Renford Bambrough Auth., Peter Geach, Jacek Holowka Eds. Logic and Ethics
Renford Bambrough Auth., Peter Geach, Jacek Holowka Eds. Logic and Ethics
VOLUME 41
R.M. Chisholm (Brown University. Rhode Island); Mats Furberg (Goteborg Univer-
sity); D.A.T. Gasking (University of Melbourne); H.L.A. Hart (University College.
Oxford); S. Komer (University of Bristol and Yale University); H.J. McCloskey (La
Trobe University. Bundoora. Melbourne); J. Passmore (Australian National Univer-
sity. Canberra); A. Quinton (Trinity College. Oxford); Nathan Rotenstreich (The
Hebrew University . Jerusalem); Franco Spisani (Centro Superiore di Logica e Scienze
Comparate. Bologna); R. Ziedins (Waikato University. New Zealand)
The titles published in this series are listed at the end of this volume.
Logic and Ethics
edited by
PeterGeach
ISBN 978-94-010-5481-2
Index 303
NOTE FROM THE SERIES EDITOR
This volume is to a large extent the result of the invitation for a term as
visiting professor at the University of Warsaw that Geach accepted in the
year 1985.
During this period Geach initiated a number of discussions on the subject
of Logic and Ethics, and it became apparent that a number of philosophers
both in Poland and in the United Kingdom were working in this area. Thus
the idea of a collective volume came into being. This became one of the
highlights of his visit to Warsaw at that time, and implemented with further
contributions from scholars invited by Geach resulted in the present book.
I.S.
1
Renford Bambrough
are sure that the two parties are fundamentally opposed while
others are still saying that there is a misunderstanding and they
are only at cross purposes.
Some conflicts are about the nature and incidence of conflict.
Wittgenstein speaks of the superstitious horror of a mathe-
matician in the face of a contradiction. There is a school of para-
consistent logicians. Can we preserve what they mean - even if
not what they say - and still remain convinced that nothing is
both true and false, that reason calls for the resolution of every
contradiction by the rejection of one or more of its elements? In
debating these issues, as in discussing most of the questions of
philosophy and morals and religion, we shall often need to
consider whether what I say is or is not in conflict with what you
say.
There is a story of a woman who kept two pug dogs, Caesar
and Pompey, and said of them: 'Caesar and Pompey are so alike
it's impossible to tell them apart - especially Pompey'. The story
is told as a joke. If the dogs are virtually indistinguishable, how
can it matter which of them is before us - matter, that is, to our
chances of knowing which is which?
But this joke is a good test of a person's attitude to more
important and wider-ranging questions. In the first place, people
divide more or less sharply into those whose reaction goes no
further than to laugh at the joke, and those who go on to wonder
what can have made the woman say what she said. The idea that
the woman was being merely stupid or confused will appeal to
one group more than to the other. The latter will look for a
plausible or at least charitable interpretation of what she said.
A first attempt might be this. Perhaps the only thing that
4
not make the most effective use of the many hands available for
chopping and shopping and mixing and carrying.
Philosophers divide on the same lines about what is to be
made of common sense, the common understanding, the
convictions of the plain man, the deliverances of the ordinary
moral consciousness and other sources and vehicles of our
inherited knowledge and beliefs.
Conflict between austere thinkers and ordinary men becomes
acute when, as in the case of Caesar and Pompey, questions of
logic (or of Logic) are at stake. Such austerity is what Walt
Whitman is repudiating when he glories in his contradictions as a
proof that he is large, he contains multitudes. And we may
mediate between the parties by asking whether he does
contradict himself. Can we not do for him what we have been
doing for the owner of Caesar and Pompey, for proverbial
wisdom? Sometimes we can, and sometimes we can do the same
for the plain man's common sense, and even for his moral sense,
his inherited values.
All this applies to some apparent contradictions that I wish
to diagnose and then resolve, or, if it should so turn out, to
characterise as merely apparent. I am concerned with some
questions rather than with any author or thinker, but it will help
me to say what I want to say if I refer now and from time to time
to a number of remarks by Bernard Williams: 1
The need and the desire for consistency with what we already
believe, the Platonic requirement of harmonia, is a constraint
that forms an important source of what Williams regards as the
illusion that moral truth is independent of us. The mistake is as
usual that of differentiating between theory and practice on
insufficient grounds. We must place the individual enquirer or
agent in the centre of the picture. When we do so we notice that
he will and must be beholden to his existing beliefs and values,
and in particular to the most fundamental among them, for any
response he can intelligibly make to a question or a dilemma,
conflict or bewilderment, theoretical or practical, that besets
him.
The more fundamental the beliefs and values are, the nearer
they are to being what constitutes the agent or enquirer. He could
not be identified and characterized without reference to these
deep commitments to beliefs and values, values and beliefs. Now
11
Notes
1 Bernard Williams, Problems of the Self (Cambridge
University Press, 1973) and Moral Luck (Cambridge University
Press, 1981).
2 Ludwig Wittgenstein, On Certainty, edited by G.E.M.
Anscombe and G.H. von Wright (Oxford: Basil Blackwell, 1969).
3 C.S. Peirce, Collected Papers, Vols. I-VI, edited by
Charles Hartshorne and Paul Weiss (Harvard University Press,
1931-1935).
4 C.I. Lewis, An Analysis of Knowledge and Valuation (La
Salle, Illinois: Open Court, 1946).
5 Bernard Williams, Ethics and the Limit of Philosophy
(London: Fontana Press/Collins, 1985) and 'Consistency and
20
NEGATIVE VALUES
Henryk Elzenberg
21
P. Geach (ed.) , Logic and Ethics, 21-31.
© 1991 Kluwer Academic Publishers.
22
for they have both to contain the generic concept of value and
some specific difference.) We have then two undefined concepts at
hand, and we cannot, of course, tell whether they are two species
of one genus. I.e., we cannot justify their common appellation of
'value'.
(We cannot then satisfactorily define negative value either in
terms of positive value, or in terms of value in general. Negative
value remains undefined. And then there is no way of telling why
it should be a value at all.)
Ad 2 and 9. The mutual exclusion of properties- assuming
their indefinability - does not turn the knowledge of the
positively value-generating character of the one into a knowledge
of the negatively value-generating character of the other. We
cannot then establish between the two concepts - of a positive
and a negative value-generating property - any relation at all.
For if an object that has property c is of value, then an object not
having c is merely without value,3 whatever the reason for its not
having it .
(Thus there is no way to show that having the opposite
property turns the object into one of negative value. An object
not having property c - whether because of having property d, or
for some other reason - does not thereby become of negative
value, but merely of no value at all.)
Our knowledge about positive values implies here a knowledge
about the negative ones, but not the other way round. For
suppose the object in question should not have d, and that having
on of those properties excludes having the other; in particular, to
have c excludes having d. This does not entail that the object
should have property c. For the debitum not to have property d
30
Annex
Notes
Peter T. Geach
33
P. Geach (ed.). Logic and Ethics. 3~8.
© 1991 Kluwer Academic Publishers.
34
punishes John'.
For some reason, this new account of the deontic operator ' 0'
gas prevailed in subsequent discussion; already in Prior's Formal
Logic (1955) this was the version of deontic logic that got
presented, and von Wright himself, for what motive I do not
know, came round to this formulation. Later developments of
deontic logic have introduced further devices and new ideas, but
this point has not come in for reconsideration: that the topic is
not what an agent is obliged to do or suffer or permit or abstain
from, but which states of affairs ought to be.
Here, to my mind, the fatal false step was taken. I am not
concerned here with tracing either possible influences on the
minds of theorists or the details of subsequent development: the
main thing is to establish that it was a false step. (It is perhaps
worth mentioning that E. Mally in his work of 1926, Grund-
gesetze des Sollens, had the same idea of a deontic propositional
operator - and the same rashness in constructing formal systems
regardless of any sensible concrete interpretation- as many
deontic logicians or our time.) To my mind, it was a false step
curiously similar to Aristotle's false step in the logic of
predication. Aristotle first made the category-<iistinction
between names and predicables, and then blurred it by setting up
the bogus category of terms; any attempted formalization of
predicate logic thereafter had to be defective until Frege
restituted the distinction between names and predicables (in
terms of arguments and functions: a name can only be treated as
an argument, not as a function), Similarly here: obligation
essentially relates to an agent, it is somebody's obligation; if
instead we try to think of the ought-to-be-ness, Sein-soUen, of a
36
Truth in Action
49
P. Geach (ed.). Logic and Ethics. 49-70.
© 1991 Kluwer Academic Publishers.
50
false assertion wrong in the sense that a bad action is wrong? But
we shall go further into this matter.)
In Wollaston's philosophy, the reason why non-linguistic
actions can be lies is that these actions have natural significance:
thus a wrong act signifies what is not the case. An example is
disposing of goods - by doing so, one signifies that they are one's
own, so that, if they are not, one lies.
Hume admits this much: "A person who takes possession of
another's goods, and uses them as his own, in a manner declares
them to be his own, and this falsehood is the source of the
immorality of injustice. But" he goes on to say "is property, or
right, or obligation, intelligible without an antecedent
morality? "3
This question is rather hard to answer: for in what way could
the relevant morality be expressed, without making use of these
concepts? We can see this by drawing an analogy with
man-made laws. Suppose one were instructed in court to give
evidence which was only about "the facts", i.e. evidence which
made no legal or moral assumptions. If one were really to obey
this instruction one would have to leave out words like
"property", "right", "obligation", except in quotation, or in
indirect speech constructions like "I thought that it was my
property". One would Similarly have to avoid the use of the
genitive in many of its senses, and many uses of the possessive
pronouns. One would have to avoid calling anything "money".
One could not refer to a man as one's host, since, if he is one's
host, that legally implies that one is only a licensee in his house.
