Parity, Moral Options, and the Weights of Reasons

The (moral) permissibility of an act is determined by the relative weights of reasons, or so I assume. But how many weights does a reason have? Weight Monism is the idea that reasons have a single weight value. There is just the weight of reasons. Weight Pluralism holds that reasons have at least two weight values and these values aren’t always equivalent. The simplest versions of Weight Monism hold that the weight of each reason is either equal to, weightier than, or less weighty than every other reason. We’ll see that this simple view leads to paradox. We must complicate the picture somehow.

altruistic ones.For example, it can be permissible to buy a car than to do the better act of giving the money to an effective charity.Since the latter is better than the former, this moral option can't be explained by equality or parity.We need a different mechanism to explain how such cases can be moral options.One such mechanism is provided by our second candidate complication, Weight Pluralism.
Following Josh Gert, (my version of) Weight Pluralism distinguishes between a reason's justifying and requiring weight.A reason's justifying weight is, roughly, how good it is at making an act permissible.A reason's requiring weight is, roughly, how good it is at making a (permissible) act required.When reasons are better at making acts permissible than making them required, they are well suited to explain moral options.You can be permitted to choose the worse selfinterested benefits of the car over the better altruistic benefits of the charitable donation, because self-interested reasons are good at making acts permissible (have lots of justifying weight).The self-interested benefits don't require you to buy the car over donating, because they are bad at making acts required (have little to no requiring weight).Hence, Weight Pluralism explains the moral options that Parity Monism can't.
Even if my objection to Weight Monism is plausible, Weight Pluralism seems to have an even bigger problem.Parity Monism is designed to explain a puzzle that I refer to as the normative significance of small improvements.Apparently, Weight Pluralism can't explain this puzzle without contradicting itself (Cullity 2018; Rabinowicz 2008, 2012).But appearances can be deceiving.The goal of this paper is two-fold: provide a Weight Pluralist explanation of the normative significance of small improvements and show that we should prefer Weight Pluralism over Parity Monism.
The first two sections of the paper reveal that Weight Pluralism can explain a puzzling case that Parity Monism can't, and so we have a reason to prefer the former over the latter.The middle two sections develop and defend the framework that we'll apply to explain the normative significance of small improvements.The final two sections apply that framework and show that Weight Pluralists can easily explain the normative significance of small improvements.
In short, Parity Monism doesn't match Weight Pluralism's explanatory power.Weight Pluralism explains the normative phenomena that Parity Monism is meant to explain and the phenomena that it can't explain.Parity is no substitute for Pluralism.

