JORDAN HOWARD SOBEL
KANT'S MORAL IDEAL ISM
(Received 28 July, 1986)
Moral Idealism identifies morality with the ideal set of moral rules,
where this set of rules is in turn identified in terms of a moral ideal.
Morality, according to a moral idealism, is that set of rules universal
obedience to which would realize this idealism's moral ideal.
Moral idealisms, as here understood, define morality in terms of
universal obedience tests. Theories of morality that instead use uni-
versal acceptance and universal allegiance test are thus not, for present
purposes, moral idealisms. Such theories encounter some of the prob-
lems of obedience-theories, but not all. Most notably, obedience in
contrast with acceptance and allegiance-theories lack resources for
imposing limits on the complexity of moral rules.
There are different ways of understanding universal obedience to a
rule or rules. To mark one distinction, let universal actual obedience to
a rule in a situation be a set of actions each one of which satisfies the
rule in this, the actual, situation given what others are doing in it, and
would do were the agent to do this thing, or this or that alternative to
this thing. Let universal projected conformity to a rule in a situation be
a set of actions each one of which would satisfy the rule in that situation
in which others were doing their actions in this set, given what, in that
projected situation, these others would be doing and would do were the
agent to do this or that alternative to this action. All references to
universal obedience in this paper are to universal obedience in the
'projected' sense. (I discuss and apply this and related distinctions in
'Everyone's conforming to a rule', Philosophical Studies, 1985, and in
'Utilitarianism and cooperation', Dialogue, 1985.)
Moral idealisms differ in their characterizations of the moral ideal.
According to Kant, universal obedience to the rules of morality would
realize a Realm of Ends:
PhilosophicalStudies 52 (1987) 277--287.
�9 1987 by D. Reidel Publishing Company.
278 JORDAN HOWARD SOBEL
a whole of all ends in systematic connection, a whole of rational beings as ends in
themselves as well as of the particular ends which each may set for himself. (Founda-
tions of the Metaphysics of Morals, tr. L. W. Beck, Chicago, 1949, p. 433. All
quotations unless otherwise indicated are from this text.)
[a union in which] everybody may pursue his happiness in the manner that seems best
to him, provided he does not infringe on other people's freedom to pursue similar ends,
i �9 on another's right to do whatever can coexist with every man's freedom under a
possible universal law. (Immanuel Kant, On the OM Saw: That May Be Right in Theory
But It Won't Work in Practice, tr. E. B. Ashton, Philadelphia, 1974, p. 58.)
The moral ideal for Kant is, I think, that state in which freedom for
each is at a maximum consistent with like freedom for all, and, subject
to the realization of this libertarian end, persons' particular ends are
satisfied and happiness is maximized. His moral ideal seems to have
combined in lexical fashion libertarian and utilitarian elements. In
possible contrast, Kant seems to have endorsed a purely libertarian
political ideal that he viewed as only 'de facto utilitarian', happiness
being an attendant of, but not a part of, the ideal.
A constitution allowing the greatest possible human freedom in accordance with laws by
which the freedom of each is made to be consistent with that of all others -- I do not
speak of the greatest happiness, for this will follow of itself -- is . .. [an] idea, which
must be taken as fundamental not only in first projecting a constitution but in all its
laws. (Critique of Pure Reason, A316/B373.)
Connectedly, while Kant may not have favored paternal, or 'good
Samaritan', civil laws, he endorsed moral principles calling for benefi-
cence, and enjoining that, when one can, one help others.
The details of Kant's conception of morality's end are not, however,
important for present purposes. What is now important is that he
conceived of morality as having an end, and as being defined by its end.
According to Kant, morality, is
that legislation through which alone a realm of ends is possible. (p. 434)
Morality is that body of laws through which a realm of ends would actually be realized
�9 if they were universally obeyed. (p. 438)
That is Kant's moral idealism.
Moral idealisms one and all, and including most notably Kant's, are
defective, logically defective. And moral idealisms, including again most
notably Kant's, can make particularly acute, for their believers, the
question, "Why be moral?". The defect of all moral idealisms is
KANT'S MORAL IDEALISM 279
established and explored in the next section. The difficulty is elabo-
rated, in the section after that.