One could not even speak of an event as having taken place in
the street, since it can be a question of law whether a particular
55
is not the case. Just as we could not paraphrase the law of the
land so that it contained no legal expressions but "It is legally
obligatory to ... " and "It is against the law to ... " so one could
paraphrase the moral law so that it contained no morally-
weighted expressions but "It is a bad act to ... " and "It is a good
act to ... " or the like.
If one takes another's goods, and treats them as one's own,
the truth to which one's act is contrary is that the goods belong
to the other. Now, people might say that "These goods belong to
Jones" means only "Anyone other than Jones who treats these
goods in a certain way is performing a bad action". However, this
analysis is only possible if the phrase "a certain way" can be
expanded and given a content which includes no reference,
explicit or implicit, to ownership. But this cannot be done, so it
is not clear that any circle exists here.
The way in which we cannot see law and morality (and,
indeed, cannot understand human conduct) without making use
of "value-laden" expressions like "rights", "marriage",
"commitment", "property" etc. is like the way in which we
cannot see the laws connecting the things of sense (and indeed
cannot understand the visible world) without making use of
expressions which suppose the existence of material objects. We
could not have a language which described the world in terms of
sense-data, though we have to see colours in order to see
material objects: and we could not describe the world of human
conduct in terms which missed out "value-Iaden t' concepts like
possession, though we have to see movements and hear words in
order to see where these "value-laden" concepts apply. 4
57
The first of these above facts shows us that when a man lies, he
makes a mistake on purpose. Now, when a man makes a mistake
on purpose, he usually alters the nature of his activity. For
example, if a musician deliberately plays a tune wrong, he is not
actually playing the tune, but an alteration ofit. And if a gunner
deliberately misses his target, he is not aiming at it. So why can't
a liar say when challenged "Oh, it's not meant to be true - it's a
lie"? Of course, it is meant to be true: that is just what makes it
a lie. So he means it as true when it is false.
We should notice also, that an assertion is an action. So
making an assertion is not like uttering a sentence. There are
64
showed how language had its meaning would be a theory, not just
of language, but of human life.
The wife of Philip Gosse thought that fictions were immoral, and
so her son Edmund (who grew up to be a literary critic) never
encountered a fairy tale until she died.
Now, I would say to Mrs. Gosse that she was wrong: fictions
are not lies, for a fiction is not a set of assertions. However, in
view of what is said above about assertions, Mrs. Gosse might
retort that I was guilty of circularity. For one asserts P if one's
action is such as would only be right if P is true. But if this is so,
a storyteller does assert what he says in his story, if Mrs. Gosse
was right, and fictions are evil.
Nor is it a peculiarity of mine to think that it belongs to the
explanation of assertion to say that it is only right if what is
asserted is true. (Dummett compares the truth of an assertion
with the winning of a game). Should one then say "The
"wrongness of a false assertions is not a moral wrongness"?
Mrs. Gosse might then point out that, under some
circumstances, it is the moral wrongness of an assertion which
makes it to be an assertion, as in my case of the false witness. So
morality seems to enter into the notion of "right" which we
employ when we say that an assertion is only right if it is true.
The circularity which Hume thought he could see in an
explanation of injustice as the using of another s property as
though it were own's own is a circularity which we might also see
67
Notes
1 See Hume's A Treatise of human Nature, Ed. L.A. Selby-
Bigge (Oxford 1960) Book III, Part I, section 1, p. 46l.
2 Ibid.
3 Ibid, p. 462.
4 See G.E.M. Anscombe, "On Brute Facts" in her
CoUected Philosophical Papers, Vol. III (Oxford, Basil Blackwell
1981), pp. 22-25.
5 See G.E.M. Anscombe, Collected Philosophical Papers,
Vol. II (Oxford, Basil Blackwell 1981), p. 116.
6 Op. cit., Vol. III. p. 28 .
7 Of course, there are other reasons for not uttering a false
proposition, than that doing so would be a lie. (There are
inappropriate jokes etc.) In the same way there can be other
reasons for not taking something, beside its being stealing (your
coat may be a man's only covering). This is one reason why an
ethical reduction of value-laden concepts is impossible.
5
Andrzej Grzegorczyk
l. Preliminaries
71
P. Geach (ed.), Logic and Ethics, 71-78.
© 1991 Kluwer Academic Publishers. -
72
components:
(4) a feeling of satisfaction (or dissatisfaction)
~ bc
a?
an intellectual judgement
a want or desire.
By the feeling of satisfaction we recognize a valuable state of
ourselves or a valuable relation to something else. This
recognition we express in a judgement of value and we realize it
by the desire of maintaining or obtaining this valuable state or
relation.
Analyzing the act of appreciation we make several
distinctions. I would like to mention the distinction between a
direct and an indirect act of appreciation:
(5) The appreciation is direct ifthe state or relation is
actually experienced. If not it is indirect.
The appreciation of eating a meal is direct. The negative
approach of threatening disaster (e.g. of possible earthquake or
nuclear war) is indirect. Likewise indirect are the appreciations of
freedom and security. Freedom is the possibility of performing
directly appreciated actions. Security is the removal of directly
negati vely appreciated states.
In order to sustain the reistic style of my considerations I
shall not introduce the notion of value as a noun, but only as an
adjective:
(6) An object X is for me (positively or negatively)
valuable if and only if I appreciate the object X
(positively or negatively).
Hence according to (5) there are directly and indirectly valuable
objects. The positively valuable objects are often simply called
values. But I prefer to speak of valuable properties and relations.
Moreover I shall limit my consideration to the properties of
74
W Properties of myself:
A. Satiation, warmth, cleanness, quiet, health, rest,
strength.
Physical fitness, coordination of behaviour, intellectual
ability.
B. Knowledge, comprehension, inner harmony, self-
control.
Relations to other persons:
A. Parentage, affiliation, erotic ecstasy.
Domination, submission.
Be1ongingness, identification with my own group and
distinction from others (strangers, deficients).
B. Respect, deference to others.
Justice, equality, autonomy.
Generosity, reconciliation, forgiveness.
W Relations to extra-human things but not to the world as a
whole. sometimes relations to the people conceived as
impersonal:
A. Comfort.
PosseSSion, manipulation, control.
B. Perception of beauty and deeper order.
Creation of beauty and deeper order.
Relations to the world as a whole and to the supposed
mechanism of it:
A. The attitude of pretension, claim, pride, self-conceit.
B. Humility, acceptance of my destiny and of external
reality, belief in the meaningfulness of existence.
after the letter B. refer rather to what may be called our spiritual
life.
The above list mentions rather positively appreciated objects.
Positive cases are often names of the whole scale. E.g. 'health'
may denote the whole scale: health-illness, but 'illness' does not.
In (b) I should like to consider only relations between
individual persons, or between a person and a group of persons.
The treatment of a human being as a tool, or treatment of the
enemy army as a disaster, should belong to point (c).
Distinguishing (d) as a separate case is justified by the
observation that the attitude of pretension, grievances, claims
and pride have a tendency to generalization upon the whole of
reality. Also the faith in the value of other human beings and
deeply experienced acceptance of a person have a tendency to be
generalized upon the whole of reality and lead to the acceptance
of reality as a whole and to the acceptance of one's destiny. The
attitudes mentioned in (d) may be called numinous.
The most relevant reflection concerning the above logical
classification of directly valuable objects refers to the distinction
between vitally (or biologically) valuable objects mentioned
under A. and the others mentioned under B. which may be called
spiritually valuable. This distinction leads to the formulation of
the principle of transcendence.
Conclusions
J. Harrison
79
P. Geach (ed.J, Logic and Ethics, 79-129.
© 1991 Kluwer Academic Publishers.
80
"if keep your promises" does not make sense. Hence we cannot
think that someone else can be mistaken, when he issues an
imperative, or that we ourselves may be mistaken when we issue
an imperative.
Though someone saying "Keep your promises" cannot be
right or wrong, he can be right to say "Keep your promises".
Whether a person who has said something was right (because,
say, he said "Promises ought to be kept" and promises ought to
be kept) is quite a different question from the question whether
he was right to say it. If I am playing a guessing game with a
disturbed child, and I guess that it is in her left hand, and it is,
then I am right (as is my guess) but it may have been wrong of
me not to say that it was in her right hand (because it would give
her pleasure if my guess was wrong). And, if I say to a lady that
she is too fat, I can be right, because she is too fat, but quite
wrong to say it, because doing so gives her pain. Moral
judgements, like assertions, can be assessed according as to
whether the person who makes them is right, and according as to
whether he is right to assert them. Imperatives can be assessed
only in the latter way.
I have said that we can think that a man asserting a moral
judgement is right if we think things are as he says they are. It
could be objected that things cannot be as he says they are,
because he has not said how things are, but how they ought to
be. This is a mistake. Someone who is saying how things ought to
be is saying how things are; he is saying, so to speak, that it is so
that promises ought to be kept. That promises ought to be kept
is the way things are. He is asserting that it is the case that
promises ought to be kept, or that promises ought to be kept. It
84
keep your promises" and "Keep your promises" is, 1 believe, this.
Since one normally says each of these in order to make someone
keep his promises, one usually cannot say the one, but the
opposite of the other, without being practicaUy inconsistent, i.e.
furthering a policy by saying the first which one is hindering by
saying the second. Since a policy is always adopted as a means to
an end, one need not be behaving in an inconsistent way if one
has unusual ends - for example, if one wants to confuse someone
or weaken him. And "You ought to keep your promises, but don't
keep them; the effort might be too much for you" is unusual, but
not exactly inconsistent, for not everybody is concerned so
exclusively about others' moral welfare as to make it inconsistent
to tell someone to do what one is at the same time telling him
ought to be done, (I shall show later that Professor Hare's view
that moral judgements imply imperatives must be mistaken,
because imperatives cannot be implied. His view that moral
judgements imply imperatives, together with another view of his,
that it is impossible to derive an imperative from a set of
premises which do not contain an imperative, (and a set of
premises which contains only one moral judgement and nothing
else would contain an imperative only if that moral judgement
was an imperative) implies that moral judgements are
imperatives.
