rest pay their due means that they adopt the practice of
never harming anybody else in the community. If they do this, the one remaining man can be
sure of getting his share in the greater good, independently of whether he pays his due or not.
He can therefore try to add to the greater advantage of never being harmed by anybody also
the smaller advantage of sometimes harming somebody-try to get the best of both worlds so
to speak. If he adds to the greater advantage of getting his share also the smaller advantage of
skipping his due, he will necessarily deprive some of his neighbours of their share in the
greater good. This I propose to call parasitic action.
-203-
The problem before us is to show how, in spite of the possibility of parasitic action, adoption
of the practice which is conducive to the greater good of the community, nevertheless may
become a practical necessity.
4. Assume that a man harms one of his neighbours for the sake of some advantage, i.e. end of
his. The man who has suffered harm may wish to revenge himself. We need not here discuss
whether revenge is or is not in origin associated with ends in acting or with primitive ideas of
retributive justice. Revenge can, in any case, be said to have a natural, self-protective purpose.
Let us suppose that people never joined hands with others to revenge a wrong, unless they had
been wronged themselves. Even then it is easy to see how threat of revenge may assume an
accumulative deterrent effect on prospective evil-doers. Fear of revenge need not deter x from
harming y and z individually. It may nevertheless deter him from harming them both. For, if
revenge on x becomes a common interest of both y and z, they may co-operate and
successfully revenge themselves. Even if x is stronger than each of them individually, it is
unlikely that he will be stronger than both of them jointly. This is a consequence of the fact
that men are approximate equals by nature, and that they can co-operate for common ends.
Fear of revenge may therefore work as an effective deterrent at least against evil-doing on a
larger scale, i.e. involving several wronged agents. It may thus go at least some way towards
making observance of the practice of notharming a practical necessity.
I shall call evil caused by revenge for evil done the natural punishment of evil-doing. It can
also be called the natural sanction attached to the practice of respecting one's neighbour's
good.
The step from these logical preforms of law to the establishment of normative relationships is
not difficult to imagine. Let there be three men x,y, and Z, who constitute a community by
subscribing to our basic inequality of goods (see sect. 3). x and y are both anxious that z shall
pay his due, for otherwise they will not get their share in the greater good of the community. x
and y can command z, being jointly stronger. If commanding z, turns out to be necessary for
making him pay, co-operation in commanding becomes autonomously incumbent on x and y.
Thus x and y impose upon z the (heteronomous other-regarding) duty of paying.
-204-
Similarly x and z impose the same duty on y, and y and z on x. Thus the very same interest
which makes each one of the three co-operate with one of the other two against the third, viz.
anxiety to get one's share in a greater good which benefits them all, becomes the power which
forces each of them to pay their due. We can imagine the men as thoroughly selfish, void of
any sense of justice or morality whatsoever. They have not the slightest desire to pay their due
for their share. But the greedier they are on their share, the stronger will the normative
pressure become under which they are themselves to pay their due. This is a fascinating
mechanism. In cooperating for the common end of imposing heteronomous otherregarding
duties on others, men come to get these same duties heteronomously imposed upon
themselves.
The considerations which we have conducted, show how men's self-interested pursuit of a
common good may engender a practical necessity of adopting a practice which is to the
mutual advantage of them all.
5. Fear of revenge or of the punishing arm of heteronomous law is, of course, far from being
the only motive which may make men adopt the practice of not doing harm to their
neighbours. The circumstances, under which there is an advantage to be gained from harming
others, are special circumstances. In the life of some men they do, perhaps, not often arise.
Observance of the practice of not-harming because of lack of motives or opportunity of
harming is, of course, not more to the moral credit of men than observance from a motive of
self-interest.
The reasons why a man wants to respect a particular man's good, may be some particular
relationship in which the two stand to one another. Depending upon the nature of this
relationship, the reasons can be self-interested or altruistic. Perhaps he expects some good in
return from that particular man. Or perhaps he just loves him and therefore is anxious to
respect that other man's good 'for its own sake'. (I am here thinking of that aspect of love
which consists in an unselfish regard for the good of another being.)
Could unselfish regard for the good of all men be a reason why a man never does evil to his
neighbour? It must not be considered self-evident that, since a man can love some or other of
his neighbours, he can love them all. Love of man is different in kind from love which is a
relationship between particular men. One could
-205-
call the second, with Kant, pathological love. 'Pathological', of course, does not here mean
'sickly'. It refers to the foundation of this kind of love in a peculiar sentiment which one
human being feels for another. We cannot exclude the possibility that a man felt this
sentiment for every other man whom he happens to meet on his life's journey. His love of men
would yet be different from that love of man which, independently of relationships among
individuals, makes a man respect his neighbour's good 'for its own sake' or treat his neighbour
'as an end in itself'.
