# Parfit’s CTM, Chapter 8: Universal Laws

This marks the seventh of eleven e-meetings of our virtual reading group on Derek Parfit’s Climbing the Mountain—see here for further details. Next week, we will discuss Chapter 9 of the June 7th version of the manuscript, which can be found here.

In chapter 8, Parfit critically examines Kant’s universalizability principle for rightness.  Despite his criticisms, Parfit urges not that we reject, but that we improve one version of this principle, by trading reference to maxims for reference to intentional actions. Parfit considers a dizzying array of versions of the principle, so I’ll only mention some more promising ones in summary, flagging some questions as I go.

Sec. 25 is focused on what Parfit calls “The Impossibility Formula,” which says “It is wrong to act on any maxim that could not be a universal law”, in contrast to the focus of Sec. 26 and 27, which he calls the “Formula of Universal Law”: “It is wrong for us to act on any maxim that we could not rationally will to be a universal law.”  While Parfit doesn’t mention this, if (as seems reasonable to suppose) we cannot rationally will what (we know) cannot be, then the Formula of Universal Law entails the Impossibility Formula.  If that is right, then if Parfit’s attacks on the Impossibility Formula in Sec. 25 are on target, it is hard to see the promise of FUL.

One last preliminary: Parfit uses the label “Formula of Universal Law” in the unusually narrow sense of being only the contradiction-in-will part of what is usually thought of as FUL.  In the usual understanding, FUL has not only a contradiction-in-the-will part, but also a contradiction-in-conception part.  Parfit means to capture this other part separately in the Impossibility Formula.  (I found all of this non-standard usage very confusing, but maybe I’m alone in this.)

Parfit says that (F) is Kant’s actual Impossibility Formula: (F) “It is wrong to act on any maxim of which it is true that, if everyone accepted and acted upon this maxim, or everyone believed that it was permissible to act upon it, that would make it impossible for anyone successfully to act upon it.”  On grounds of both doctrine and defensibility, I’m partial to the view that we should eliminate the “or everyone believed that it was permissible to act upon it” clause (though, Parfit rightly notes, there is some evidence for it being Kant’s view), so I’ll go directly the version that removes that clause (Parfit’s (H)).

Parfit rightly cites Herman’s “convenience killing” maxim as a potential counter-example: “Kill or injure other people when that would benefit me” (p. 144). If everyone acted on this maxim, it wouldn’t be impossible for anyone to successfully act upon it (since there will be at least one person left standing), so it incorrectly comes out permissible.  On this basis, Parfit rejects (F) (and it applies to (H) too).

But another option is just to come up with a more sensitive way of describing willings in maxims that doesn’t get this result.  In general, this escape strategy means that at worst Parfit’s counterexamples leave the plausibility of the formulas open.  He saves discussion of action description for Sec. 27 (to which I’ll return), but it seems that it must be sorted out before rejecting the formulas considered in Sec. 25.

Consider, for example, the maxim “Steal when that would benefit me” (145).  In a world where everyone acted on this maxim, there would still be property, so, says Parfit, in this world theft could sometimes still happen, and therefore this maxim is incorrectly rendered permissible.  But maybe an action redescription is appropriate here: “Steal when that would benefit me for the purpose of reliably having goods at my disposal.”  Once we incorporate this purpose (and it seems unreasonable to suppress it—if there is some other purpose that avoids what I’m about to say, Parfit should tell us what that is), the action is undermined upon universalization.  If everyone’s stealing, I cannot reliably have goods at my disposal, and therefore the maxim is impermissible.

Parfit presents other false positives, but while I have similar concerns about those cases, I won’t go on about that here.  Instead, consider his false negatives to (H), e.g. “Give generously to the poor.”  If everyone did this, the poor would disappear, and then we would fail in our attempts to give to them, apparently making that maxim wrong, incorrectly.  However, since people would achieve their aim of eliminating poverty, Parfit changes the maxim to “Give generously to the poor in order to gain admiration.” This aim would be frustrated upon universalization.

Once again, though, this maxim could be rendered differently: “Give to the poor in order to gain admiration when doing so is possible.”  This is universalizable, but Parfit doesn’t consider such action descriptions.  The same problem arises in his discussion of coordination maxims, such as “Never take the first slice.”  Such a simple maxim would fail the test, but such simple maxims might not be the ones that should be tested.  Rather, the ones to be tested could also (at least implicitly) reference circumstances and purpose, such as “Never take the first slice, unless it’s my turn.”  This is universalizable.