It may also be the case that someone saying "Promises ought
to be kept is contextually implying that he has, or that he believes
he has, a disposition to issue imperatives to other people
enjoining promise-keeping upon them. (What is contextually
implied is not the imperative "Keep your promises", but that the
speaker has the disposition to issue such imperatives.) It follows
86
that someone who says "Men ought to keep their promises", but
does not have the disposition to issue imperatives such as "Keep
your promises", does not sincerely think that men ought to keep
their promises, as someone who says "It is raining", but does not
believe that it is raining, is being insincere. But the statement
that anyone saying "Promises ought to be kept" has the
disposition to issue imperatives enjoining promise-keeping must
be much qualified. For one might be disposed not to issue such
imperatives because on thinks that people have a right to go to
the devil in their own way, thinks that constantly telling them to
keep their promises may have the opposite effect to that which is
intended, or thinks that it is improper or unwise to issue
imperatives to others except in unusual circumstances.
In the following pages I give twenty-nine further reasons for
rejecting the theory that moral judgements are imperatives. The
jejune nature of some of them simply underlines the
implausibility of the view that is being attacked. As I have said, I
have as a subsidiary target the belief that there is such a thing as
imperative logic. Indeed, one argument against the view that
moral judgements are imperatives is simply this: there is no logic
in which the conclusions and premises are imperatives; therefore
moral judgements are not imperatives. The reasons for thinking
that moral judgements are not imperatives are as follows.
1. It is correct English to speak of moral judgements, not of
moral imperatives. A judgement is just a species of proposition or
statement the assertion of which involves judgement, good sense
or the ability to estimate, as opposed to involving apodeictic
proof, intuitive certainty, or accurate measurement.
2. We can have the same array of attitudes - i.e.,
87
believed.
One can, of course, make mistakes in action as well as in
thought, and the former do not consist in having mistaken
beliefs. However, making a mistake in action cannot be what
holding a mistaken moral belief consists in. A mistaken action
may be produced by a mistaken belief, but it is not itself a
mistaken belief. One may be doing something mistaken when one
issues an imperative, and one may perform this mistaken action
because one has a mistaken belief that it is the right imperative
to issue, or that it will produce a desired result, or that to issue
that imperative is ordained but a given piece of ritual, but the
imperative itself, as opposed to the act of issuing it, cannot be
mistaken. Moral judgements, however, can be mistaken, as well
as its being possible for one's asserting one to be a mistake.
There can be reasons for thinking that one ought to do
such-and-such. There can be no reason for thinking (that) do
such and such. You cannot think do such and such, and a fortiori
cannot have reason for thinking "Do such-and-such".
Imperatives themselves do not belong to the category of things
for which there can be reasons. There can, of course, be reasons
for issuing imperatives, but moral judgements resemble
statements, in that they can be assessed both according to
whether there are reasons for (believing) them, and according to
whether there are reasons for asserting them out loud.
10. Moral judgements (sincere ones, at any rate) are moral
beliefs spoken out loud. Most moral beliefs are seldom, and
sometimes never, uttered. The imperative theory must have great
difficulty in explaining those moral beliefs which we do not
publish.
91
about them).
(One element of truth in the view that ought-judgements
must be action-guiding is that if it is true that a person ought to
do such-and-such in order to achieve some end that he wants,
there need be no extra motive for doing such and such, over and
above his wanting the end to which his doing such- and-such is
the means in order for him actually to do such- and-such.)
12. It has been held that only imperatives can be "action-
guiding"; beliefs are cold and inert, and cannot be. Moral
judgements, since they are action-guiding, must be imperatives,
and not beliefs.
This is the absolute reverse of the truth, however. Beliefs are
only inert when they do not inform me how to satisfy my desires.
Since there are desires to do what one ought, or to obtain those
ends which will result from doing what one ought, moral beliefs,
which are about what one ought to do, are not inert. Imperatives,
on the other hand, do not move us to action (unless perhaps
sometimes our obeying them is species of conditioned reflex);
what does move us is the belief that an imperative has been
issued, coupled with a desire to do what has been enjoined.
Advice, for example, does not by itself move us to action; what
moves us to action is the belief that we will lose something we
want if it is not followed. The fact that imperatives cannot move
us to action, though moral beliefs can, is reflected in the fact that
I cannot sensibly say "I did it because ... " where the gap in the
preceding sentence is filled in with an imperative, but I can say
"I did it because I thought I ought and wanted (for once) to do
what 1 (thought I) ought. (I can, of course, say that 1 did it
because someone had ordered me to, but this is an assertion
94
modified form . One could argue that, if one says to one person
"Do A" and to another person "Do not do A" then one is being
inconsistent, in the sense of performing inconsistent actions, if
one believes that there is no difference between the two people,
the actions they are asked to perform, and the circumstances in
which they are both placed.
There are two difficulties with this contention. What is
inconsistent, and what it is inconsistent to say, are two entirely
different things. If I say "There were survivors" to my own
superior officer, but "There were no survivors" to an enemy
interrogating me, I am asserting inconsistent propositions, but
my saying one thing to one man and another thing to another
may be part of a perfectly consistent policy of telling the truth to
one's friends, but not to one's enemies. Though the propositions
asserted are inconsistent, the actions of asserting them are not.
It is fairly obvious that I am being inconsistent in the first
way (being theoretically inconsistent) rather than in the second
way (being practically inconsistent) if I say to one man that an
action is right, and to another man that the same action is
wrong, although I believe that there is no other difference
between these two actions such as to make the one right and the
other wrong. For, as we have seen, I need not be being practically
inconsistent at all in saying one thing to one person and its
contradictory to another. The first person may be my employer,
and the other my wife, and it may be important for me to keep
on good terms with both of them. Whether I am being
theoretically inconsistent will depend upon my beliefs about the
circumstances concerning the action judged of; whether I am
being practically inconsistent will depend upon my beliefs
96
when, before the Ides of March, I said that what Brutus was
going to do was going to be wrong. There is no form of words
which makes it possible usefully to repeat the imperative
"Brutus, do not assassinate Caesar" after Brutus has assassinated
Caesar.
16. Not only are moral judgements neutral with respect to
time and place: they also have the feature of being neutral with
respect to the person who asserts them. It does not matter who
asserts the moral judgement that killing people is wrong. If the
person who asserts it is right, any person who asserts the same
judgement (though not necessarily in the same words) must also
be right.
It is, of course, true that if two men both say "It is right for
me to drive on the left hand side of the road" one can be right
and the other wrong. This is because, though they are using the
same sentences, they are, since they are different people driving
in different countries, using them to make different assertions, so
there is no reason why what the first man says should not be
right, and what the second man says be wrong.
It is impossible for the condition (which is controversial) that
I have just stated to be fulfilled by imperatives, because it is
senseless to say of a person who issues one that he is right (or
wrong) at all. One can say that a man is right (or wrong) to issue
an imperative, but it is not even remotely plausible to say that, if
one man is right to issue an imperative, any other man issuing
the same imperative must be right to issue it. One may be
entitled to issue orders to one's own children that one is not
entitled to issue to anyone else's children.
The truth of what I have just said presupposes that the same
100
said to the one who said that Brutus was right to assassinate
Caesar that he was right, he could be right. This would follow
from the general principle that, if I say of someone who says what
is right (or true) that what he said is right (or true) I must be
right (or what I say must be true). The remark IIThough Brutus
was wrong to assassinate Caesar, you, when you say that he was
not wrong to assassinate Caesar, are right II is absurd.
Furthermore, if I say to someone who has just said that
killing people is wrong that I agree with him, I cannot
consistently say that killing people is not wrong. If we could both
be right, however, it would not be inconsistent to say this.
Imperatives cannot have the above-mentioned feature,
because I cannot say to someone, who has just issued an
imperative, that I agree with him at all. If I do so, I imply that I
think that when he says this he is expressing an opinion which is
right. If I disagree with him, I am then saying that he is
expressing an opinion which is wrong. If one man says to another
"Shut the door II , it would be entirely inappropriate to reply "I
agree". One can say II Yes II , but this means IIYes, I will do what
you tell me toll, not "Yes, what you say is SOli. A man and his
wife can agree in the sense that what one wants them to do is
what the other wants them to do, but this is compatible with a
vast amount disagreement of opinion on academic matters,
(including moral matters, provided they do not attempt to put
different moral opinions about what they should do into
practice). Wanting the same thing, however, as opposed to
wanting them both to do the same thing, is a source of disagree-
ment rather than of agreement. If Jack Sprat and his wife had
wanted the same thing, e.g., lean, instead of one wanting lean
102
and the other wanting fat, they would have got on much less well
than they did.
It is a feature of moral judgements, which they share with
indicatives, that if a person has made a moral judgement, any
other person can either agree or disagree with that judgement. If
someone says, for example, that Brutus was wrong to assassinate
Caesar, anyone at all may agree with him or disagree with him,
and say that he does. That is to say, anyone at all can say "Yes,
I agree", when someone has made a moral judgement. Only the
person to whom an imperative is addressed, however, can say
"Yes, I will do what you ask" or, alternatively, "Shan't".
18. Moral judgements and imperatives have different
contextual implications. My commanding someone contextually
implies that I have or believe that I have power or authority over
him. My requesting someone contextually implies that he is or
that I believe that he is in a position to do something that I want
him to do. My advising someone contextually implies that I am
or believe I am in some respect wiser, or possess more expertise,
than he does. Neither making statements nor making moral
judgements has these contextual implications. I can say to
someone "You ought to keep that promise" without contextually
implying that I believe I have power or authority over him, that I
believe that I am wiser than he is, or that I believe that he has
power to do something that I want him to do. (Contextual
implication differs from ordinary implication in that what does
the contextual implying is my saying something, not what I say.