In order to get a firmer grasp of this 'abstract' love of man, let us go back to our problem of
the greater good of the community and the individual's share and due.
We have seen how wanting one's share can engender a practical necessity of paying one's due.
But we also know that paying one's due is not, as such, a necessary means to getting one's
share as an end. It becomes a necessity only thanks to the intervention of heteronomous norm-
pressure, one could say.
Suppose a man refuses to receive the smaller of two advantages, if he can have it only at the
expense of another man losing the greater advantage. He gives as a reason for his refusal, that
his receiving the advantage would hit his neighbour harder than it would benefit himself. Is
such comparison of goods possible? We have touched upon the problem before. We then
suggested that a balancing of goods against one another is possible only from the point of
view of one and the same valuating being. Therefore, if a man thinks that an action of his
injures another man more than it benefits himself and on that ground abstains from this act,
then he respects the good of that other man as though it were his own. This peculiar respect
for the good of another being as one's own is a motive which, without the intervention of
heteronomous norm-pressure, will put a man under an autonomous necessity to pay his due
for his share in the greater good of a community, of which he is a member.
It would be quite wrong to call this attitude to one's fellow humans self-interested. It would be
misleading, if not wrong, to give to it the name of self-love. But it could be called a man's
love of himself in his neighbour. This is the 'abstract' love of man, which I was contrasting
with the 'pathological' love of one man for another.
6. I have tried to show under what conditions there could exist
-206-
a duty to abstain from evil-doing. We must now stop to reflect on the nature of the argument
which we have been conducting.
That something is duty, we have said, means that it is a practical necessity with a view to the
welfare of some being or beings. To ask under which conditions something or other is duty,
entails asking what the wants of men must be in order that something or other shall be a
practical necessity.
It is not off-hand clear, which wants of men would necessitate abstention from evil-doing.
Mere wanting to escape harm from others cannot affect this -- for one thing because men
cannot as individuals command one another to reciprocal respect of one another's good.
Wanting not to harm others would not make abstention from evil duty -- wanting the very
thing is not practical necessity of doing the very thing. These considerations show that our
problem is not trivial.
The method which we employed for its solution was a general method. The derivation of the
duty to abstain from evil was only an illustration of how the method works in a particular
case. We shall now describe the method in abstraction from this particular application of it.
We started from the assumption of a community of men, who agree about that which we
called a basic inequality of goods. A certain good G
1
is for them (individually) greater than a
certain other good G
2
This basic inequality is subject to the following three conditions:
(i) If all men in the community relinquished the smaller good G
2
, then all of them will get the
greater good G
1
.
(ii) It is logically possible for a man to get both goods, the greater and the smaller one, for
himself.
(iii) It is logically necessary that, if one man gets both goods for himself, then some other man
will lose the greater good.
The existence of the greater good for one man we called his share in the greater good of the
community. Giving up the smaller good by one man we called his paying his due. In this
terminology, the first condition means that the greater good of the community
-207-
will exist, if each member pays his due. The second means that it is possible to get one's share
without paying one's due. The third means that, if a man gets his share without paying his due,
another man will lose his share.
To get both goods for oneself or, which means the same, to have one's share without paying
one's due, can be regarded as a basic form of injustice. The idea of justice has many roots and
has accordingly assumed many conceptual forms. We shall not here discuss the varieties of
justice. The only form of it with which we are concerned is embodied in the following
principle: No man shall have his share in the greater good of a community of which he is a
member, without paying his due. (It is essential that 'greater good', 'share', and 'due' are
defined in such a way that having one's share without paying one's due logically entails that
some other member of the community loses his share in the good.)
This Principle of Justice I would regard as the cornerstone of the idea of morality.
The Principle of Justice has as many applications as there are communities of men, for whom
a basic inequality of goods satisfying our three conditions holds true. I shall call any such
community a moral community of men and a duty to act in accordance with the Principle of
Justice a moral duty. Thus, e.g., in the community of men who have more to gain from never
being harmed by anybody else (in the community) than from sometimes harming somebody
else (in the community), a duty to abstain from evildoing would be a moral duty. This
restriction of a moral duty to a moral community we must discuss presently (see sect. 7).
The problem is now: How can action in accordance with the Principle of Justice become duty
in a moral community?
