Kagan I

Shelly Kagan's

The Limits of Morality

Initial Remarks

I.  I first read this book in the spring of 97, largely because Shelly Kagan was one of my teachers in grad school, and I had been hearing so much about his book.

II.  On the first reading, I realized that it was a powerful attack on a view of morality to which I myself subscribe, and that if I was going to continue to hold, much less defend, that view, I was going to have to get very clear on what was going on in this book.

III.  When I got the offer to teach an advanced moral theories course at SFU in the fall of 97, I decided that I should include this book, since it was already reshaping the debate over utilitarianism which has always been a central topic in moral philosophy.

III.  This will be my third reading of the whole book. In fact, the more I read both this and Scheffler’s book, the more I think that Kagan’s book is far better than Scheffler’s book, although it probably could not or at least would not have been written had Scheffler not written his.

IV.  This is not to say that we need to hang on every single word of Kagan’s book : We can skip or skim quite a lot of it, including a long portion in the middle. When the time comes, I'll just summarize for you what he does in the middle part, and have you read a shorter paper that argues for more or less the same thesis that Kagan defends.

V.  Kagan basically says that one can skip most of chapter 2. This, however, is an invitation I'm recommending we not take up. For I suspect that the weakness in Kagan's argument may lie somewhere in this stuff that he invites us to skip. That, anyway, is my hunch.

I. Let’s start out with a translation guide: Kagan is responding in large part to stuff that goes on in Scheffler, but he uses somewhat different (and, I think, better) terms.

II. Rather than Scheffler's agent centered restrictions, we have constraints rather than agent centered prerogatives, we have options.

III. The Moderate is someone who holds both options and constraints, so he is like Scheffler's deontologist, or the proponent of what S calls at one point the Fully agent-centered conception.

A. It is worth remembering that S himself holds, as we remember, the hybrid view.

B. Part of Kagan’s argument will be for the claim that the hybrid view does not do all that S thinks it does.

C. S seems to want to offer the hybrid view as a quasi-moderate view–one that gives us everything that the moderate should want, but without the paradox of the ACR. Kagan will argue that the ACR is not nearly as expendable as S thinks.

IV. The extremist, of course is the consequentialist BUT WITH THE FOLLOWING ADDITION: in Kagan's terminology, the extremist COULD accept constraints. Thus he defines extremism as: the thesis that one must perform whatever action that it not prohibited by constraints, if any, that one can reasonably expect to produce the best SOA. Quote on page 2: it is extreme because it "strikes us as outrageous in its demands. The claim is deeply counterintuitive. But it is true."

I. Kagan is going to defend the extremist position.

II. He's not going to argue directly for E here, but rather he is going to argue against Mod, for Mod is the thing most people believe instead of believing E.

III. If he can show that M cannot be defended, or that it is incoherent, he will have at the very least narrowed the field down to E and Min.

IV. Minimalism is the other possibility: it claims that morality requires either nothing or at least far less than what the Mod says and way in the hell less than what the E says.

A. Egoists, for example, would be min.

V. Kagan claims that mod's position is midway between E and min.

VI. And he's going to try to show that this middle ground is unstable.

A. If the mod is walking a fence between min on one side and E on the other, Kagan's goal is to go get an ax and start chopping away at the fence.

VII. Now, the way E and Mod and Min are related is this:

A. E says that we always have a decisive reason (=we are always morally required) to promote the best SOA so long as doing so does not violate any constraints, if any such there be.

B. Mod says that we sometimes have a decisive reason (=we are sometimes morally required) to promote the best SOA, but very often we do not, and the reason for this is the existence of both constraints and options.

C. Min says that we never have a decisive reason (=we are never morally required) to promote the best SOA

VIII. By the way, we’ll use the term "best" here the same way we did when we discussed Scheffler: The BEST STATE OF AFFAIRS may not be our favorite state of affairs: to be the best state of affairs means that it is best from an impartial point of view.  

A.  Also, remember that saying "best state of affairs" is the same as saying "best consequences."  And, as if that were not enough, the term "promoting the good" is yet another way to say the very same thing.

B.  So in other words, the following phrases all mean the same thing:

    Promote the best state of affairs

    produce the best consequences

    promote (or maximize) the good

C.  Finally, remember that each of these phrases refers to what is good or best from the impartial/impersonal point of view.  

IX. Kagan's argument will work by (p. 6): "forcing the mod to defend his position against the E's demands without collapsing into the arms of the min."  He thinks that the moderate will be unable to do this, and that this will show that the moderate position is in fact incoherent and therefore irrational and therefore mistaken.

I. Some of the beginning moves of the argument we can move past fairly quickly, since we've seen the same moves and counters in Scheffler.

II. Thus on p. 7 we get a version of the question about whether E is self-defeating because by requiring us to always produce the best SOA, it might make us too tired to do so.

A. But of course this objection--a common mistake, by the way, in papers on utilitarianism in intro ethics--simply misses the definition of E:

B. For E says to do whatever it takes to promote the best SOA (and if that means stop doing good to get a decent night's sleep, or maybe even go see a movie or drink a beer now and then to recharge your batteries in such a way to make you better good-promoter the next day, then THAT is what E will require.