It is my saying "It is raining" that implies that I believe it is
raining, so that I believe it is raining can be false, even though
that it is raining is true. Hence my issuing an imperative can
103
false.)
One might wish to argue that, though the principle of
bivalence does not apply to imperatives, an analogue of that
principle does. Just as every proposition is either true of false, so
every imperative must be either satisfied or not satisfied. (I am
using the word "satisfied" to cover a heterogeneous collection of
actions like obeying a command, complying with a request,
taking advice, following instructions, etc., though in what follows
I shall for the sake of simplicity sometimes consider only
obedience or disobedience to commands.)
This analogy with the principle of bivalence is not as close as
it might seem, however. Every proposition, whether asserted or
not, is an instance of the principle of bivalence. I doubt very
much, however, whether there are such things as uncommanded
commands. It seems as difficult to hold that there are commands
which have never been issued as it is to hold that there are
headaches that no-one has ever had. If there were, every
individual, every moment of his life, would either be obeying or
disobeying an infinite number of unexpressed commands, just as,
every moment of his life, he is the subject of an infinite number
of unasserted true of false propositions. But though it is true of
me that either I am sitting in my chair or true of me that I am
not sitting in my chair, it is not the case that I am either obeying
the command "Sit in your chair" or disobeying it. Even if there
were unissued commands, as there are unasserted propositions,
unasserted propositions can be true or false, but unissued
commands cannot be obeyed or disobeyed.
Furthermore, what is the analogy with the principle of
bivalence, even when it is limited to issued imperatives? Is it the
106
Though "If don't shut the door, don't shut the door" does
not make sense, "Either shut the door or don't shut the door"
does. The formula in the calculus of propositions that the first
allegedly instantiates, however, (the formula "p -+ p") is
equivalent to the formula the second allegedly instantiates
(pv-p).
This anomaly is only apparent, however. If we use the
artificial expression "make it the case that", "Make it the case
that if the door is shut the door is shut", is equivalent to "Make
it the case that either the door is shut or the door is not shut".
This suggests that if p, then do A, is not of the form "Make it the
case that if p then q" but of the form "If p, then make it the case
that q." "If p, then do A" does not normally "command a
conditional" (p -+ q) so to speak, but conditionally commands a
categorical (q). This is why "If p, do A" (make it the case, if p,
that q) is not equivalent to "Do -A or B" (make it the case that
-p v q). "Either -p, or make it the case that q" does not even
make sense. Any expression of the form "Make it the case that pIt
is equivalent to any expression of the form "Make it the case that
q" if p can be shown to be equivalent to q by means of the
calculus of propositions. At least, what we are being told to make
the case is equivalent.
In this respect hypothetical imperatives (not hypothetical
im perati ves in Kant's sense) resemble hypothetical predictions,
e.g., "If it rains, England will win". Here I have not committed
myself to anything true or false if the antecedent is false. I use a
form of words which is such that I am committed to something
(England's winning) only if the "antecedent" is true. If I am
arguing with someone else, who thinks that England will lose if it
108
door" and "Shut the door because "but the function of the
words "so", "therefore" and "because" is not in this context to
indicate that "Don't shut the door" is a conclusion. They have
some other function, which philosophers have confused with that
of indicating that what follows them is a conclusion.
24. Not only can imperatives not be conclusions. they
cannot be premises either, whereas moral judgements can be
premises. This follows because (a) they cannot be true, and
because a fortiori (b) they cannot imply anything in such a
manner that it can be truly said that if they are true, their
alleged implicates are true. If one wishes to say that moral
judgements cannot be true, it must be pointed out that they can
be reiterated, and that if they are reiterated, their conclusions
may be reiterated. For example, you can say "if it is Sunday, I go
to church, and it is Sunday", instead of saying "It is true that it
is Sunday" . Imperatives, as we have seen, cannot be re-iterated.
Since imperatives cannot imply anything at all in the
required sense, they can imply neither other imperatives nor
indicatives, in that sense of "imply" in which the premises of a
valid argument imply its conclusion.
It may be instructive to consider a case where it has been
held that imperatives do imply indicatives. A king says to one of
his courtiers "If you are a faithful subject kneel down; but rise
up". From this it has been supposed to follow that the courtier in
question is not a faithful subject.
Even if the king had uttered the indicatives "If you are
innocent you will not drown", and "You will drown" one would
not be able to conclude from his saying these things that the man
addressed is guilty. You would be prepared to deduce this
112
conclusion only from the truth of what the king said, not from,
his saying it. If you conclude that the man in question is not
innocent, you deduce this from facts which the king has
communicated to you, and which you accept on his authority. In
the alleged example of imperative argument mentioned above,
the king's remarks cannot be true. Hence the king cannot
transmit to anyone else something true which this other person
can then proceed to use as the premises of an argument .
Imperatives do not enable people to transmit premises.
Where indicatives (including moral) judgements are
concerned, it makes no difference who has communicated to us a
piece of information which we use in order to draw our
conclusion. In the celebrated example where one man (a priest)
says that his first penitent was a murderer, and another man says
that he was the priest's first penitent, it does not matter that the
two pieces if information which we use as our premises have been
given us by different people; we arrive at our conclusion just the
same.
However, if one man tells me to shut the door, and another
man tells me to open the window if I shut the door, what can I
conclude? The most one can "conclude", so far as I can see, is the
indicative that if one obeys the first man's command, and shuts
the door, one will then be disobeying the second man's command
if one does not then open the window. (I do not conclude this
from these commands, but from the fact that these commands
have been issued, which I know because I heard them being
issued.) If, however, both commands are issued by the same man,
one's conclusion is different (though one's conclusion is not
different if indicatives have been issued by the same man or
113
rained" and "It rained" does not entail that he said that England
would win. What he has done is to say something (i.e. some one
thing) which is such that if it does rain and England does not win
it is false. Parallel to this is the fact that someone who says
"Take all the boxes to the station" has issued an imperative
which is such that, if this is a box and this has not been taken to
the station, it has been disobeyed. Someone who says "Take your
umbrella if it rains" has issued an imperative which is such that
if it does rain, and the person commanded has not taken his
umbrella, it has been disobeyed. The function of logic here is to
work out what counts as obeying (or complying with), and what
counts as disobeying, this one imperative. There is no question of
a different imperative, e.g. the imperative "Take your umbrella",
following from it. For this reason obedience cannot be handed on
from "premises" to conclusion of an alleged imperative argument.
For though an unasserted proposition may be true, an unissued
imperative cannot be obeyed (or complied with, or whatever).
What I have just been saying does not apply to moral
judgements. "All husbands ought to support their wives" and
"Smith is a husband" do together entail "Smith ought to support
his wife", and "If Mary has a child, she should look after it"
together with "Mary does have a child" do entail that she ought
to look after it . It is entirely irrelevant whether anyone has
asserted the moral judgement in question or not .
I do not, however, wish to say that logic is irrelevant to
imperatives. I have said that you cannot conclude "Take this box
to the station" from "Take all the boxes to the station, and this
is one of the boxes". What you can do is, from hearing someone
saying to you "Take all the boxes to the station" conclude that
119
one is to take all the boxes to the station, and then argue that,
since this is one of the boxes, one is to take this box to the
station. (Actually, I do not think that one can even conclude this.
All I think one can conclude is the subtly different fact that this
is one of the boxes which one is to take to the station.)
Again, one cannot conclude "beat him" from "Beat him if he
sneezes", and the fact that he has sneezed. What one may
conclude from the facts that one has been told to beat him if he
sneezes and he has sneezes, is that one is to beat him. (One
cannot conclude that one has been told to beat him, for this
categorical command has not been issued.)
Where moral judgements are concerned, as with other
indicatives, one can conclude that one ought not to kill Smith
from the fact that one ought not to kill people, and Smith is a
person. One concludes this directly, from "One ought not to kill
people", not from the fact that someone has said "One ought not
to kill people".
Logic is also relevant to imperatives in the sense that some
imperatives are incompatible with others, and some imperatives
make others redundant. If I assert that all men are mortal, there
is no logical need (though there may be a psychological one) for
me to assert of Smith, a man, that he is mortal. Similarly, if I tell
someone to take all the boxes to the station, there is no need to
tell him also to take this box to the station. The first proposition
and the first command make the second proposition and the
second command redundant. Again, just as, if I assert that all
men are mortal and that Socrates is a man, I cannot consistently
assert that Socrates is not mortal, so, if I tell all the men in the
room to stop smoking, I cannot consistently permit Smith (a man
120
gaps in the schema "If ... , then ... " and get a result which is
necessarily true, or even makes sense, and the fact that there is
no way of distinguishing between its being a valid inference and a
proof by saying something analogous to what one certainly can
say about an indicative inference, viz., that not only does its
conclusion follow from its premises, its premises are also true,
there are other defects as well.
The main one is that its minor premise does not contradict
one of the disjuncts in a disjunctive command, as the minor
premise in the allegedly analogous indicative inference, the
modus tollendo tollens, contradicts one of the disjuncts in the
disjunctive major premise. It is, indeed, incompatible with this
command. This is because posting-the-Ietter-or-burning-it is
the intentional object of the command, which intentional object
is, so to speak, single and indivisible. Whereas what is asserted in
the major premise of the modus tollendo tollens is that there are
two things you can do, post the letter on the one hand or burn it
on the other, there are three things you can command, posting
the letter, burning it, and indifferently posting it or burning it.