In Sec. 27, Parfit focuses on maxims, in order to ultimately argue that the best Kantian formula—some version of what he is calling the Formula of Universal Law, including the “Law of Nature Formula” and “Moral Belief Formula” from Sec. 26—should replace talk of maxims with talk of intentional actions.  Parfit understands maxims to be either “policies,” which are descriptions of actions, or “policies and motives” which make maxims descriptions of both actions and intended purpose of the action (he doesn’t explicitly recognize a circumstance component).

Parfit has multiple objections to testing maxims, rather than intentional actions.  One is the “Rarity Objection” (154).  If my maxim is “to steal some wallet containing \$63 from some woman dressed in white who is eating strawberries while reading the last page of Spinoza’s Ethics,” then this would be permissible, because it would be “unlikely that anyone else could ever act in this way” (the test of the Law of Nature Formula).

Note first, though, that this maxim might be made impermissible on the Law of Nature Formula if Kantians can come up with a principled reason to render features such as what the woman is reading and wearing irrelevant to the maxim that is to be tested by the Law of Nature Formula.  Such solutions have been proposed, but Parfit doesn’t consider them.

Note also that rare-because-specific maxims can translate into rare-because-specific intentional actions.  I am not sure how Parfit thinks that replacing maxims with intentional actions solves this problem.

Parfit also objects that an Egoist who has only one policy of “Do whatever would be best for me,” where this policy would be rejected by the Law of Nature Formula, would then be wrong to act under this policy not only when he, say, steals, but also (implausibly) when he, say, keeps his promises for self-interested reasons.  Generally, the “Flawed Maxims Objection” (157) is that when we act under a very general policy like the Egoist’s, we sometimes act rightly and sometimes wrongly, but a focus on maxims can’t make this distinction.

But a maxim is, plausibly, just a description of what one intends to do (calling into question Parfit’s whole conception of maxims, I guess).  So if the Egoist intends to keep his promises for self-interested reasons, then this, not simply “Do what would be best for me,” is his maxim.  More broadly, maxims in the right cases bear some feature that is different from similar maxims in the wrong cases. I.e., the more precise maxims are not “unnecessary clutter” (n. 239), but really different maxims.

Maybe another way of putting the concern is this: Parfit holds that we should replace maxim-talk with talk of intentional actions.  But if maxims just are descriptions of what one intends to do, this might be a case of much ado about nothing.

## 34 Replies to “Parfit’s CTM, Chapter 8: Universal Laws”

1. David Shoemaker says:

Josh: Nice summary, and interesting comments and questions along the way. I have two responses/questions regarding what you’ve said, and then two questions of my own.
First, in providing a possible rebuttal to Parfit on his counterexample to (F) and (H), you say, “But maybe an action redescription is appropriate here: “Steal when that would benefit me for the purpose of reliably having goods at my disposal.” Once we incorporate this purpose (and it seems unreasonable to suppress it—if there is some other purpose that avoids what I’m about to say, Parfit should tell us what that is), the action is undermined upon universalization. If everyone’s stealing, I cannot reliably have goods at my disposal, and therefore the maxim is impermissible.” Now perhaps this is a general question about methodology, but on the basis of what is a redescription of the maxim “appropriate” here? Seems to me (a) that the maxim as it stands is perfectly possible and yields the false positive Parfit’s after, and (b) a change in your suggested direction is essentially ad hoc and unmotivated without some *antecedent* and independent criterion of rightness in mind, which is explicitly ruled out by Parfit’s project.
Second, you write, “Note also that rare-because-specific maxims can translate into rare-because-specific intentional actions. I am not sure how Parfit thinks that replacing maxims with intentional actions solves this problem.” But I don’t yet see why this would be the case. In describing the intentional action in the strawberry woman case, e.g., you stop after the physical action-token (stealing money) is described, with no need for the underlying policies to be explicated. Why think the rarity objection would then apply, given the rarity comes only with respect to the specific underlying policies?
My own first question is the point you raised the other day on a comments thread from two weeks ago’s post, having to do with the point of the entire enterprise. Doug has often urged that what Parfit’s doing here is looking strictly for a decision-procedure, and Parfit’s comments to this point have seemed to bear that interpretation out. Nevertheless, here in Ch. 8, the tune has been changed, with Parfit in two places writing that there are *two* goals here, viz., to see if Kantian formulas (a) can help us decide what acts are wrong, *and* (b) “help to explain why these acts are wrong” (152). So what’s changed (if anything)?
But then this leads to my second question. In offering his own improvement over Kant’s universal law formulas, Parfit offer LN3: “We act wrongly unless what we are intentionally doing is something that we could rationally will everyone to do” (p. 159). The focus here is now on intentional actions, and not maxims, but my worry is about what role *universalization* is now supposed to play. In other words, wouldn’t the reason that we couldn’t rationally will certain intentional actions be only that we believed them to be wrong, and not because there’d be some *contradiction* involved in doing so? If so, then there’s just no *explanation* of wrongness provided by LN3; it could at most be a mere decision procedure.