III. A similar quick dismissal is available to an argument he gives on pages l5f- Why can't the Mod just say, look, we are not motivated to always promote the good, so we cannot be required to do so.

A.  Now if we construe this argument as resting on the implicit premise that we can only be required to do what we are motivated to do, then we won't get mod, but min, since many people will never be motivated to promote the best SOA, and on this view, they would never be so required. (it's worth pointing out that a more sophisticated version of something like this argument will re-appear later on.)

 

The Pro Tanto Reason to Promote the Good

I.  Having dismissed a couple of oddly placed non-starting arguments, Kagan sets out to sharpen his methodology

II.  First, he wants to elaborate this idea of a PRO TANTO reason to promote the good.

A.  To call a reason "pro tanto" means that it has some weight, though it might be outweighed by other, stronger reasons to do something else.

B.  The picture of reasons he develops here and later is this:

1.  Reasons are like forces, and they can push for different actions

2.  They can have their magnitudes added together.

3.  There will be a "resultant" or balance of reasons, that will be the outcome of all the different reasons and their weights pulling in all the various directions in which they might pull.

III.  Now, according to Kagan, the mod accepts the pro tanto reason to promote the good.

IV.  Why does he think that?

V.  The very short answer is that if she does not, she'll end up being a minimalist rather than a moderate.

VI.  The longer answer is that mods will typically want to claim that there are situations where one must promote the good.

A.  Typically those situations will be ones involving little cost to you and in which one does not have to violate any constraints.

Thus throwing the rope to the drowning child.

So it is not, on Kagan's view, that the Mod doesn't hold that there is a pro tanto reason to promote the good.

VII.  An important feature of the pro tanto reason is that it does not disappear if it gets outweighed.  (This makes it different from Ross's prima facie duties).

A.  There are actually some other, more technical arguments for this claim. Esp pages 49 and thereabouts.  However, here is what I hope will be a simpler :

B.  To make things a bit simpler and, I hope, clearer, suppose that you go to the beach and see splashing coming from two locations in the water.

1.  Your days as a lifeguard tell you that at least one person is drowning at both locales. So you know that there is reason to swim to both places. Yet once you get out your telescope., you see that at location A one person and at B two people are drowning.

2.  Now, you have most reason to go to B,

3.  But that does not mean that the reason you had a second ago to go to A simply disappears.

4.  Indeed, if it did, then there would be no reason to call for backup.

5.  Rather, the best way to make sense of the commonsensical approach to morality that the Mod represents is to say that the fact that doing something would improve the world creates a pro tanto reason, a reason that can be over-ridden, but a real reason nevertheless.

C.  Similarly, the Mod holds that there is A reason to give one's money to the poor. For that is what makes doing so a good thing, even though it is, according to the mod, not required.

VIII.  These sorts of considerations suggest (and I want to lean heavily on that word, suggest) that the mod is committed to the existence of the pro tanto reason to promote the good.

IX.  This means that the moderate and the extremist BOTH agree that there is a pro tanto reason to promote the good.  In this, they both differ with the minimalist, who denies that we have even a pro tanto reason to promote the good.  The difference between the mod and the E is that the mod thinks that in many if not most cases, that pro tanto reason will be outweighed by other moral considerations.

X.  Like what? The most plausible idea is that it is the COST TO THE AGENT that will outweigh the pro tanto reason (p 21fff).

A.  That is, the most plausible basis for defending M is the idea that the pro tanto reason to promote the good will get overridden by the need to take into account the personal point of view.

B.  Or, in Kagan's terminology, the pro tanto reason will be outweighed by the cost to the agent of acting always to promote the good.

1.  What kind of cost to the agent are we talking about here?

2.  The same basic stuff Williams and Scheffler talk about: the freedom to pursue one's projects out of strict proportion to the amount of good they will do from the impersonal point of view.

A Preview of Things to Come

I.  Kagan is going to talk a lot more about the appeal to cost later in the book, but right now he wants to make some preliminary remarks as a sort of preview.

II.  One big problem that Kagan sees on the horizon for the M who wants to defend options on the basis of the cost to the agent (e.g. the cost to the agent's integrity) is that the argument might work TOO WELL.

III.  That is, the appeal to cost might prove TOO MUCH.   In particular, it might prove that there are options of a kind that the Moderate may not like.

IV.  To see how this might be so, we must first remember that --as we saw from Scheffler--you don't get constraints "for free" --that is, if you show there are options, that does not in and of itself show that there are constraints.

V.  Now, if the FORM of the argument from cost to options is going to go like this:

A.  It is unreasonable for morality to impose great costs on agents (see how much this is like Williams's charge that utilitarianism is "unreasonable")

B.  Not recognizing options imposes a great cost on agents

C.  Therefore it is unreasonable for morality to fail to give us options.

VI.  The problem is that if this is the argument that the M wants to use, then it also seems to support the conclusion that we have options to HARM others.