Commanding any of these three things is incompatible with any
other. Hence issuing the command analogous to the minor
premise in the corresponding indicative argument does not, as it
does in the indicative version of the modus tollendo ponens, deny
one of two possibilities and so make necessary the truth of the
other. For in the command to post the letter or burn it there are
not two possibilities, but one disjunctive one. Hence anyone who
first says: "Post the letter or burn it" and then says "Don't burn
it" has done something that someone who first says "The letter is
burnt or posted; the letter is not burnt, therefore the letter is
122
posted" has not done, i.e., changed his mind. Since he has
changed his mind, he must have withdrawn the original
command "Post the letter or burn it", and hence one cannot infer
that one is disobeying any command which remains in force if one
does not post the letter.
In this respect, incidentally, moral judgements resemble
commands. The statement (a) that one has a duty to burn-
the-book~r-post-it, it does not matter which (O(p v q)), would
be incompatible with the statement (b) that one has a duty to
burn the book (Op), and with the statement (c) that one has a
duty to post it. This is because (a) does not entail that you would
not be doing your duty if you posted the book nor that you would
not be doing your duty if you burnt it, while (b) does entail that
you would not be doing your duty if you posted it, and (b) entails
that you would not be doing your duty if you burnt it. Hence
anyone trying to argue: You have a duty to post it or burn it,
you do not have a duty to burn it, therefore you have a duty to
post it, would be using a minor premise, and coming to a
conclusion, both of which were incompatible with his major
premise. It may be that this alleged argument has been regarded
as satisfactory because it has been confused with the perfectly
valid argument: Either you have a duty to post it or you have a
duty to burn it (Op v Oq); you do not have a duty to burn it
(--op); therefore, you have a duty to post it. The major premise
of this latter argument is a genuine disjunctive proposition, not
an atomic proposition attributing to someone a disjunctive duty.
The element of truth in the view that: post the letter or burn
it; don't post it, therefore burn it, is a valid argument is that the
person to whom these commands are addressed (or, indeed, any
123
other person) can deduce something from the statement that they
have been issued, which he knows to be true because he heard
them being issued; he can deduce that the only way in which the
person addressed can do what has been enjoined by both
commands is by posting the letter. But again, he deduces this
indicative conclusion not from the imperatives, but from
propositions stating that these imperatives have been issued.
I pOinted out earlier that, though one could not say "I
conclude burn the letter" (or "I conclude take that box to the
station") one could say "Therefore post the letter" or "So post
the letter" or "Post the letter because .. . " I shall now explain
why this is so. I believe that the reason is of considerable
importance. It is that, though when I say "conclude that" I am
indicating that I have been giving reasons for coming to the
conclusion that something or other is so, when I precede "Post
the letter" with "so" or "therefore", or follow it with "because", I
am not giving a reason for concluding that anything is so, or
even, impossibly, for "concluding an imperative", but giving
reasons for an action, the action enjoined by the imperative.
When I say "So post the letter" or "Therefore post the letter" or
"Post the letter because ... " I am not indicating that I am giving
reasons for the imperative "Post the letter", but for the action of
posting the letter. If, impossibly, the reasons that precede the
words "Post the letter" were reasons for a conclusion at all, then
the conclusion would be an action - Aristotle thought actions
could be conclusions - rather than a proposition or imperative.
Even when I say "If you ought to do it, do it, and you ought to
do it" I have not given reasons for the imperative "Do it" , but for
the action of doing it.
124
reasons have been given for the proposition, but "therefore" and
"so" can precede either propositions or imperatives. When "so"
and "therefore" precede imperatives, however, they are not
reasons for the imperative, as the unwary might suppose, but for
the action enjoined, advised, recommended or directed by it.
It does not, however, follow from the fact that, when we issue
an imperative, and back it up with reasons, we are giving reasons
for the action enjoined by the imperative, not for the imperative
itself, that moral judgements are not imperatives. This is
because, when we way that an action ought not to be done, and
back this statement up with reasons, we are also giving reasons
for an action, the action we say ought to be done. But, in the
case of judgements about what ought to be done, at any rate, the
reasons which back up such judgement are also reasons for the
action which it is said ought to be done. For example, if I say to
Smith that he ought to visit his aunt, because he promised he
would, that he promised he would is just as much a reason for the
action of visiting his aunt, as the reasons which precede the
imperative, "Visit your aunt" are reasons for the action. (It may
be just for this very reason that some modern moral epistemo-
logists have supposed, wrongly, that moral judgements are indeed
imperatives, or something like them.) But that he promised he
would is a reason for thinking that Smith ought to visit his aunt,
as well as for his visiting his aunt, whereas there can be no such
things as reasons for "Visit your aunt". To say that there were
would be to utter a piece of gibberish.
The only way I can think of to explain this apparently
anomalous fact is by saying that what ought-judgements state is
just that there are reasons, of good reasons, or conclusive reasons,
126
for the action which it is said ought to be done, or that there are
better reasons for doing it than for omitting it. (I am not sure
which of these would be correct.) To say that I ought to move
my pawn to QR4 is just to say that there are good or conclusive
reasons for moving my pawn to QR4. 1 show that such
judgements are true, then, just by showing that there are good
reasons for the action of moving my pawn, which I really do by
producing the reasons.
One must then look not for reasons for the judgement, but
reasons for the action. To suppose that that 1 promised I would is
an immediate (and academic) reason for believing that 1 ought to
go, rather than an immediate reason for going (and so, mediately,
a reason for believing that 1 ought to go) is to make a mistake.
On the other hand, to suppose that ought judgements are
imperatives, because they are not backed up by immediate
(academic) reasons for believing them, but by practical reasons
for doing the things which they allege ought to be done, is
another mistake. One can steer a course between these two rocks,
by saying, quite simply, that you do in fact back up an ought-
judgement, not by academic reasons for believing it, but by
practical reasons for acting on it, for the very simple reason that
the ought-judgement just says that there are reasons for
performing the action that it says ought to be done. Hence one
shows that a moral judgement is true by producing reasons for
the action, and not the judgement, because what the moral
judgement says is just that there are such reasons. It is obvious
that imperatives cannot state that there are reasons for the
actions they enjoin.
Hare founders on both these aforementioned rocks
127
Bibliography
Jennifer Jackson
The Problem
According to what I shall call the liberal view, the doctor should
neither impose nor be imposed upon, should neither be
obstructive to the patient's wishes nor be obliged by her to do
what offends his conscience.
In other words, the doctor should not try to dissuade (a). But
he may refuse to advise (b), provided he also refers to a willing
colleague (c), since unless the patient knows that she can get
advice she wants elsewhere, this doctor simply by refusing to
131
P. Geach (ed.), Logic and Ethics, 131-144.
© 1991 Kluwer Academic Publishers.
132
Tolerating Injustice
Asymmetry
really credible that all the horrific practices which are said to be
or have been customary in this or that society can be discounted
as traveller's tales, lent credence only by naive colonialists? And
would those horrific practices that are now extinct not still be
with us if some had not seen fit to forcibly put an end to them
e.g. the annual "customs" (of human sacrifice) celebrated in the
city of Benin until 1897, the Slave Trade, the Press Gangs?
Notes
1 Another version of this paper, title Moral Toleration, was
presented at the Annual Conference of the United Kingdom
Association for Legal and Social Philosophy, April 1986, and was
published in Mark Ockelton (ed), Medicine, Ethics and Law,
Steiner, Stuttgart (1987) together with a reply by R.D. Mackay
(Department of Law, LeIcester Polytechnic).
8
Jacek Holowka
145
P. Geach (ed.), Logic and Ethics, 145-165.
© 1991 KLuwer Academic Publishers.
146
with partial satisfaction of the claims of all parties. But unlike the
zero-sum games they do not have neat mathematical solutions.
They can only be solved with the use of additional postulates. Such
solutions can yield illuminating and ingenious results, but the
validity of such methods is rather limited, I think. We have no
independent criteria of choosing between different conceptions of
justice.
Let us consider two famous examples, the Prisoner's Dilemma
and the Matthew and Luke's Game. The former was defined by
A.W. Tucker, the latter by R.B. Braithwaite (compare R.D. Luce
and H. Raiffa, Games and Decisions, John Wiley and Sons, Inc.,
New York 1958, chapters 5 and 6).
Player B
b1 b2
a1 0.9 0.9 0 1
Player A
a2 1 0 0.1 0 .1
Player B
b1 b2
a1 1 2 7 3
Player A
a2 4 10 2 1
148
In every box the first number stands for the payoff of Player A, the
second for the payoff of Player B. It is characteristic in both these
games, in fact this is what makes them interesting, that neither
player finds it advantageous to use one strategy all the time. In
Game 1, Player A initially wants to play a2 to get the payoff of 1.
But as he chooses a2, Player B, who acts independently, will play
b2, hoping for the payoff of 1 for himself, and both will get the
lowest payoff of 0.1 for each {a2, b2}. Now, if both shift to the first
strategy they will get 0.9 each, and may be satisfied for some time.
But sooner or later one of them will be tempted to play his second
strategy again to get his top prize. He may quite easily do so the
first time. He will win one point and his partner will lose badly,
viz. get zero. Naturally the partner will go back to his second
strategy now, and enforce the payoff of 0.1 on both players. He will
do so to punish his rival and in order to improve his own score.
More than that, he may feel or play being offended and stick to his
second strategy to hoard a number of full points whenever his
partner makes and overture at cooperation. If he plays this game of
intransigence too long, his opponent, i.e. the player who was first
tempted to play against his partner, may feel he has been punished
too severely and will neglect all further offers of cooperation. Now
both will find themselves in a stalemate of envy and non-<o-
operation. In this game, as in life, being greedy is less of a vice than
feeling jealous and offended.