2. Robert Johnson says:

Josh,
Very nice summary.
I wonder what you think of Parfit’s reading of G 424: “But, as Kant points out, wrong-doers do not in fact try to universalize their maxims, so ‘there is really no contradiction’ in these people’s wills.”
Parfit uses this to reply to O’Neill’s claim that you can’t take maxims out of the FUL because that is the only way for there to be a contradiction in the will of wrongdoers (and hence for wrong action to be irrational).
The relevant passage that Parfit refers to seems to me to require maxims anyway. Kant says “not a contradiction here *but instead a resistance….through which the universality of the principle is changed into mere generality and the practical rational principle is to meet the maxim half way*. I’m not confident of my reading of this, but whatever it ultimately means, it still requires maxims. What I *think* it means is that wrongdoers don’t represent what they’re doing as permissible exception to some general rule that isn’t so strict. In any case, it’s no reply to O’Neill

3. Robert Johnson says:

I meant to add that it’s no reply because although wrongdoers hide the contradiction from themselves, were they to consider things rationally, they would see it was there.

4. Doug Portmore says:

Josh,
Thanks for the great critical summary. Below, I want to press you on your proposed redescription of maxims to avoid false positives and false negatives. (Dave has already pressed you on this, but I believe that I have something to add).
You write,

Consider, for example, the maxim “Steal when that would benefit me” (145). In a world where everyone acted on this maxim, there would still be property, so, says Parfit, in this world theft could sometimes still happen, and therefore this maxim is incorrectly rendered permissible. But maybe an action redescription is appropriate here: “Steal when that would benefit me for the purpose of reliably having goods at my disposal.” Once we incorporate this purpose (and it seems unreasonable to suppress it—if there is some other purpose that avoids what I’m about to say, Parfit should tell us what that is), the action is undermined upon universalization. If everyone’s stealing, I cannot reliably have goods at my disposal, and therefore the maxim is impermissible.

On Parfit’s understanding, Kant is claiming that “the wrongness of an act depends on the agent’s actual maxim” (i.e., policy of action). “It is a factual matter what the maxim is on which someone is acting.” So it’s not up to us to redescribe the maxim. We have to go with the maxim on which the agent is actually acting. Parfit also says, “When describing someone’s maxim, as O’Neill claims, we should not include any details whose absence would have made no difference to this person’s decision to do whatever she is doing.” Surely, then, there is someone who acts on the maxim “Steal when that would benefit me.” This person will steal even when doing so doesn’t serve to place any goods at her disposal. Such a person would steal even when she knows that that which she steals will be taken back immediately, provided doing so would provide some net benefit to herself. Perhaps, stealing gives them a rush. Perhaps, stealing will get them membership in some club. The point is to benefit oneself, not necessarily to possess goods. Thus I don’t see how showing that there is some different maxim that can be universalized saves Kant’s formula in question. The fact that some actual agent does, or at least that some hypothetical agent could, act on the maxim “Steal when that would benefit me” (so described) is a sufficient counter-example.
You also suggest revising “Give generously to the poor in order to gain admiration” to “Give to the poor in order to gain admiration when doing so is possible.” But adding “when doing so is possible” seems unwarranted. What difference would that make to the person’s decision that the first part of the maxim doesn’t already cover? If it makes no difference, then this is not the correct way to describe the agent’s actual maxim.

5. Doug Portmore says:

Dave and Josh,
Dave writes:

Doug has often urged that what Parfit’s doing here is looking strictly for a decision-procedure, and Parfit’s comments to this point have seemed to bear that interpretation out. Nevertheless, here in Ch. 8, the tune has been changed, with Parfit in two places writing that there are *two* goals here, viz., to see if Kantian formulas (a) can help us decide what acts are wrong, *and* (b) “help to explain why these acts are wrong” (152).

