Robert Cohen wrote:

Hello and thank you for this article which I am reading now. (rest of
article below my insert) I particularly like Kant's logic and hardly
focus upon his ethics, this since I read the view below and other like
it. But I have crossposted this into alt.kant and some there may know
what Kant might really think about affirmative action if alive.

Philosophical Problems and Arguments: An Introduction
by James W. Cornman, Keith Lehrer, George Sotiros Pappas
http://www.amazon.com/exec/obidos/tg/detail/-/0872201244/
http://hume.ucdavis.edu/phi102/lecmenu.htm

The First Formulation of
the Categorical Imperative

You are permitted to act on a
principle P only if you can
will P to be a
universal law.

The Second Formulation of
the Categorical Imperative

Act so that you treat humanity,
whether in your own person or
in that of another, always
as an end and never
as a means only.

S) A Proposal for a Satisfactory Standard:
A Utilitarian Kantian Principle

Our job is to find some way to graft the utilitarian principle onto
Kant's second formulation. We know two things: First, the basic
prescription is, if possible, to treat no one merely as a means, but if
this is not possible in a particular situation, then we should treat as
few people as possible as mere means. The overcrowded lifeboat example
illustrates a situation in which someone must be sacrificed, treated as
a mere means, in order to save the others. In such a situation it is
obvious that as few as possible should be sacrificed. Second, we should
treat as many people as ends as possible. We have interpreted this to
imply that we should actively promote the well-being of those affected
by the action in question. However, because promoting the well-being of
as many people as possible could conflict with treating as few people
as possible as mere means, and because the most basic imperative is not
to treat people as mere means, the second imperative must be restricted
so that it is consistent with the first.

At this point an objection can be raised. We can avoid treating a
person as mere means by doing nothing at all. Consequently, in any
situation we can avoid treating anyone as a mere means. If we accept
the preceding imperative as basic, we should sacrifice no one in the
lifeboat example, because that would be to treat as mere means as few
people as possible. But that would result in a needless loss of life.
We must, then, find a different basic principle.

We can avoid the objection by construing the treatment of someone as
mere means to include doing nothing to help him when he truly needs
help, especially when his life is imperiled. Not to do anything to help
someone in such a situation is to respond to him as something with no
intrinsic worth. This amounts to treating him a mere means. We can,
then, take the basic imperative to be

In any situation, (a) treat as mere means as few people as possible,
and (b) treat as ends as many people as is consistent with (a).

We have claimed that promoting someone's happiness is important for
treating him as an end. We should, then, incorporate into our
imperative a prescription to promote the happiness of those affected by
an action. However, because promoting as much happiness as possible
often conflicts with the previously stated basic imperative, any
prescription to promote happiness must be restricted, so that following
it is consistent with what our basic Kantian imperative prescribes.

Although this gives the essential skeleton of the principle, there is
still the question how we are to relate the treatment of as many people
as ends as possible to the promotion of happiness. The problem is that
there are several conflicting ways we could do this. We treat one
person as an end by promoting his happiness. We could, then, require
the action that promotes to some degree the happiness of the greatest
number of people, or we could be utilitarian at this point and require
that it maximize the total amount of happiness, counting, of course,
each one as one and no one more than one. Let us initially choose an
act utilitarian interpretation that gives us the following principle:

An action ought to be done in a situation if and only if

1. Doing the action, (a) treats as mere means as few people as possible
in the situation, and (b) treats as ends as many people as is
consistent with (a), and

2. Doing the action in the situation brings about as much overall
happiness as is consistent with (1).

As the reader can discover for himself, this principle seems to meet
all of the first five conditions that any satisfactory ethical theory
must meet, except for (3), which concerns special duties. By applying
the act utilitarian principle to treatment of people as ends, we have
allowed the problem of the special duties of teachers and others to
arise again. However, because this problem can be handled by rule
utilitarianism, we can accommodate special duties by applying the rule
utilitarian principle. Here again we have a choice to make. We can
assume, as a rule utilitarian does, that there are utilitarian rules
covering every situation involving a moral choice. Or we can make
provision for the existence of some situations not covered by these
rules by requiring that the act utilitarian principle apply in these
situations. Let us here, however, accept the rule utilitarian's
assumption. What we can call the utilitarian Kantian principle will be

An action ought to be done in a situation if and only if

1. Doing the action, (a) treats as mere means as few people as possible
in the situation, and (b) treats as ends as many people as is
consistent with (a), and

2. Doing the action is prescribed by any utilitarian rule that (a) does
not violate condition (1) in the situation, and (b) is not overridden
by another utilitarian rule that does not violate condition (1) in the
situation.

To help understand this principle, let us see what it would prescribe
in one particular lifeboat example. Let us assume that you are the
captain of a ship that has just sunk, and you are in charge of the one
remaining lifeboat, which has too many people crammed into it and three
others, who are taking their turns in the water, hanging onto the sides
of the boat. Suppose further that a dangerous storm is quickly
approaching, and the boat will capsize unless five people, at minimum,
are cast adrift. You must decide what ought to be done. The utilitarian
Kantian principle requires you to sacrifice some people, but as few as
possible, in the situation in order to save the rest. In this way you
would treat as few as possible as mere means, and as many as possible
as ends in this situation.

Once this decision is made you are faced with the problem of finding a
procedure for deciding who is to be sacrificed. One decision procedure
which clearly treats no one as mere means is to draw straws, but
another one is to ask for volunteers. The basic Kantian requirement
expressed in condition (1) provides no way to choose between the two
procedures. Thus, you must consider any relevant utilitarian rules. To
see which rules apply, let us further assume that five people in the
boat have publicly volunteered to be sacrificed. Consider now the
following rule: Whenever it is required that some people be sacrificed
to save others, and some people have publicly volunteered to be
sacrificed, then there is a prima facie obligation to sacrifice the
volunteers. This rule clearly applies in this situation and it does not
violate what the basic Kantian condition requires. Furthermore, it is
reasonable to think it is a utilitarian rule, because its being in
effect tends to maximize the overall happiness of those to whom it
applies. Indeed, it is quite likely that if this rule were not followed
when it applies, there would be great unhappiness, and strong
resistance, or even mutiny, when those who did not volunteer, but know
others did, are asked to take a chance on being sacrificed. And, given
the additional plausible assumption that this rule is not overridden in
this situation, your obligation is to ask for volunteers, rather than
have the passengers draw straws.

The principle we have finally reached is complex. As can be seen from
the preceding example, it requires of anyone that he consider and
relate many factors in order to decide what he ought to do in any
particular situation. In many situations, it is practically impossible
to complete such a complex task. Each of us should, of course, do the
best he can, and where anyone has done a reasonably good job but failed
to decide correctly, no blame or guilt should attach to him. As brought
out in the beginning of this chapter, the standards appropriate for
morally evaluating actions are different from those appropriate for
morally evaluating persons. Although we have not considered the latter
kind of standard here, one thing is clear: Many actions that are quite
clearly wrong do not reflect blame or guilt upon the doer.

Philosophical Problems and Arguments: An Introduction
by James W. Cornman, Keith Lehrer, George Sotiros Pappas
http://www.amazon.com/exec/obidos/tg/detail/-/0872201244/
http://hume.ucdavis.edu/phi102/lecmenu.htm