Game 1 has an option which is advantageous to both players,
and equally advantageous to both of them: {aI, b1}. If they are
interested more in winning than in winning over the partner, they
can play {aI, b1} and collect a safe 0.9 in every round. But this
strategy works only if both of them are greedy rather than
149
in the end his score will either be equal or higher than the score of
A. (Compare Douglas R. Hofstadter, Mathematical Themes in
Scientific American, June 1983.)
'Tit for tat' is a good strategy for players who are more
interested in winning against their partners than in maximizing
their own score, who do not care that their payoffs are low as long
as their partner's payoffs are not higher. They are guided by the
relative and not by the absolute results of the game. For anyone
who has different goals in the game, i.e. is more greedy than
envious, the 'tit for tat' strategy serves no purpose.
This strategy has also another inconvenience. Its 'planning and
reaction span' is very short, it focusses on the partner's last move
and chooses only the next to come. It cannot be matched or
transformed into either mixed or random strategy by either player.
It stabilises relative results of the game but makes absolute results
unpredictable.
Suppose that A and B play the Prisoner's Dilemma, and both
use the 'tit for tat' strategy. If they happen to open with {a2, b2}
each will see the other's move as aggressive, or defecting, and thus
repeat his strategy 2 in every next move. If, on the other hand,
they happen to open with {aI, b1}, each will see the other's move
as cooperative, and both will continue with cooperative moves, i.e.
strategy 1. If they open with opposite moves, {aI, b2} or {a2, b1},
they will proceed 'out of sync' in every next move, and one will
always try to cooperate when the other defects - alternating in
these roles after every round.
In a non-symmetrical game, like the one played by Matthew
and Luke, a 'move by move' response is not the best either.
Suppose the players open with {aI, b1}. That gives them a low
152
score. If they both back out and play their second strategy, which
would be good assuming that the partner does not switch at the
same time, they will have {a2, b2}, and the result is as bad as the
opening move. Without a longer 'reaction and planning span' the
players will alternate between these two joint reactions
indefinitely, for neither is allowed to vary his response to the
partner's last move. They will be locked in the 3 points result
forever and will never adopt a pattern of coordinated responses.
This shows that a longer reaction and planning span, reaching
beyond the partner's last move and one's next is called for. Justice
may be an elusive objective in such games, but the players may at
least settle for coordination, i.e. a pattern of serial responses that
are either pure, i.e. consist of a number of identical moves, or
mixed, when the two simple strategies are used in a particular
proportion, or at random.
min strategy for its own sake. It makes sense to use it only if a
player has a reason to believe that his partner intends to ruin him
rather than cooperate with him. Thus, if at all, a mixed strategy
could be used to indicate that a player is ready to cooperate. He
may begin, for instance, from a proportion of 10% of al to 90% of
a2. What does it mean to B1 First of all, it will take some time
before B knows that A is playing this mixed strategy, perhaps 50
moves. Then he will have to guess what A is telling him. He may
assume, in this case rightly, that A is ready to use one move of
strategy al for nine moves of strategy b2. If the game they are
playing is the Prisoner's Dilemma this means that A proposes to
play for his highest score 9 times out of 10 and is ready to assist B
in getting his best score at every tenth move. If open negotiation or
exchange of messages is excluded, B has no way of telling A that he
has received the message before he begins to respond to it. It will
also take some time before A can assume that B has got this
message and is making an offer of his own. But this time will
finally come about, and A will find out that B plays, say, a mixed
strategy of 60% of bl to 40% of b2. They will have to begin to
bargain now, i.e. they will exchange messages about different
proportions between each one's pure strategies until eventually
they agree on the same proportion between their pure strategies.
At that moment they can begin looking for a pattern of
coordination, so that the strategies constituting the same fraction
of the whole could be used simultaneously. We must admit that it
is a cumbersome, ambiguous and costly method of bargaining. No
player ever knows whether his partner has received his message,
how he understood it, when he has started to respond to it, if he is
ready to read a new offer, how long it will take him to respond to it
154
that he may use, i.e. as he plays bl 10%, 20%, 30%, etc., of times,
his returns monotonically rise or fall. Thus he has no incentive to
use a mixed strategy, i.e. stop short of the 100% value. He had
rather playa pure strategy, the one which promises his highest
payoff in the matrix. In the example just given, by playing bi all
the time, and assuming that A does not change his strategy, B's
payoff will amount to U[B]=6 in the long run. If he played any
other strategy his payoff would be lower.
Why should anyone playa random strategy at all? One reason
is that a player may not care about the game or may not know how
to play it, or even that he is playing it at all. Then his moves may
very well be chaotic or random. Another, but a very dubious,
reason is that one player may want to induce his partner to adopt
the maximum payoff strategy in order to punish him for that. In
the Matthew and Luke Game B might play randomly to lure A
into adopting the ai strategy and then respond by bI, which will
give 1 point to A and 2 points to B in every round. But B has a
better strategy than that. He can simply play bi and then either
get the same result, i.e. punish them both but his partner more
than himself, or he can get his highest payoff. Thus he need never
resign from bl in order to adopt a random strategy. A does not
have any such option in this game, because by inducing B to play
for the maximum payoff, A will make him play bI, which, as we
have just seen, punishes A more severely than B, or gives B high
returns. A comparable situation cannot even arise in a symmetrical
game like the Prisoner's Dilemma. One can neither hope to win
more than the partner nor to maximise one's own payoffs by
playing randomly. The only viable objective in a symmetrical game
is to maximise profits symmetrically, and that can be obtained by
157
A Coordinated Game
his pattern first. The two middle moves may either be {aI, bI},
{aI, bI} or {a2, b2}, {a2, b2}. The first case has just been
analysed. If B switches first the results will be slightly different.
Initially they will have: U[A]=4.8 and U[B]=5.4; and A's returns
will amount to 0.88 of B's. After dividing the two middle moves
between themselves, the results will be the same as before, i.e.
U[A]=5.5, U[B]=6.5 and the ratio will be 0.84. Thus depending on
who made the first courtesy switch they will either see the relative
difference between the returns grow or fall. If A happens to be
interested both in increasing the absolute value of his returns and
in closing the gap between them, B would do well to offer a
courtesy switch after the first stretch of the game. Then meeting
half way, by making an equal number of concessions, would look
doubly attractive to A. Conversely, A would be smart to make the
switch first, if he wants to bargain for a different division of the
middle moves, say, one for him and the other for nobody.
If a conciliatory gesture by one player to move one step in the
direction of the other and to extend the succession of moves which
give the partner his maximum payoff does not evoke a kind
response, the player who has made the offer may take it back in the
next round of ten moves. If this in return produces a retaliation by
the less conciliatory player, the more conciliatory one should
answer in kind. These hostilities may easily break down the pattern
of the game and both players will get their lowest payoffs. This
. cannot be helped. It is hard to induce an uncooperative man to
engage in cooperative behaviour. Some players will rather seek the
ruin of themselves and their partner than acquiesce in results that
offend their sense of fairness, justice of simply their expectation of
success. There is no strategy in the game that an opponent can use
161
A Failure to Coordinate
Julie Jack
Introductory Remarks
167
P. Geach (ed.), Logic and Ethics, 167-197.
© 1991 Kluwer Academic Publishers . .
168
saying, viz that one is to say that p only if it is the case that p.
The existence of such a ru1e does not consist in the fact that
people's behaviour is invariably in obedience to it, nor even in
the fact that people cannot be justified in breaking it (for they
can be)j it consists simply in the fact that one cannot be aware of
something as a saying, without being aware of the rule; anyone
who identifies an act as a saying attaches this ru1e to the act as a
relevant standard of its assessment.
It is in virtue of the fact that people's behaviour shows that
they associate such a rule with an utterance that we can attribute
to them the judgement that the utterance in question is a saying.
It is possible for the practice annexed to a particular utterance to
consist of no further activities than the utterance itself.
Generally, however, the practice annexed to an utterance will
involve further activities, activities of producing other utterances,
as well as non-verbal accompaniments. It is in such further
activities that agreement and disagreement in the specific
character of different participators' judgements regarding a
particu1ar utterances can be expected to emerge. The practice
annexed to a particular utterance X might, hence, include such
further activities as the following: the addressee of X asks (in
connection with X) why its being the case that p is relevant to
the situation, or reminds the speaker of counter~vidence to the
truth of p, or denies that p, or asks why the speaker believes that
p, or how he is qualified to judge that pj or the addressee asks for
further evidence of the truth of p, or asks the speaker to justify
having said to p in view of the falsity of p or in view of the fact
that the speaker did not know or believe that p. Of course, the
addressee may simply pursue the conversation in a way which is
181
equipped observer.
Reductionist aims having been foresworn, objections about
the learnability of a first language will of course arise. It might be
thought particularly objectionable to claim that a child must.
have aquiredd the notion of saying by the time he or she can
participate in language use. This thought, however, does not bear
inspection; the objectionability must attach to certain
unacceptably recondite analyses of 'saying', rather than to the
claim itself; for full participation in language use is intentional,
and what could count as participation if not saying things and
attending to sayings? The general problem of concept-acquisition
in the course of learning a first language does not need discussing
here. However, there is an aspect of the account of the previous
section which is worth stressing in connection with the topic of
language-learning. This is the distinction, which is crucial to
cognitivism about meaning, between direct observations of
utterance-meaning, based on hearing what is said, and
observations of behaviour from which is learned the meaning
which another attaches to an utterance. Any practice-conception
of meaning should allow for this distinction, because, if the sole
basis (even for those familiar with the language being spoken) on
which to observe the significance of an act of speech were to
observe the speaker's other behaviour, this would not yield a
conception of that other behaviour as constituting a practice;
such behaviour, ex hypothesi, would not be answerable to what
holds true of an utterance independently of the speaker's current
attitude towards the utterance. Granted, more needs to be said
about how a 'seems/is' distinction can be drawn for direct
perception of utterance significance; but helping ourselves to the
185
hearing.