This can happen in two ways. We can call them the outer and the inner way. The first imposes
the moral duty as a heteronomous other-regarding duty on the members. The second imposes
it as an autonomous other-regarding duty.
Self-interest is sufficient as a moving force behind heteronomous moral duties. This is so,
because men's anxiety to get their share in a greater good of the community can engender a
practical necessity for them to pay their due. The logical mechanism works as follows:
Wanting one's share in the greater good of the community is a motive for joining hands with
other members to enforce a duty
-208-
to pay their due on such members, who might otherwise be unwilling. It is not a motive why
one should pay one's due oneself. But fear of revenge or of punishment for breaking the very
same laws, which are upheld by one's anxiety to make others pay, is such a motive. We can
imagine a community, in which everyone is anxious to add the smaller of the two goods of the
basic inequality to his share in the greater good, but is prevented from doing so by his even
greater anxiety to prevent anybody else from trying the same unjust trick lest he should
become the victim. It should namely be observed that from subscribing to the basic inequality
of goods it follows that one's anxiety to get one's share is greater than one's anxiety, should
one feel it at all, to profit from injustice. This is the very 'point' of the inequality: it provides a
check on a man's desire to treat others unjustly in the form of his even greater desire to be
treated justly by others. We can, in other words, make us a picture of a society, in which
justice and morality are 'kept going' -- even perfectly -- through mere self-interest. It is
essential that such a society could exist. In it men would observe their moral duties -- but not
from what we would call moral motives.
I shall say that an agent acts from a moral motive, when he observes his moral duty, neither
from a self-interested motive such as fear of revenge or punishment, nor from an altruistic
motive such as love or respect for the neighbour, whose welfare might become affected
through his action, but from a will to secure for all the greater good which similar action on
the part of his neighbours would secure for him. Action prompted by a moral motive is thus
beyond both egoism and altruism. The moral will is, in a characteristic sense, a disinterested
and impartial will to justice. Its impartiality, however, consists in treating your neighbour as
though his welfare were yours and your welfare his. Therefore the moral will is also a love of
your neighbour as yoursef.
When action takes place from a moral motive, observance of the Principle of Justice has
become autonomous other-regarding duty.
The question may be raised: Are moral duties, in the sense defined, ever heteronomously
imposed? Or is the possibility 'purely theoretical'?
The answer is that the possibility is not 'purely theoretical'. It is an important aspect of the
working of the huge normative power,
-209-
which we call the state, that it serves the interests also of moral communities of men. It would
be a great misunderstanding to regard the laws of the state as essentially void of a moral
content. The very example of a moral duty which we have been discussing, viz. the duty to
abstain from evil, is probably the most important single concern of the laws of all states. What
may nevertheless make us think of the law of the state as essentially different from 'the moral
law' is probably not so much the fact that there are branches of law which are not immediately
concerned with matters of public or private welfare, as the fact that, although laws of the state
need not be void of a moral content, the power of the state as an authority of norms is founded
on self-interested co-operation of men for common ends and not on a disinterested will to
observe the Principle of Justice.
Since self-interest can be the safeguard of morality, why should a man ever act from a moral
motive? To raise this question is to express astonishment at the fact that men should pursue
other than self-interested ultimate ends. But even if one cannot give reasons why men should
act morally from moral motives, one can try to make a man respect the good of another as if it
were his own by the use of arguments, which look like an appeal to ends. One can ask
questions like this, 'What right have you got to put yourself in a privileged position? If you get
your share without paying your due, then somebody else, who is equally anxious to get his
share, will necessarily be without it. Don't you see that this is unfair.' One could almost call
this appeal to a man's sense of justice an appeal to a man's sense of symmetry. 'If my wants
are satisfied at the expense of another man's, then why not his wants at my expense?' This is
like saying: 'For symmetry's sake you must want to be just. And for justice's sake you must
make yourself pay your due for your share in the good of the community.'
Such argumentation as this need not be without effect upon men. To resort to it is not
cheating. But it is not an appeal to further ends, for the sake of which we should aim at acting
justly. It is making the nature of justice as an end clearer by making us see logical
relationships between concepts.
That which is here called action from a moral motive has a certain resemblance to that which
Kant calls action from a motive of duty (Handeln aus Pflicht as distinct from pflichtmssiges
Handeln). The resembling features are those of disinterestedness and im-
-210-
partiality, and the detachment of the moral will both from selfinterested concern and from
altruistic sentiment.
The moral will is a will to do to others something which we want others to do to us. It is not,
however, a will to do this, because we count upon or demand a return of service. It is a will to
do this because we think it but fair that we too should contribute to the agreed good of all.