I never urged that the only thing that Parfit is doing is looking for a decision procedure. I only urged that this is what is most important for us to know if we are to figure out whether we ever have most reason to act wrongly, which is, perhaps, the central aim of the book.
More specifically, my main point was that (and I quote myself) “the basic argument [of chapter 6] is not that the Formula of Humanity is not the best (or correct) account of what makes acts right or wrong, but that the claim that it is wrong to treat people in ways that are incompatible with respect for them ‘would seldom help us to decide, in difficult cases, whether some act would be wrong’.
I don’t see how this quote from chapter 8 conflicts with what we said in the thread on chapter 6, as the two of you seem to be suggesting.

6. Josh, you write:

While Parfit doesn’t mention this, if (as seems reasonable to suppose) we cannot rationally will what (we know) cannot be, then the Formula of Universal Law entails the Impossibility Formula. If that is right, then if Parfit’s attacks on the Impossibility Formula in Sec. 25 are on target, it is hard to see the promise of FUL.

While FUL as stated at the start of section 26 may entail the Impossibility Formula, the revised version of FUL as stated in Parfit’s LN3 and MB2 doesn’t entail the Impossibility Formula, does it? So why is it hard to see the promise of FUL if Parfit’s attacks of the Impossibility Formula are on target?

7. Dave and Doug,
Regarding the “stealing” maxim, you guys ask a couple of pressing questions. Let me try to separate them out; tell me if I’ve missed anything.
1) Isn’t it ad hoc to redescribe maxims in a way that works?
I think the short answer is: it might be. That’s why I tried to put such suggestions in tentative terms, such as “maxim redescription might be appropriate”. But it also might be very principled, and not ad hoc at all, if, first, there is a principle for relevant descriptions, and, second, the redescriptions follow this principle. There’s a cottage industry debating these questions, so it seems that Parfit is too quick to dismiss FUL with counter-examples, unless those examples have taken into account possible redescriptions based on plausible principles for relevance. Notice, finally, that fans of FUL don’t need an antecedent principle for rightness; they need an antecedent principle for moral relevance (such as one that rules out the color of the woman’s clothes and rules in the agent’s purpose in the stealing case).
2) The maxim Parfit uses as a counter-example, while perhaps convertible into some maxim that is better for the Impossibility Formula, is itself a maxim, which some agent might will. So isn’t that maxim a problem, even if the converted one is not? (Millgram has a nice statement of this objection to the redescription strategy at the end of his recent Phil Review paper.)
Doug is right that we should look at what the agent is actually willing, but it is not true that “it’s not up to us to redescribe the maxim.” Or, at least, it’s not entirely up to the agent, either. There are constraints on maxim description, so it’s not totally discretionary. Plausibly, one constraint is that it include reference (at least tacitly) to all of act, purpose, and circumstance. Another is that it include only morally relevant features (thereby, I think, eliminating Parfit’s strawberry-eating reference from the one maxim, for example).
Now Parfit’s original stealing example had the act, “stealing,” and the circumstance, “when it would benefit me,” but no purpose. Under what I consider a plausible constraint on maxim description, I stuck a purpose in there, namely “to reliably have goods at my disposal.” This one gets the right result, which confirms (to me, at least) that Parfit’s original example is corrupted (i.e., it is not a “sufficient counter-example”). But Doug is right that there might be other purposes that are possible for people who steal, such as getting a rush or gaining membership to a club. It is possible that one or more of these won’t get the pre-theoretically intuitive result. We’d have to sort that out one at a time, in a way that comports with the proper constraints on maxim description (whatever those turn out to be). The point, then, isn’t that there are certainly no false positives and negatives. There might be; I was trying to leave that open here (but hopefully there aren’t any!). Rather, the point is more limited: a critic can’t just throw out any kind of maxim and say, “See, that doesn’t work, therefore Kant’s universalizability principle doesn’t work!” Instead, the critic should offer a maxim that describes the agent’s intended act in a way that comports with the defensible constraints set in the most recent literature.
3) A specific worry concerns “Give to the poor in order to gain admiration when doing so is possible.” Doug, you think the “when doing so is possible” part is irrelevant because it makes no difference to the person’s decision to give to the poor.
I don’t follow: if I am thinking about gaining admiration through some means (giving to the poor or anything else), and I learn that those means are unavailable, won’t I (or isn’t it a constraint on rational planning that I) decide not to pursue those means?

8. Josh,
You say, “Now Parfit’s original stealing example had the act, ‘stealing’, and the circumstance, ‘when it would benefit me’, but no purpose. The purpose was I thought implicit. But let me make it explicit: I will steal in order to gain some benefit whenever doing so will benefit me. Isn’t this a sufficient counter-example?