MORAL IDEALISM'S DEFECT, AND POSSIBLE
REMEDIAL MODIFICATIONS
1. Every moral idealism presupposes that there is a unique set of ideal
moral rules, that set of rules universal obedience to which would give
rise to the state of affairs that is the moral ideal of the idealism. But
there cannot be a unique set of rules that satisfies the universal
obedience test.
Proof: Assume for purposes of indirect argument that M is the one
and only set of rules that satisfies a certain universal obedience test
(that is, a universal obedience test that incorporates some specific
moral ideal). Let M' include every rule in M and also some rules that
prescribe specifically and only for cases in which rules in M are not
obeyed.
If the rules in M' are universally obeyed, then so, of course, are the
rules in M. For M' includes M- M is a subset of M'. And, conversely,
if the rules in M are universally obeyed, then so are the rules of M'.
For rules of M' that are not also rules of M positively apply only when
not all rules in M are universally obeyed. When all rules in M are
universally obeyed, rules in M' that are not also rules of M are satisfied
'by default' or 'vacously' somewhat as is the rule to keep promises in a
world in which no promises are made.
Though M' is different from M, given the defined relation between
M' and M, the rules of M are universally obeyed if and only if the rules
of M" are, and M satisfies a universal obedience test if and only if M'
also satisfies it. So M is not the one and only set of rules that satisfies
the given universal obedience test, and regardless of the exact details of
a given moral ideal, there cannot be a unique set of rules universal
obedience to which would realize it.
2. What can be done? Probably any one of a number of things. The
following suggestions take moral idealism some steps towards what can
be termed Humean Realism or Practicalism.
As a first pass at remedying the defect one might propose that
280 JORDAN HOWARD SOBEL
morality is the intersection or the common part of ideal morafities. But
morality, on this suggestion, will be incomplete. Most notably, it will
contain no rules that would resolve issues in one way when there are
other equally good ways in which they could be resolved -- e.g., no
rules relating to many of the details of property and contract, if as it
seems, there are details here that, though they need to be settled, can be
settled equally well in a number of different ways, Also, morality will
not include rules that say what is to be done is cases of breaches of
moral rules, if, as seems true, there are always several things that might
be prescribed in the event of various breaches. Morality, on the present
proposal, will not include any rules that concern responsibilities of
law-breakers tomake amends, or that concern responsibilities of others
to punish law-breakers. Morality on the present proposal will, it seems,
include few if any rules for cases in which, through earlier transgres-
sions, an agent finds himself under conflicting moral demands and
prohibitions. And perhaps most disturbing, morality on the present
proposal will include few if any rules for situations, possibly awful
situations, that can arise for agents only given the immoral doings, past,
present, or future, of other agents.
Is morality, understood as the intersection of ideal moralities, there-
fore hopelessly incomplete? Must we consider some entirely different
line of remedy for the radical defect of moral idealisms? No. Morality
as currently understood, that is, morality as the intersection of all ideal
moralities, though not including rules that deal directly with many
important matters, can by a small addition be made to address itself to
these matters indirectly, through the actual norms of well-arranged
societies.
Proposal -- Let morality be the intersection of ideal moralities,
together with the rule:
Defer to those norms of your society that are ideal norms.
Let a norm be an 'ideal norm' if it is a member of at least one ideal
morality.
This, while perhaps a move in the fight direction, does not really
solve the problem that provokes it. For consider that the norms of
many societies, while doing good work and being better than no norms,
may even so not be ideal norms for the cases they cover and the
KANT'S MORAL IDEALISM 281
institutions they structure -- they may well not be members of any ideal
moralities universal obedience to which would realize an ideal moral
state. It should suffice here to recall that ideal norms can be of any
order of complexity and can be expected often to elicit the response,
That may be all very well and good in theory, but it would never work
in practice. Even with the admixture of realism just proposed -- that is,
even with this reference from within morality proper to actual positive
moralities of agents' societies, agents will be without moral guidance in
cases in which they most need it, since actual rules, even good and
useful ones, will seldom if ever be ideal rules. Here is a new proposal
that responds to this difficulty. Replace the rule displayed in the
previous paragraph by the following rule:
Defer to the norms of your society in so far as they are
promoting morality's end and are in this sense improve-
ments on the State of Nature.