One could say that the moral will is a will to observe the principle known as the Golden Rule
from a motive, which comes at least very close to that which is a Christian love of our
neighbour. But one could also call it a will to treat our fellow humans as ends in themselves.
Here is another resembling feature with the ethics of Kant.
7. Someone may wish to raise the following objection against the argument which we have
been conducting: This argument makes the existence of moral duties depend upon the wants
of men. It may not be futile or uninteresting to try to describe the peculiar nature of the wants
behind such duties. What men want, however, is contingent. Is it not a most unsatisfactory
position in moral philosophy to let questions, whether something or other is moral duty,
depend upon such contingencies?
There are, moreover, two ways in which moral duties, on our understanding of the term,
depend upon the wants of men. The first is that moral duties exist only within moral
communities defined in terms of that which I have called a basic inequality of goods. The
second is that, even within the moral community, the moral duty exists only to the extent that
it is either autonomously imposed by the member upon himself or heteronomously imposed
upon him by others.
The restriction of moral duties to members of a moral community may seem very narrowing.
This impression, however, is deceptive. Let us go back to the basic inequality in the example
which we discussed, and consider how big the moral community might be, which it
determines.
The moral community under consideration consists of all those who have more to gain from
never being harmed by anybody else than from sometimes harming somebody else.
Some men may stubbornly refuse to acknowledge that they have more to gain from never
being harmed than from sometimes harming and yet be mistaken. But they may also be right.
In order
-211-
to see under which conditions they would be right, we must inspect more closely the nature of
the goods, which are balanced against one another in the inequality.
It is conceivable that a man never suffers any harm from his neighbours all through his life. It
is also conceivable that such a man has done harm to some other man for the sake of attaining
some end and that he attained his wanted end and never regretted it. This man has then gained
an advantage from harming his neighbour without having suffered a corresponding loss in the
form of revenge or punishment. Since our assumption was that he never suffers harm from
anybody else, there is no such harm which he could balance against this advantage so that he
might say to himself: 'I should have rather been without that which I gained from harming my
neighbours than have had this harm done to me.'
If these are the assumptions, does it follow that the basic inequality does not hold good for our
man? This does not follow. All that follows is that this man actually gained more from
harming others than he lost from being harmed by others. But this possibility is not denied in
the basic inequality. It says that a man has more to gain from not being harmed than from
harming. Therefore, in order to decide whether the basic inequality holds true for our man, we
must also consider how much he has gained from escaping harm. How shall this quantity be
estimated?
The harm which, on our assumption, the man has escaped, is the harm which would have
resulted to him had his neighbours never hesitated to wrong him, when this could have been
to their advantage. Escaping this harm is the advantage which a man has to gain from never
being harmed by anybody.
The advantage against which this has to be put in the scale is not the advantage which, on our
assumption, the man actually gained from doing harm to somebody. It is an advantage greater
than that, viz. the advantage which would have resulted had he never hesitated to wrong his
neighbour for the sake of anything which was coveted by him. This is what a man can be said
to have to gain from sometimes harming others. Unless our assumption is that the man is very
ruthless, we shall have to admit that he probably did not gain the full measure of what he had
to gain from doing evil.
In considering whether the basic inequality in our example is
-212-
true for a man or not, we have thus to balance against one another the following two --
negative and positive goods: On the one hand, the loss which a man can be expected to suffer,
if others take as much advantage from harming him as they can. On the other hand, the gain
which a man can be expected to harvest, if he takes as much advantage as he can from
harming others.
Having thus made clear the nature of the goods, which are balanced against one another, we
go back to the question, whether the basic inequality is true for all men or not. The question
is, in other words, whether the moral community which the inequality defines will comprise
the whole of mankind.
To this question is relevant that which we have said before about the approximate natural
equality of men and about the possibility of overruling natural inequalities by co-operation.
If there existed a giant among men, who could treat the rest of mankind as insignificant
worms, who can do nothing to harm him, then this man would not belong to the moral
community as determined by the basic inequality in our example. Men sometimes imagine
that they are such giants -- metaphorically speaking. They are usually soon taught that they
are not. Also when the lesson is ineffective, the truth may nevertheless be that they too have
more to earn from escaping harm, which could befall them, than from harming others.