9. Robert,
I like your reading. Maybe this is another spin on, or elaboration of, your point: when Kant says that immoral maxims aren’t universalized, he is simply recognizing that immoral maxims can’t comport with his universalizability principle(s). Thus, if the agent with the immoral maxim were to try utilize the universalizability principle, they would get a contradiction in their wills. That’s just to say that the principle is a (rational) norm; it remains true that some people might not utilize that norm. Looked at from this direction, we get your conclusion again: “[this is] no reply [to O’Neill] because although wrongdoers hide the contradiction from themselves, were they to consider things rationally, they would see it was there.”
And, I agree that it’s hard to see how any of this motivates eliminating maxims.

10. Doug,
I like that one, but let me propose two ways in which it might not work as a false positive, i.e., that it might be rejected by a mainstream Kantian story. Note that these are put forth tentatively, but that this discussion, even in its tentative tone, illustrates why I’m dissatisfied with Parfit’s approach: these are exactly the kind of issues he should (I think) be sorting through before dismissing (H), even if he’s ultimately right that (H) should be rejected.
1) If everyone was stealing from everyone when it benefitted them, Parfit is right that there would still be property (nobody benefits by stealing, say, my used handkerchief). But now it is at least quite likely that the benefit I get from stealing is undermined by the fact that everyone else is stealing from me to benefit themselves. Thus I can’t achieve the very purpose of my maxim, namely to benefit myself, when my maxim is universalized (illustrating again how the purpose can play a key role in the permissibility of a maxim). If memory serves, this is what Korsgaard calls a “practical” contradiction.
2) Even if (1) doesn’t work, a common Kantian approach (the target of, again, Millgram’s recent paper) is that the universalized maxim can’t undermine anything necessary for effective normal agency in a social world with a division of labor and inter-personal coordination. (I.e., not a self-sufficient hermit.) But if every piece of property is up for grabs, that kind of agency might be undermined.

11. Josh,
Regarding (3), okay I see that I was mistaken.
But couldn’t some agent act on the following maxim: “I will give to the poor in order to gain admiration whenever others are watching and giving to the poor is possible.” Couldn’t this and not your suggested redescription be the accurate description of an agent’s actual maxim? That is, couldn’t it be that this person’s policy to try to gain admiration by giving to the poor whenever it is possible to give to the poor and others are watching whether she thinks that she will succeed in gaining admiration or not. And wouldn’t this maxim wrongly fail Kant’s formula, thereby entailing that this agent’s giving to the poor was wrong?

12. Josh, You write, “But now it is at least quite likely that the benefit I get from stealing is undermined by the fact that everyone else is stealing from me to benefit themselves.” I don’t see why. As I suggested earlier, the benefit I’m seeking might be the rush I feel from stealing or the admittance I get into a powerful gang who requires stealing as part of its initiation process. How is that benefit undermined by the fact that everyone else is stealing from me?
You also write, “Even if (1) doesn’t work, a common Kantian approach (the target of, again, Millgram’s recent paper) is that the universalized maxim can’t undermine anything necessary for effective normal agency in a social world with a division of labor and inter-personal coordination.” But that would be some other formula, right? I don’t see where any of the formulas that Parfit is giving counter-examples to says that the universalized maxim can’t undermine anything necessary for effective normal agency in a social world.

13. Doug,
Could you explain why everyone can’t act on your “and others are watching” maxim? (Why would it wrongly fail Kant’s formula?)

14. Doug,
Going back to the stealing one, right, sorry, I misunderstood, thinking that the benefit was the property gained. (See, this is exactly why giving the proper description is so critical!) So now the purpose is not generically “to benefit,” but is specifically, “to benefit by getting into the gang” or “to benefit by getting a rush.”
Could those be (wrongly) universalizable? Maybe. Lemme think on it some more. In the meantime, let me add another suggestion (that Parfit does not consider). An old false positive is “to steal in order to undermine the credit economy.” (Or “to steal in order to undermine capitalism” or whatever). Here, the agent achieves her very purpose if everyone does it, so it seems permissible. One response (possibly from O’Neill?) is to hold that one should test one’s purposes as separate maxims, in addition to testing the maxims that include them. So you’d have to succeed in testing “undermining the credit economy,” or “joining a powerful gang,” or whatever, before the original maxim can be rendered permissible.
Re: the other point, that “I don’t see where any of the formulas that Parfit is giving counter-examples to says that the universalized maxim can’t undermine anything necessary for effective normal agency in a social world.” That depends whether the “everyone” in (F) and (H) refers to ordinary agents in a social world. As Millgram highlights, it is at least true of “New Kantians,” like O’Neill, Herman, or Korsgaard, that this is exactly who “everyone” is. Arguably, that wasn’t Kant’s view, of course. In any case, maybe we should separate (F) and (H) into (F1/F2) and (H1/H2), where “everyone” is understood differently in (1) and (2). If that’s done, then, you’re right, only one of the formulas so understood could utilize the “New Kantian” solution.