We are, one may feel, nearly done. It remains only to consider that
moralities of actual societies vary in their coverage, and that at the
extreme there is the State of Nature, a condition of man in which there
are no actual social rules or conventions. Focusing for simplicity on this
extreme, must morality in the State of Nature come down to nothing
more than the common part of all ideal moralities? No. The inspiration
for the course we have taken makes one last Step nearly irresistible.
Final proposal -- Let morality be the intersection of all ideal moralities,
together with the rule stated at the end of the just previous paragraph,
as well as the following rule:
Subject to obedience to every other rule of morality, do
whatever will promote morality's end at least as much as any
other thing you can do.
3. We began this section with the observation that Moral Idealism is
logically defective, and that it must be changed in some way if it is to
be even a candidate for a correct theory of morality. One line of
correction was then followed to its natural culmination. Elements of
realism were introduced in the form of referrals to actual moralities,
idealism was tempered and actual norms allowed to be relevant if they
282 JORDAN HOWARD SOBEL
are good norms, even if they are not best or ideal norms, and the
account was provided with a kind of completeness through a rule that
would have decisions directly serve morality's end subject to conformity
to all other moral rules.
And what is this theory that our remedial action has produced? It is,
we can say, Moral Idealism Humeanized, but not completely so.
Suppose we were to operate in the manner described on Kant's moral
idealism. The result, while closer to, would not be Hume's theory of
moral actions as actions of virtuous persons, just and benevolent
persons. One difference would be in the conception of morality's end.
For Hume it is the pure utilitarian end of maximum happiness. For
Kant it was, I think, first the libertarian end of maximum freedom, and
only secondarily and subordinately the utilitarian end of maximum
happiness. Another difference, a more important one in the present
context, is that the account we have reached retains an element of
idealism not found in Hume's theory. Morality on the present qualified
idealistic account includes the common part of ideal moralities, whereas
morality for Hume includes, at the level of rules only certain actual and
generally observed rules of the agent's society. To dramatize the
contrast I observe that morality on Hume's completely non-idealistic
account includes no rules (for example, no rules of justice relating to
property or promises) in a State of Nature, in which state active moral
virtues are confined to species of benevolence, and moral principles
(as distinct from rules of thumb) reduce to the single principle, the
Supreme Law, that you should do what you can to make people happy.
I am, of course, not saying that Kant was Hume. Kant was Kant. He
was first and last an unremitting moral idealist. Kant's pronouncements
on the subject are, I think, unequivocal. He endorses moral idealism
without qualification, and without complicating additions. 'But what are
you implying here about the logical defect of moral idealisms? Do you
mean to say that Kant didn't notice it?' Yes. And I have no idea what
he would have done about this defect, and related problems, if he had
been confronted with them. It would have depended, I suppose, on who
arranged the confrontation, on how resourceful and relentless he was in
this connection, and on what remedial possibilities suggested them-
selves in the conversation.
KANT'S MORAL IDEALISM 283
MORAL IDEALISM'S PROBLEM WITH MORALITY'S
JUSTIFICATION
1. Moral idealism -- the identification of morality with rules universal
obedience to which would realize some end or ideal -- can seem to
make particularly acute the problem of morality's justification. Suppose
that someone wonders why he should be moral, why he should commit
himself totally and without reservation to morality's rules, and scrupu-
lously satisfy morality's every demand. Suppose, to take what one might
have thought would be the easiest case, that this person is already inter-
ested in morality's end -- that ideal that universal obedience would
realize. Suppose, for example, that he is already excited by the nobility
of the conception of a 'realm of ends' and prepared to do everything he
can to promote its realization. What can be said to this person that will
suffice to interest him in Kantian morality and move him to submit
without reservation to its constraints?