1
Against Weight Monism

Simple Weight Monism and the Normative Significance of Small Improvements
Consider Simple Weight Monism, which adds things to Weight Monism: Three Comparatives: Every reason is either weightier than, equally weighty as, or less weighty than every other reason; Monist Permissibility: φ is permissible if and only if the reason for φ is not less weighty than the reason for ∼φ.Monist Requirement: φ is required if and only if the reason for φ is weightier than the reason for ∼φ.
This simple position leads to paradox in at least two different ways.The first paradox is what I call the normative significance of small improvements.It involves an apparent tension between two ideas: the pro tanto maximization of certain reasons and the stability of moral options involving those reasons.Start with the former.Altruistic reasons (to prevent suffering 1 ) are pro tanto maximizing: if you don't have to trade altruism off against some other morally relevant consideration, then you are required to choose the biggest altruistic benefit that you can.If you could costlessly save an additional life, then you are required to do it.
Simple Weight Monism-thanks to its commitment to Monist Requirement-has no problem accommodating the pro tanto maximization of altruistic reasons.Suppose that you have a choice between altruistic A (saving 5 lives) and altruistic A+ (saving those same lives and an additional life).Intuitively, you are required to choose A+ over A. Monist Requirement, if true, can explain why.It says that weightier reasons give you a requirement.And you have weightier reason to save all 6 rather than just 5.
Our reasons to respect rights are also pro tanto maximizing in the sense that, if you don't have to trade respecting rights off against some other morally relevant consideration, then you are required to respect as many rights as you can.If you can costlessly respect an additional right, then you are required to do it.Suppose that you have a choice between respectful R (respecting Jerry's right not to be beaten up) and respectful R+ (respecting both his right not to be beaten up and his right not to be insulted) Intuitively, you are required to choose R+.Monist Requirement can also explain this requirement, as you have weightier reason to respect both rights than to respect only one.
But consider the tradeoff between the altruistic prevention of suffering and the respect of rights.Suppose that you are forced to choose between A and R, i.e., the only way that you can save five people is to beat up Jerry.Intuitively, the choice is a moral option.You can permissibly choose A and you can permissibly choose R. The moral option remains if we consider the choice between the same altruistic A and the slightly more respectful R+.If the only way to save 5 is to beat Jerry up and insult him, it is still permissible to do the altruistic action.Respect of an additional small right is not enough to make one required to choose respect over saving five lives.The moral option between altruism and respect is stable insofar as increases/decreases to one of the opposing reasons doesn't convert the moral option into a requirement.
Simple Weight Monism can't handle the stability of moral options.The problem emerges from its implicit assignment for moral options.It's commitment to Three Comparatives tells us that there are only three possibilities concerning the relative weight of the reasons for A and R: the reason for A is weightier, the reason for R is weightier, or they are equally weighty.Its commitment to Monist Requirement tells us that you are required to do whatever you have weightier reason to do, and so the first two possibilities give us requirements, not moral options.Simple Weight Monism entails, then, that the choice between A and R is a moral option exactly when the reasons for A and R are equally weighty.
Simple Weight Monism's assignment for moral options makes moral options fragile.Small increases/decreases to one of the opposing reasons breaks the tie and so makes one of the two alternatives required.For example, if the choice between A and R is a moral option, then Simple Weight Monism entails that you are required to choose R+ over A. To see that this conditional is true, consider: 1 For simplicity, I generally suppress the parenthetical phrase.I assume only that altruistic reasons to prevent suffering are pro tanto maximizing.Perhaps altruistic reasons to provide pure benefits are not pro tanto maximizing.For example, suppose that Imani already has a great life and you can give her a nice gift or a nicer gift.Maybe you aren't required to give her the nicer gift even if you could costlessly do so.I take no stand on this.The first premise follows from Simple Weight Monism's assignment for moral options and the assumption that the choice between A and R is a moral option.The second premise came up in our discussion of pro tanto maximization.The only relevant difference between R and R+ is that you would respect an additional right in R+, and so you have weightier reason for it.
The third premise follows from the conjunction of the first premise (the reasons for A and R are equally weighty) and the Substitutability of Equality: "if two items are equally good with respect to V, one can always be substituted for the other in comparisons with respect to V" (Chang  2017: 4).When V concerns weight, the principle tells us that equally weighty things always compare equally to the weights of other things.If the reasons for A and R are equally weighty, the principle tells us that the third premise is true, that no reason can be weightier than R without being weightier than A. Analogously, since the square root of 25 equals 5, 10 can't be greater than 5 without being greater than the square root of 25.
The fourth premise is an application of Monist Requirement.Together the four premises reveal that Simple Weight Monism can't handle the stability of moral options.Any small increase in one of the reasons breaks the tie and converts the moral option into a requirement.If the choice between A and R is a moral option, then you are required to choose R+ over A. Intuitively, however, moral options are often stable.If it is permissible to beat Jerry up to save five lives, then it is permissible to beat him up and insult him to save five lives.Jerry's right not to be insulted is too small to convert the moral option into a requirement to choose respect over altruism.Since Simple Weight Monism can't explain the stability of moral options, it can't explain the normative significance of small improvements.
Small improvements, such as A+ and R+, pose a puzzle or a paradox because their liability to generate requirements depends on whether we are trading off different morally relevant considerations.When we aren't trading off different considerations, then we are required to choose the small improvement.We are required to choose A+ over A and R+ over R. On the other hand, when we are trading off different morally relevant considerations (altruism vs respect), moral options are stable and the small improvement isn't liable to generate a requirement.If the choice between A and R is a moral option, then the choice between A and R+ and the choice between R and A+ are also moral options.
To explain this differential significance, a normative theory must be able to explain both the pro tanto maximization of some reason and the stability of moral options involving that reason.It is difficult to do both.Simple Weight Monism explains pro tanto maximization by Monist Requirement, by holding that a weightier reason to choose φ over ∼φ always generates a requirement to φ.It rules out the stability of moral options when it further assumes Three Comparatives, that the only way to avoid one reason's being weightier is for the reasons to be equally weighty.Perhaps Three Comparatives is the problem.Perhaps the weight of two reasons can be comparable even though they aren't comparable by the traditional three comparatives (weightier than, less weighty than, equally weighty as).If there were a fourth comparative, perhaps we could combine pro tanto maximization with the stability of moral options, thereby explaining the normative significance of small improvements.