15. Josh, You ask, “Could you explain why everyone can’t act on your “and others are watching” maxim? (Why would it wrongly fail Kant’s formula?).” Before long, if this maxim was universally acted upon, it would be impossible for anyone to succeed in being admired for such a common and mundane act.

16. Josh,
Why can’t the purpose still be generically “to benefit.” It may be that I have no goods that anyone could steal from me, but that I want to be in a powerful gang or want the rush that stealing gives me. It seems, then, I would likely still benefit by acting on my maxim even if my maxim was universally followed.

17. Doug,
I guess I’m not seeing why it would be impossible to be admired in such a situation.
In the other case, your maxim could be “to benefit generically” if that was your purpose. But you described it as “to benefit by getting a rush,” I thought.

18. I should add, “to benefit generically” strikes me as an almost incoherent purpose of some specific action. Usually, at least, we perform the action to secure a benefit because of some specific relation between the act and the benefit, in this case getting a rush (or joining a gang, or whatever).

19. Dave,
You wrote:

Second, you write, “Note also that rare-because-specific maxims can translate into rare-because-specific intentional actions. I am not sure how Parfit thinks that replacing maxims with intentional actions solves this problem.” But I don’t yet see why this would be the case. In describing the intentional action in the strawberry woman case, e.g., you stop after the physical action-token (stealing money) is described, with no need for the underlying policies to be explicated. Why think the rarity objection would then apply, given the rarity comes only with respect to the specific underlying policies?

Right, I should have been clearer here. What I should have said was that the very nearby problem of relevant descriptions applies to peoples intentions (as it does to maxims, if maxims are descriptions of peoples intentions, and to beliefs in the generality problem, etc.). So maybe Parfit can find a reason to, as you say, stop at the physical act-token. But, note, so could one do this with maxims. The question is (as you noted above) what principled reason there is for including only so much information. If my intention is not just to steal, but steal from a strawberry-eater (as Parfit supposes it is), why not include all that information?
I think there are good reasons not to (viz., it’s not morally relevant that she’s eating strawberries), but, again, that point can be made just as well by the fan of maxims.

20. Dave,
You also write:

In offering his own improvement over Kant’s universal law formulas, Parfit offer LN3: “We act wrongly unless what we are intentionally doing is something that we could rationally will everyone to do” (p. 159). The focus here is now on intentional actions, and not maxims, but my worry is about what role *universalization* is now supposed to play. In other words, wouldn’t the reason that we couldn’t rationally will certain intentional actions be only that we believed them to be wrong, and not because there’d be some *contradiction* involved in doing so? If so, then there’s just no *explanation* of wrongness provided by LN3; it could at most be a mere decision procedure.

I sympathize with the claim that FLN (including LN3) is only a decision procedure. (I follow, not surprisingly, Timmons on this, where FH is the moral criterion.) But in defense of LN3, I would say that universalization is playing a role because we’re seeing what would happen if everyone had the intention in question.
It’s worth adding that preserving a role for universalization is important even if LN3 were only a decision procedure.

21. I have a general question I forgot to ask:
Is anyone clear on exactly which of the various universalizability formulas Parfit wants to keep? Maybe I’m just being dense, or maybe I just missed it, but this wasn’t clear to me. Was it LN3? Or MB2? Or both?

22. Josh,
Here’s the dialectic as I understand it:
I said: The maxim is “I will steal in order to gain some benefit whenever doing so will benefit me.”
You replied: “But now it is at least quite likely that the benefit I get from stealing is undermined by the fact that everyone else is stealing from me to benefit themselves.”
I replied: Not so. It may be that the benefit I get from stealing is not at all undermined by the fact that everyone else is trying to steal from me, for the benefit in this particular instance could be, say, gaining admittance to a powerful gang.
You replied: “So now the purpose is not generically ‘to benefit’, but is specifically, ‘to benefit by getting into the gang’ or ‘to benefit by getting a rush’.
I replied: Not so. The purpose stated in my maxim (my policy) is “to benefit.” There’s a difference between the general purpose in my maxim/policy of stealing and the specific instance of that purpose being achieved in this particular instance of action.
You reply: “‘to benefit generically’ strikes me as an almost incoherent purpose of some specific action. Usually, at least, we perform the action to secure a benefit because of some specific relation between the act and the benefit.
Now the end for which we act can be described at various levels of abstraction. You could say that I work to get money. But if I want the money only as a means to get happiness, then it seems that we should say that my ultimate aim is to be happy and so I work in order to increase my happiness. Why, then, is it incoherent to suggest that my ultimate aim in performing a specific act is to further my self-interest (i.e., to benefit myself)?