The question of whose dispositions are relevant now arises,
and three positions seem possible. The first is that only the
dispositions of the judger of significance are relevant. The second
is that the dispositions that are relevant are those inhering in
members of some community, to which the judger must belong.
The third is that the dispositions that are relevant are
dispositions related in some way to the utterance to which
meaning is being ascribed. With respect to the first position:
Surely it cannot be a general truth about the nature of
judgements of significance that the truth of each judgement is
relative to the judger. If this were so, two people could not make
judgements of significance which had the same true--<)r-false
evaluable content; so not even speaker and hearer could endow
their judgements of significance with the same complete
propositional content. This is the sort of view one adopts only if
forced to, by sceptical arguments. But I know of no good
arguments to this conclusion. The second position, which
espouses a conception of judgements of significance whereby they
have an 'us'-relative content - relating, through the judger of
significance, to dispositions which are authoritative within some
community that he belongs to would allow for shared judgements
of significance. But it is far from clear that such a conception
would allow for that degree of objectivity in judgements of
significance which our view of language actually possesses. The
reason is that in this picture, only co-members of a given
community can make specific meaning-judgements in common,
since every judgement one makes about utterance-meaning, if it
has a complete truth-€valuable content, is, by this hypothesis,
191
in use - i.e. what the speaker acted with the intention of doing
(the aim of his most basic speech-act intention). If the addressee
gets the speaker's meaning right, and the speaker's meaning is
different from the significance as judged by the authoritative
standards of the language in use (because, e.g., of a slip of the
tongue, or a misuse of vocabulary where the speaker must
acknowledge a disposition more authoritative than his own), an
'off-beat' concordance is reached which is nonetheless
satisfactory for the immediate purposes of discourse. Self-
correction on the speaker's part, or the correction from the
addressee, or later discourse, shows that it is off-beat nonethe-
less. The importance of speaker/addressee concordance for the
immediate purpose of discourse, and the fact that an addressee
must coordinate his judgement with the speaker's if concordance
is to be reached, may dictate that 'speaker's meaning' in the
sense outlined above is a secondary standard of the 'real'
significance of an utterance. However, speaker's meaning in the
above sense incorporates a concept of 'significance' according to
which significance must be judged by an epistemically public
standard. Moreover, judgement (perhaps mistaken) of utterance
meaning as determined by the latter standard is ordinarily, for an
addressee, epistemically prior to any judgement on his part of
speaker's meaning. It must be so, whenever language use is
instrumental in the sense of providing the first step of an
inference.
To the extent that either doctrine is acceptable, what is true
in subjectivism about significance is closely connected with local
non-realism about significance: Subjectivism and non-realism in
their plausible versions centre on the same point, that there is
196
Notes
Stephan Korner
199
P. Geach (ed.), Logic and Ethics, 199-224.
© 1991 Kluwer Academic Publishers.
200
1. S + [S + p] -i [S + p] -i Pi S + [S - p] -i [S - p] -i "'Pi
S + [S ± p] -i [S ± p] -i Up
2. S ± [S + p] -i [S + p] -i Pi S ± [S - p] -i [S - p] -i "'Pi
S± [S± p] -i [S± p] -i Up
3. S - [S + p] -i [S - p] -i "'Pi S - [S - p] -i [S + p] -i Pi
S- [S± p] -i [S± p] -i Up
The step-by-step realization of third and higher level practical
attitudes proceeds in the same manner.
We are now ready to extend our definitions of "conjunction",
"negation", "practical inconsistency" and "practical implication"
to practical attitudes of any level. The definition of a conjunction
of attitudes remains the same, since we may assume that a
person may have practical attitudes of the same or of different
level. As regards the negation of practical attitudes, we must
distinguish between the case where a practical attitude and its
negation are of the same level and the case where they are of
different levels. In the former case our definition is the same as
the definition given for two practical attitudes of first level. That
is to say two practical attitudes of the same level are negations of
each other if, and only if, one is a practical pro-attitude and the
other a practical anti-attitude towards the same practicability or
if each of them is a practical attitude of indifference towards the
same practicability. I call this kind of negation "opposition".
If two practical attitudes are of different level, then they are
negations of each other if, and only if, one of them, say a,
dominates the other, say b, if a is a practical anti-attitude and b
is not a practical attitude of indifference. Thus a person's anti-
attitude towards his pro-attitude to smoking and his proattitude
to smoking are negations of each other. So are a person's
209
(So + PI & P2) & (SO + PI) & (SO + P2)' In other words we
must - and can easily - distinguish between practically or
evaluatively inseparable conjunctions of practicabilities on the
one hand and practically or evaluatively separable ones on the
other. The difference might be symbolized by distinguishing
between So + PI & P2 and So + (PI & P2)' Similar remarks
apply to disjunctions. Thus a person may have a pro-attitude
towards paying his debt by cheque or paying it by cash, so long
as it is being paid. This disjunction, which may be symbolized by
So + (PlfP2) differs from a disjunction So + (PI v P2) in the
case in which So has a pro-attitude towards PI and in which P2
is irrelevant. 10
The preceding remarks on the general strategy of
constructing a formal system of practical evaluation cannot, of
course, replace its actual construction. They are meant to show
that - although the informal analysis of evaluative or practical
inconsistency, consistency and implication is defensible by itself -
the construction of an adequate formal system is no less a
practicability than the construction of a logical system under-
lying theoretical judgement or of a number of such systems which
have the same core but differ in their peripheries. 11
Notes
1 The essay develops some ideas expressed in Experience
and Conduct (Cambridge 1976), henceforth referred to as E & C,
especially on the relation between the stratification and
universalization of practical attitudes, on the relation between
the ascertainment of practicabilities and their evaluation and on
224
Jja Lazari-Pawlowska
225
P. Geach (ed.). Logic and Ethics. 225-240.
© 1991 Kluwer Academic Publishers.
226
statements.
According to Albert value judgements and norms should be
understood as "hypotheses", i.e. as provisional assumptions,
which we are ready to give up in the light of new data or
reflection. It is to be stressed that this postulate covers all value
judgements and norms - including the basic ones. 10
A more moderate view is represented by Christian Bay. He
requires that all moral opinions be tentative with the sole
exception of the postulate of humanism ("militant humanism"),
which is subject to analysis and interpretation, but cannot be
challenged. The postulate of humanism acknowledges the
intrinsic value of every individual and his right to free develop-
ment. The author sees all other directives as means to that aim
- means which should constantly be subject to analysis and
evaluation, adapted, reformed, or refuted, according to the
changing situation or improved knowledge. "Opinions and
attitudes must be, if we are rational, tentative only. But, and
this is where I differ profoundly with most pluralist liberals, there
has to be one basic, a priori commitment, subject of course to
analysis and interpretations but not to challenge; and that
commitment, in my view, can only be the value of each human
life itself, freely evolving." 11
The conception of Paul Kurtz has much in common with
Albert's radical stand. Kurtz calls for "critical and revisionary
enquiry" with regard to all moral postulates. Moral ideals and
principles, he says, should be taken as hypotheses whose useful-
ness is confirmed or questioned by experience and scientific
advance. By taking such an open attitude we continue to be
receptive to alternative forms of morality which may deserve
238
acceptance. 12
In the deductive model of ethics, says Albert, the process of
substantiation ends at a definite point - at the norm
acknowledged as a basic one; however, the point where
substantiation stops is chosen arbitrarily. According to Albert,
the question: Why stop here? remains sensible and legitimate. He
who proclaims a statement as the "last" assumption reveals that
he is not ready to subject his basic moral ideas to discussion; he
has dogmatized them. The decision to break off the chain of
substantiations is harmful for psychological, and also for moral,
reasons. For people are prone to take the final element in a chain
of substantiations as something certain, something which does
not require any further justification - in fact they take it as a
dogma. The process of substantiation in the deductive model,
says Albert, ends therefore at a dogma: de principiis non est
disputandum. The impression that there exist some "final"
assumptions in ethics is the result of fascination with the
axiomatic method which aims at establishing exactly such
assumptions.
Albert is therefore against the axiomatization of ethics
because it favours dogmatic attitudes and is incompatible with
the broadly conceived postulate of criticism. According to Albert
dogmatization expresses the tendency to give the need of
intellectual certainty priority over the need to achieve optimal
solutions of actual, important problems. A deductive system in
ethics, even if it were possible to realize it, would take the form
of a closed system; this prompts the author's objections, because
a closed system cannot be improved upon. The open character of
an ethical system - says Albert - is itself a matter of morality.
239
Notes
1 R.B. Brandt, Ethical Theory, Prentice Hall 1959, p. 295.
2 H. Reichenbach, The Rise of Scientific Philosophy,
Berkeley 1951, p. 279.
3 J . Ladd, The Structure of a Moral Code, Cambridge 1957,
p. 152.
4 A. Schweitzer, My Life and Thought, London 1966, p. 13l.
5 Cz. Znamierowski, Oceny i Normy, Warszawa 1957,
p.51!.
6 A. Ross, On Law and Justice, London 1958. p. 297 ff.
7 W.D. Ross, The Foundation of Ethics, London 1939, p. 168
ff.
8 T. Nordenstam, Sudanese Ethics, Uppsala 1968, p. 27 ff.
9 H.D. Aiken, Moral Reasoning, "Ethics" 1953, p. 27 ff.
10 H. Albert, Traktat iiber kritische Vernunft, Tiibingen 1968.
11 Ch. Bay, "Human Development and Political
Orientations", BuUetin of Peace Proposals, 1970, No.2, p. 179.
12 P. Kurtz, "The Uses and Abuses of Science", The
Humanist, 1972, No.5. p. 7.