A.  For if it is costly to benefit others as much as possible--e.g. by giving all your time and money to charity--then surely it is also going to be costly to refrain from actively harming others sometimes.

B.  Thus if you have a million dollars, E will tell you to give it up to charity. IF you say, no no no, that is too costly, for it deprives me of my fortune.

C.  But if that is a good answer to E, then why does that same appeal not apply to the person who can bump off uncle Milty and inherit a million dollars?

D.  Either way, you are out a million bucks you could otherwise have had.

VII.  So one challenge for the M who wants to use the appeal to cost is going to be to show that it will not also give us options to DO harm.

VIII.  Rather, moderates typically think that we have options to refrain from helping, but not options to harm. Notice that Scheffler's formulation of the ACP seems to fall prey to Kagan's worry here, for it seems to allow us to harm others at least a little bit (i.e. within the limits of whatever the weighting factor is for our projects).

IX.  So it would appear, says Kagan, that if the moderate is going to defend the ordinary morality that most of us know and love, she will not be able to stop--like Scheffler does--at options, for those options will also, it seems, be options to harm, and most moderates do no think that we have such options.

X.  Thus for that reason, if we are to justify the kind of morality that most people believe in--the kind that allows us the option of NOT BENEFITING, but not the option of ACTIVELY HARMING, then we will also need to argue for constraints as well, for this seems like the only way that we are going to be able to have options but not options to harm, i.e. by having constraints that rule out harming.

Problems with Constraints

I.  For reasons that are much the same as we saw in Scheffler, Kagan points out that constraints are going to be extremely difficult to justify.  

II.  Of course Kagan has read Scheffler, and essentially rephrases lots of the arguments we dealt with earlier.

III.  Kagan gives his version of the moral that Scheffler draws, on page 29: "constraints cannot be erected on considerations of badness."

IV.  This is because constraints have the paradoxical feature that we discussed earlier:  it seems irrational to forbid the performance of one horrible action in cases in which doing so would prevent several equally horrible actions.  

V.  What Kagan says, however, is that this should worry the moderate.  

VI.  Kagan argues that a theory that has options but not constraints will have no way to rule out options to harm.  

A.  Now this may be a point of disagreement between Scheffler and Kagan.  

B.  Scheffler suggests that the distribution sensitive conception of the good will be able to do much or most of the work that the constraint would have done.  

C.  That is, Scheffler seems not to be particularly worried about the possibility of options to do harm because he seems to think that the distribution sensitive theory of the good will ensure that most if not all deliberate harms will NOT produce a good SOA, since part of what makes an SOA good is the equality of the distribution of happiness, and Scheffler thinks that actions that deliberately harm others will usually produce a less equal (and therefore less good) SOA.  

D.  If that is true, then Scheffler's hybrid theory will prohibit most deliberate harms not because of a constraint, but because the distribution sensitive theory of the good will normally rule the consequences of such actions to be bad because they make the distribution of utility less equal.  

E.  A good term paper topic would center on this debate:  Is Scheffler wise to reject the ACR/constraint in favor of the distribution sensitive theory of the good as a way to get his theory to prohibit actions which deliberately harm others?  

Why the Moderate Cannot Use a Two-level or Indirect Consequentialist Approach to Justify Her Position

I.  The Moderate may try to claim that extremism is self-defeating.  That is, the moderate may claim that people will actually do MORE to promote the good if they hold a Moderate position than if they hold the extreme position.

II.  The reasons they may think this are basically the same reasons we talked about before when we discussed the idea of "esoteric morality": 

A.  The Moderate may argue that if people subscribe to extremism, they will become discouraged because of its impossibly harsh demands, and so they will end up giving up on morality altogether.  However, if we give them an option to SOMETIMES take "time off" from morality, then they are less likely to quite altogether.  So the moderate might defend options as a means to promoting the good more efficiently.

B.  Similarly, the moderate may defend constraints on familiar rule-utilitarian grounds:  If we have people decide in each situation what action will produce the best consequences, there will be problems of measurement and prediction, as well as problems with bias.  It would be better, then (so the Moderate's argument might go), to have a morality with constraints.  

III.  While Kagan goes through endless variations on the idea, the basic problem is really inherent in the idea of using consequentialism as a basis for defending the rejection of consequentialism. 

IV.  That is, the very DEFINITION of an option and of a constraint MAKES it anti-consequentialist.  The most you can defend on consequentialist grounds is a consequentialist dispensation or a consequentialist presumption.  These are simply MEANS to producing the best SOA.  A genuine option or constraint would operate even in cases in which it leads to a WORSE SOA. 

V.  In short, Consequentialist reasoning (ie, arguments based on the idea that doing such and such will produce better consequences) can only justify a consequentialist dispensation to pay more attention to yourself WHEN DOING SO LEADS TO THE BEST SOA.

VI.  And Consequentialist reasoning can only justify a consequentialist presumption against certain kinds of actions that is to be ignored when doing so leads to the best SOA.

VII.  In either case, we do not have genuine options or constraints, for they OVER-RULE the duty to produce the best SOA.  Consequentialist presumptions and dispensations are simply tools for bringing about the best SOA.