23. Doug,
Okay. I should clarify a bit. I assumed that Parfit was not working exactly with the problem of relevant descriptions. That is, he wasn’t saying: “Here’s a maxim, it can be described in one way (which fails the test) or another (which passes).” Instead, he was giving us some maxims that have determinate, already relevant (ex hypothesi) descriptions attached to them.
But if we put it, as you do (and maybe as Parfit does–maybe I was reading too much into him there), as a matter of levels of abstraction, we’re doing something else. I wanted to say that the purpose “to benefit generically” was incoherent. You say “to benefit” is the agent’s purpose at the most abstract level. Those are two different purposes: “to benefit” (abstracting) and “to benefit generically.”
So, to get to your question, I agree with you that it is coherent to say either that you can abstract away from the specific kind of benefit and just describe it as “a benefit”, or, as you put it in your most recent post, that you can understand “benefitting myself” as my “ultimate aim.”
The question, then, is whether that’s the appropriate level of abstraction, i.e., the problem of relevant descriptions. Now I don’t know what the best solution is (though I have my suspicions), but one that would rule out the very abstract description is Herman’s point that we should capture maxims “as they are willed.” That is, since I’m trying to join a gang, then, if that’s morally relevant by all other principles of relevance, it should be included in my maxim, and we shouldn’t go too abstract, because that conceals what I’m willing. In any event, though, I would have liked to see Parfit address these kinds of consideration before dismissing the project.

24. Josh,
I don’t think that Parfit has failed to “address these kinds of consideration(s).” He explicitly states that we should capture maxims “as they are willed.” He says, for instance, “It is a factual matter what the [actual] maxim is on which someone is acting.” So he agrees with Herman on this point, I think. Moreover, he addresses the issue of how to determine the relevant descriptions: He says, “When describing someone’s maxim, as O’Neill claims, we should not include any details whose absence would have made no difference to this person’s decision to do whatever she is doing.”
In your original post, you wanted him to consider re-descriptions of the maxims that he gives that yields false positives and false negatives. But we don’t have to consider re-descriptions. It is enough that it is possible that some agent might actually will the maxim as he describes it to show that Kant’s formula yields false negatives and false positives. He says, for instance, “Even if actual wrong-doers never act on such highly specific [or, in our stealing case, highly abstract] maxims, we can imagine such people. Kant’s formula ought to be able to condemn these imagined people’s acts. So how does Herman’s point that we should capture maxims “as they are willed” rule out the possibility that some imagined person has the highly abstract maxim “I will steal in order to benefit myself whenever stealing will benefit me”?

25. David Shoemaker says:

Sorry for the late reply, Josh. To your most recent question of me, that you think there’s still universalization work being done by LN3, viz., that we can “see what happens” were this to be everyone’s intentional action, that still doesn’t answer my question. For what’s the problem we’re looking for now supposed to be? A contradiction of some sort? But what, exactly, would that contradiction consist in? The only clue Parfit gives us is that, with respect to some actions now, “I may not be able rationally to will it to be true either that everyone acts in this way, whenever they can, or that everyone believes such acts to be morally permitted” (160). But why not? Will it just be because of some really bad consequences? But that wouldn’t be Kantian at all. Would it be because of some *other* criterion of wrongness? But that would undermine the point of appealing to the formula as a criterion of rightness. So now I’m just not getting what would *make* it some action I couldn’t rationally will.
As to your question about which formula Parfit’s advocating, I think it’s both LN3 and MB2. Did you think they might conflict?

26. Robert says:

David,
Perhaps Parfit is advocating a version of the ‘practical contradiction’ interpretation, viz., that I can’t rationally will to x and will that everyone x, since everyone’s xing precludes me from xing. Or, I can’t achieve my ends when everyone acts that way. That seems irraitonal.