'But what could possibly be this" person's problem? He makes
morality's end his own end, so isn't he bound, if consistent, to be
prepared to make morality's rules his rules, and to do his part in that
practice of universal obedience that by hypothesis would realize that
end?' No. He is not, by his acceptance of morality's end, bound to
accept morality. For consider:
... though he scrupulously follow the rules of morality, [he] cannot for that reason
expect every other rational being to be true to it . . . . Still the law: Act according to the
maxims of a universally legislative member of a merely potential realm of ends, remains
in full force . . . . And just in this lies the paradox that merely the dignity of humanity as
rational nature without any end or advantage to be gained by it, and thus respect for the
mere idea, should serve as the inflexible precept of the will. (pp. 438--9)
A moral idealist who makes morality's end his own end cannot
automatically take his interest in this end as a ground for interest in and
commitment to morality's rules. For obedience by a person to rules,
universal obedience to which would realize some end, can, of course, in
the absence of obedience to these rules by other persons, or by himself
at other times, be "without any end or advantage." In such cases,
scrupulous adherence to morality's rules can seem not only pointless,
but positively counter-productive, and can seem so even, perhaps
284 JORDAN HOWARD SOBEL
especially, to a person who makes morality's end his own end. Such a
person could think, correctly it seems, that in some bad circumstances
morality's end, and morality itself, could be best promoted by immoral
means that would violate some of its rules. For example, a little 'noble
lying' about the nature and grounds of morality to get others to join in
adherence to it, though against morality, that is, though against its rules,
might be just the thing that in some imperfect circumstances would best
serve and promote its objective.
2. I digress to observe that the present problem with moral idealism
does not arise for a certain related, though in spirit very different,
conception of morality. Let Practical Moral idealism identify morality
with rules or prescriptions that satisfy two obedience-conditions: First,
for idealism, universal obedience to these rules is to realize some
specified moral end or ideal; and second, for practicality, individual
obedience to them is (both in the presence, and in the absence of
obedience by others or by the agent at other times) to serve that end or
ideal at least as well as would any other course open to the individual
obeying agent.
Generalized Act Utilitarianism (see B. C. Postow, 'Generalized Act
Utilitarianism', Analysis, 1977), and, arguably, Cooperative Utilitari-
anism (see Donald H. Regan, Utilitarianism and Co-operation, Oxford
1980), are examples of what could be termed 'utilitarian practical moral
idealism'. These practical moral idealisms depart in one way or another
from exclusively individual-act-oriented conceptions of morality. Ac-
cording to the first, GAU, morality includes prescriptions for actions of
groups of agents, as well as for actions of individual agents. According
to the second, CU, morality includes prescriptions for attitudes and
motives (in particular, it requires that one be ready and willing in a
certain manner to cooperate) in addition to prescriptions for actions
proper. It is demonstrable that every practical moral idealism that
satisfies a certain condition must depart in one way or another from the
'exclusive individual-act-orientation' exemplified by act utilitarianisms.
The condition is this: That the ideal of the practical moral idealism
order interactions in situation in a manner that is sufficiently inde-
pendent of the actions involved in them to allow for the possibility of a
situation of the structure,
KANT 'S MORAL IDEAL ISM 285
C1 *C2
,,
R1 ! 2 0
1
*R2 I, 0 1
where numbers order possible interactions from worst, 0, to best, 2,
according to the moral ideal, asterisks indicate what each agent will do,
and actions are causally independent so that, for example, even if Row
were to do R1, Column would still do R2.
Proof: suppose that there is an exclusively individual-act-oriented
practical moral idealism, API, for which a situation of this kind is
possible. Since API is a practical moral idealism, individual obedience
to its rules in this situation would include Row's doing R2 -- that is the
best Row can do given that Column is doing C2 -- and would include
Column's doing C2, which is the best Column can do given what Row
is doing. Since API is an exclusively individual-act-oriented practical
moral idealism, total obedience to it would include nothing more than
these obedient individual acts, so that there is universal obedience