12
Marian Przelecki
241
P. Geach (ed.). Logic and Ethics. 241-253.
© 1991 Kluwer Academic Publishers . .
242
Note
Jan Srzednicki
255
P. Geach (ed.) , Logic and Ethics, 255-274.
© 1991 Kluwer Academic Publishers.
256
cost is surely too high". But on what basis is the cost calculated?
The participants in norm-vendettas do not think it too high. An
Australian academic barracking for one side in one of them once
said to me: "It does not matter whether people are murdered,
what matters is for what cause they are murdered!" Since he said
it I cannot now write: "Clearly no one would seriously accept
such a statement." This inability is both annoying and
frustrating, so we are back to square one. But one wishes
fervently for an objective norm or norms that would put a stop to
all that nonsense.
One does not need to be an advocatus diaboli to see that
there is genuine plausibility in the subjectivistic position, and
that it has some depth.
Let us make some preliminary assumptions:
Assumption 1 At least in some areas there are objective facts
- this assumption ignores metaphysical idealism, but this can be
accepted since our main task is to contrast and compare the
realm of facts with the realm of norms.
cases the particular insight must take preference over the rule
- if it is our only check on that rule's limits of application.
There are of course constructive rules that are meant to be
absolute - the ten commandments are a prime example. But
should I respect my father and mother to the extent of helping
them to commit major crimes? It could be retorted that the ten
rules together delineate a system that prevents such an effect
- but then none of them is a strict action-<iirective, and if they
are absolute they can at best be absolute rules of thumb. Each
rule has limits to its application, but if we are strict about the
system these must be derivable from one or more of the other
nine commandments, and the need for such derivation, let us
suggest, will always be indicated by the nature of a particular
case, and our intuitions regarding it. Since the case where general
norms are not limited by particular insights is favourable to our
present case, we will assume the opposite without further ado.
It is perhaps easier to reach non-collusive agreement on a
general norm, because it is less intimately bound up with our
ulterior motivation, but also because it is expressed mainly in
terms of the decisive insight categories. We are not then
distracted by fog or lesser and multifarious reactions. This
drowning of main considerations in trivia is not a normative
problem in itself, as anyone with any experience in committee
work knows only too well. But the strength of the general norm is
also its weakness - it tends to be indirect vi~a-vis any
particular action-relevant situation ergo the fit is loose. Mostly
of course this does not matter, else the rule would be useless, but
sometimes it does, and sometimes this is painfully obvious. In
that organized technology of norms, the law, injustices and
267
miscarriages of justice are the hallmark of these cases, for the law
cannot easily appeal to direct normative insight. Perhaps we can
offer a picture reconciling these tensions via the moral monster
technique. A moral monster case, such as constructed above, is a
case where the general normative rule fits a case precisely - there
is no slack, if the author of the case can prevent it. There is no
problem of interpretation and the insight comes into full force
- underscoring the objectivity of the result. The case is also
likely to be so odd and unusual that motivational trivia remain
inactive, they do not interfere with our reception of the norm as
objective.
So we have made rather more than just a prima facie case for
objectivism: firstly - it would be very difficult indeed to mount
any purely logical case against it, and this is strengthened by the
application of the principle of uniformity of explanation - on the
objectivist model the same type of explanation fits both
constative and normative discourse.
Secondly - We made out a case for the objectivity of
significant general assessments, such as that it is objectively
wrong to maximize suffering for the sake of maximizing suffering,
and similar ones. The case for this is in accord with our best
intuitions and with an overwhelming majority of our practices.
The case we made out justifies the special weight we give to
serious, normative assessment, and supports the idea of moral
standards.
It is perhaps important to note that beside the question
whether one's: reaction; assessment; opinion; decision; etc., is
objective or subjective, there is also the question whether in this
or that case, or type of case, it is reasonable to be subjective of
268
Kingfoil and Bach interfering with the purity of the samples. But
other cases are possible, e.g. the transition from acceptance to
abhorrence of slavery, etc. The collector's value is always:
market; box-office; or social acceptance value. We could read
some of the examples of putative objective value as pure
collector's value, where this value is all that there is to it. We
could read other cases as involving intrinsic (objective) values
which are either recognized or not only as collector's values. If so,
some might be fully recognized, some underrated, some over-
rated, some unrecognized. A subjectivist would have to say that
all value is but collector's value, and his story is plausible - this
kind of value is almost always independent of individual tastes, it
almost always has a structure all of its own, almost always it has
significance for those individuals who do not share the taste. The
kind of feature that the objectivist might wish to claim in
support of his claim, is so often, and possibly always, paralleled
by well established and entrenched collector's value, that this can
be effectively claimed as a counter-argument by the subjectivist.
But the subjectivist must admit that on his interpretation no
value can be unrecognized, underrated, overrated, or perceived
accurately. This admission is counter-intuitive, and it
impoverishes normative discourse in a drastic fashion. The only
complexity left is one whereby an individual, or individuals is
(are) in or out of step with other individual(s). Quite clearly, in
those terms it it possible to produce a fair facsimile of: correctly
judging; missing; underrating; or overrating the value of some-
thing - provided only that we can have the basic concept of
recognizing a value, and therein lies a difficulty. We can put this
difficulty fairly succinctly by saying that in order to create via
271
Notes
1 Using 'fact' in the widest possible sense.
2 Please note here the bias of the language - it is actually
difficult to state the present points in a neutral fashion.
3 Extreme animal-rights fanatics may replace in this
example the chasing of a piglet with the digging up of a large
yam - there is no reason to also substitute the piglet for the
daughter.
4 I was taught at school that it cannot, on the quite
reasonable ground that at that speed the vortexes created would
create an infinite force (but sonic boom intervened to retrieve the
situation).
5 Any theory claiming its norms to be absolute would then
fall foul of these arguments, and nothing making a softer claim
- the criterion is thus not theory-specific. Should the failures of
a type of norm be relatively very numerous, the norm would be in
jeopardy - this however is a different kind of criticism.
14
Klemens Szaniawski
275
P. Geach (ed.). Logic and Ethics. 275-288.
© 1991 Kluwer Academic Publishers . .
276
References
Boguslaw Wolniewicz
Non-Naturalism
289
P. Geach (ed.), Logic and Ethics, 289-302.
© 1991 Kluwer Academic Publishers.
290
race.
Thus there are as many philosophical anthropologies as there
are different views of human nature. For the purpose at hand let
us divide them into two large groups: the naturalistic and the
non-naturalistic ones. The former are characterized simply by
the tenet that 'man is just part of nature'. There are many
adherents of this tenet in contemporary philosophy, and they
belong to at least three distinct trends: to the positivist, to the
materialist, and to the Freudian one, the latter being a hybrid of
the two former. A fair sample of the naturalistic bend of mind are
the following statements by Ernest Nagel (Logic without
Metaphysics, 1956, pp. 9, 11, and 27):
A utonomous Value
Doubts on Naturalism
Indeed, if the only values are utilitarian ones, then they are all
relative, depending on the needs and desires which determine
them. However, these needs and desires are given only in so far as
the biological constitution of the human organism is given too. In
axiology naturalism is a plausible position only as long as it
always can fall back upon that constitution as a kind of a
biological absolute which gives its values the requisite stability.
But things are changing now, and we are being confronted with
the technical feasibility of reconstructing human organisms at
will, and thus also their needs and desires.
One of the first to have fully realized the axiological
implications of that momentous change was the Polish writer and
philosopher Stanislaw Lem. In 1969 he wrote in an enunciation
entitled Philosophy on a Floe:
We see that Lem and Milosz are quite close to each other in their
axiological conclusions. This closeness is remarkable, as otherwise
they are people of very different philosophical persuasions: Lem
sticks to a rugged evolutionary physicalism, whereas Milosz's
position might be described as a kind of Manichean existen-
tialism. There should be thus some good reason for the
convergence of their views on that particular point.
Lem's diagnosis was soon corroborated by events, and in fact
it has been shown by them to have been rather conservative. Let
us cite just one telling example of that: in October 1983 we could
again read in our papers about the sinister Dr. Robert White of
Cleveland, Ohio, who is to have announced then in Rome that
'modern medicine makes it surgically feasible to transplant the
head of one man to the body of another'. Without commenting
on this ominous news let us merely observe that it is certainly a
297
Counte1'- Values
constant need for animals, this is just short for saying that being
constantly supplied with free oxygen is such a need; and the
phrase 'being constantly supplied with free oxygen' clearly
denotes a state of affairs, not a thing. We do not know whether a
transformation like that is feasible in all cases where one does
speak of 'needs'; we merely assume it is. Consequently, our
universe of discourse will be the totality of possible - though not
always mutually compossible - states of affairs. Let us mark it by
SA.
Secondly, we assume with Elzenberg that every desire creates
some needs. We assume even more, i.e. that whatever is desired,
is thereby needed too, though not conversely, of course. Thus any
state of affairs which gratifies some desire is such that at the
same time it satisfies the corresponding need of the subject in
question. In view of this the set D of possible states of affairs
gratifying some desire, and the set N of those satisfying some
need, are both subsets of the universe SA, with D being a subset
of N. Observe that the subjects of needs and desires, whatever
they are, are not members of SA. They form a universe of their
own, S, and the needs and desires of the members of S may then
be regarded as two different projections of S into SA. I.e., we
have d: S -+ SA, n: S-+ SA, and D = d(S), N = n(S).
A third universe which is to be taken into account in our
context is the realm AV of autonomous values, and in view of our
first assumption it will have an image in the universe SA too.
But here an important reservation is called for, namely that the
realm AV splits clearly into two: into the set of positive values
(i.e. values proper), and that of negative values (i.e. counter-
values). If a state of affairs comes into existence which
299
Diagram
303
304
Volumes J~ previously published under the Series Title: Melbourne International Philosophy Series.
Nijhoff International Philosophy Series