27. David Shoemaker says:

Thanks, Robert. I thought of that, but then I immediately thought of the kind of counterexamples raised against the caricatured Kant in intro classes, the “what if everyone did that?” parodies of Kant. For example, if my intentional action is now to run to the store to buy a Coke, I couldn’t rationally will that everyone do that because if everyone did that it would frustrate my ends of getting a Coke (all the Cokes would be gone by the time I got there, perhaps), in which case running to the store to buy a Coke is wrong. Presumably Kantians can get out of these cases by focusing on the particular maxims involved in each case, but if Parfit’s abandoned maxims for merely intentional action, I don’t see how he can now avoid these caricature counterexamples.

28. Robert Johnson says:

So really he just needs a solution to the problem of relevant act descriptions.
Couldn’t he say that the relevant act description is the one included in the content of the intention, rendered as general as it can be rendered without making a difference to whether or not he intentionally did the act. He jettisoned maxims, but not all propositional content to intentional actions. So perhaps he can make use of that.

29. Doug,
As I said above, while I agree with the claim that it is necessary that a description reflect what the agent actually wills, it is not true that “It is enough that it is possible that some agent might actually will the maxim as he describes it.” Parfit’s description must also comport with the other criteria for relevance besides merely that some agent might will it. For obvious example, the description must elide the temporal location of the act, even if the agent (or Parfit, or anyone) were to insist that the temporal location was key to her maxim. So sometimes we do have to consider re-descriptions: it’s not enough that some agent could will the maxim as Parfit describes it; it is enough that some agent could will the maxim as Parfit relevantly describes it.
You’re right that Parfit does consider some elements of the debate over relevant descriptions, but as far as I can tell, he doesn’t consider enough to conclusively dismiss the formulas he wants to dismiss. That doesn’t mean that we couldn’t aid him, or that the formulas are all correct; it just means that there are others things to consider before dismissing at least some of the formulas. I think that the most obvious case of that was with the strawberry-eating woman case, where some folks have proposed criteria to rule out the fact that she’s eating strawberries, etc., as morally irrelevant, even if the agent makes a big deal about this fact.
Now you’re right that Herman’s “as they are willed” point doesn’t necessarily rule out the very abstract maxims. I didn’t mean to suggest otherwise (at least I don’t think I meant to). Rather, I was bring that point to bear on your suggestion that someone might seek to benefit by joining a gang or getting a rush. If that’s what they’re seeking, then Herman’s point suggests that we should include in the maxim those particular ways of benefitting. But, you’re right that it’s possible that someone might just say “I want to benefit in no particular way.” That’s possible, but that’s what I think might be nearly incoherent. If someone says “I will go to the market to benefit in no particular way,” I would think that person was either insincere or confused. If, instead, this agent were just abstracting from her particular ways of benefitting, we might have to get more specific about her particular purpose(s) and include them in the maxim to satisfy Herman’s criterion for relevance.

30. Thanks, Dave. I didn’t have any view on whether LN3 and MB2 might conflict (or, for that matter what their relation might be more generally), I just wasn’t clear on where Parfit ended up on that question.

31. David Shoemaker says:

Robert: That would have to be the idea, I guess, but then it seems a lacuna remains regarding what principled reason there is to articulate the content of an intention one way over another, if there are multiple general (and accurate) descriptions of the content of the intention. This is especially important given that Parfit remarks that what I think I’m doing may be irrelevant. He also says (160) that while O’Neill thinks her appeal to maxims can resolve the problem of morally relevant descriptions, that is actually not the case. But I don’t yet see how his favored replacement that doesn’t appeal to maxims fares any better.
On a different note, I’m puzzled by Parfit’s formulation in LN3 of wrongness as a kind of default state of our actions: “We act wrongly *unless* what we are intentionally doing is something that we could rationally will everyone to do.” That strikes me as carrying with it too much of the “born into a state of sin” religious hangover in some of Kantian ethics. I think of the default state of our actions as permissible unless they violate some moral constraint (this is the way in which it comes across in Scanlon, for instance).

32. Robert Johnson says:

Maybe you’re reading the ‘unless’ as ‘except when’. But ‘unless’ just makes this into a necessary condition.

33. David Shoemaker says:

I realize it’s simply being rendered as a necessary condition. I was just puzzled by it more as a matter of rhetorical style than anything else.
On the other note, as it turns out, Parfit begins the next chapter (“What if Everyone Did That?”) by addressing the concern I’d been raising about what it is that renders some intentional action one that I could not rationally will. Should’ve read ahead, I guess. But here’s where he’s heading in a decidedly consequentialist direction (although I’ll leave this for the next chapter summary).