Yet it is hardly an a priori necessity that an individual man should be weaker than all other
men jointly. (I say 'hardly', since we may wish to deny the name of 'man' to a being of
superhuman powers.) Suppose a man invents some fearful weapon, to which he alone has
access and by means of which he can wipe out any number of men who withstand his wishes
or encroach upon his privacy. He could be a kindly man and never want to do harm to
anybody. Could he be just and moral? I shall not attempt to answer the question. As long as
he keeps his secret weapon and is aware of the superiority, which it confers upon him in
relation to the rest of humanity, he, in any case, does not belong to the moral community,
which is determined by the basic inequality in our example.
The fiction of the superman, although logically possible, is yet highly unrealistic, someone
may say. Is it highly unrealistic? I would ask. Substitute, in our example, a team of men for an
individual and the example is less unrealistic. Substitute a state -- and we shall
-213-
recognize it as thoroughly realistic. The basic inequality from which we derived the moral
duty to respect the neighbour's good, holds true for practically all individual men. It is
therefore, even if not logically necessary, yet practically certain that the moral community,
which it defines, will comprise the whole commonwealth of men. This is a consequence of
men's natural equality and powers of co-operation. On the level of relationships between
states the same basic inequality is not undisputably true. Therefore it is not practically certain
that the moral community, which it defines, will comprise the whole commonwealth of
nations. For, although states can co-operate, they are not individually approximate equals. For
this reason, too, the problems of justice and morality in staterelationships are, to a great
extent, logically diferent from the problems of justice and morality in the relations of
individual men.
I hope to have succeeded in showing that the restriction of the moral duty, which we derived,
to a community of men is of no practical importance. For it is a practical certainty that no man
can put himself outside the community as we have defined it.
But what about the imagined superman, who is not in the community? Is it not his moral duty,
too, never to do evil to his neighbour? Must it not be everybody's duty? I would answer No.
The superman cannot be commanded to respect his neighbour, i.e. fear of punishment is no
motive for him to obey such a 'command'. Moreover, he cannot even feel that love of man,
which puts him under the autonomous law of respecting all other men. For this love 'of thy
neighbour as thyself' is conditioned by the similarity of needs and wants and powers of men.
If a being stronger than all men together shows benevolent concern for the welfare of humans,
this would be more like an act of mercy than an act of justice.
8. In conclusion, I wanted to raise the following question: In which sense, if any, can justice
and observance of moral duty be said to possess a 'utilitarian justification'?
The Principle of Justice, as we formulated it, says that no man must have his share in the
common good of a community, of which he is a member, without paying his due. Assent to
this principle autonomously imposes a moral duty on the members of the community to adopt
the practice of paying. The same duty may also become heteronomously imposed.
-214-
Observance of a moral duty, on this view, is a kind of 'public utility'. This means: To do one's
moral duty is a necessary condition for the existence of a common good of a community, of
which the dutybound agent is a member. In this sense, justice and the observance of moral
duty, trivially, has a utilitarian foundation.
From this does not follow, however, that the man who acts as justice and moral duty demand,
will fare better than the man who acts unjustly and neglects his moral duties. If all men in the
community act justly, they will all get a profit from their good conduct, which is necessarily --
this follows from membership in the community-greater than any profit they could have
gained from acting unjustly. But, as we have seen, it is also possible for a man to add to the
profits of justice the profits of injustice -- at the expense of another man's loss of the first
profit. This is a fearful possibility and its occurrence cannot be ruled out by logical argument.
Of course, to try to draw the extra profits of injustice and retain the profits of justice is an
extremely difficult game and it is probably right to say that few are successful at playing it. It
is even logically impossible that all men could thus profit from both justice and injustice.
(This can be deduced from the conditions, which the basic inequalities have to satisfy.) But
some can, theoretically.
The possibility of adding to the blessings of the reign of justice the profits of unjust action, in
short: the possibility of that which we have called parasitic action, constitutes an important
sense in which justice and morality can be said to be essentially void of a utilitarian
justification.
Justice and morality are of public utility in a sense in which, for reasons of logic, injustice and
immorality could never be. But injustice and immorality can be of private utility. If any
attempt to increase one's well-being by immoral means would necessarily be doomed to be
unprofitable, then moral action would be what it is not, viz. a practical necessity under
considerations pertaining to the agent's own welfare alone. Moral action would be
autonomous self-regarding duty.
There have been many attempts to 'reduce' the moral imperative to self-regarding duty. This
attempted reduction is, I think, one of the roots of the idea that there will be a last judgment,
when the unjust shall suffer and the just triumph.
-215-
I fear I cannot share this optimism -- if it be called by that name -- in the ultimate triumph of
justice. Injustice may as a matter of fact become revenged. Or it may become punished. But
from the standpoint of a secular morality it must remain a contingent matter, whether the
unjust man will prosper or perish.
-216-