Simple Weight Monism and Dorsey's Car Case
Simple Weight Monism faces a second paradox.Its commitment to Monist Permission tells us that φ is permissible exactly when the reason for φ isn't less weighty than the reason for ∼φ.Three Comparatives tells us that the only way to not be less weighty is to be at least as weighty, i.e., equally weighty as or weightier than.Since the relation at least as weighty is transitive, it follows that the makes-it-permissible-to-act-against relation is transitive.In other words, Simple Weight Monism entails: Justification Transitivity: If R1 makes permissible acting against R2 and R2 makes permissible acting against R3, then R1 makes permissible acting against R3.
We get a version of the Justification Intransitivity Paradox whenever our intuitive judgments about certain cases conflict with Justification Transitivity.I focus on a version of the paradox inspired by Dorsey (2016: ch 4). 2 Suppose that purchasing a car would cost $17,000, and it would allow you to accept a more satisfying position and participate in your preferred leisure activities.Intuitively, it is permissible to buy the car for these self-interested reasons, as long as you've already made sufficient efforts to help others and give to charity.Yet, instead of buying the car, you also have the option to donate that $17,000 to the Malaria Consortium.GiveWell (2020) estimates that every $3,373 donation to that foundation's chemoprevention program saves a child from death.$3,373×5 is approximately $17,000, the cost of the car.Hence, it is morally permissible to get a car rather than save the lives of five people.
The lives of five people nonetheless have significant weight in their own right, as it is permissible to beat up an innocent person, Jerry, to save the lives of five people.So far so, so good.The paradox emerges when we notice that it is impermissible to beat up Jerry to get a car.Let's refer to your self-interested reasons as Car, the altruistic reasons as 5Lives, and the rights-based reason as Respect.The following claims are individually plausible but jointly inconsistent: Car Justifies Against 5Lives: It is morally permissible to get a car rather than save the lives of 5 people.5Lives Justifies Against Respect: It is morally permissible to save 5 people's lives even if you have to beat up an innocent person, Jerry, to do it.Justification Transitivity: If Car makes permissible acting against 5Lives and 5Lives makes permissible acting against Respect, then Car makes permissible acting against Respect.And yet: Respect is Required Against Car: It is not morally permissible to beat up an innocent person, Jerry, to get a car.
Dorsey's Car Case is the conjunction of the three deontic verdicts, i.e., Car Justifies Against 5Lives, 5Lives Justifies Against Respect, and Respect is Required Against Car.If these three verdicts are true-as I and others assume them to be3 -Justification Transitivity and Simple Weight Monism are false.They must be replaced, but with what?Three Comparatives tells us that weights are always comparable by weightier than, less weighty than, or equally weighty as.This assumption guarantees that weight comparatives are always transitive.Perhaps that assumption is the problem.

1.2.1
Parity Monism and the Normative Significance of Small Improvements The problems for Simple Weight Monism seem to suggest that the weights of reasons can be incommensurable, or not precisely comparable, i.e., not comparable by weightier than, less weighty than, or equally weighty as (cf.Chang 2017: 6).Incommensurable weight values can still be comparable if there is a fourth comparative that imprecisely compares them.Ruth Chang's notion of parity is designed to make such imprecise comparisons.In this sub-section, we will extend her argument for the claim that parity exists to see whether it can address the problems for Simple Weight Monism.Chang (2002) originally used parity to explain the axiological significance of small improvements (e.g., how do we explain the intuitive value comparisons involving A, A+, R, R+?).Yet we are focused on how she uses parity to explain the normative significance of small improvements (how do we explain the intuitive deontic verdicts of pairwise choices involving A, A+, R, and R+?).On her view, the axiological and normative significance are tightly linked.
Chang assumes that "Evaluative relations correspond in a one-to-one way with distinctive [moral4 ] responses " (2017: 9; cf.Rabinowicz 2008: 37-39).For our purposes, the most important correspondence is: Required iff Better: one is required to choose X over Y iff X is better than Y (equivalently, Y is worse than X).(Chang 2017: 9;cf. Rabinowicz 2008: 37-8) 5 This correspondence allows us to derive an explanation of the normative significance of small improvements from her argument that parity exists.
Recall that A and R are different kinds of good actions.Altruistic A is saving 5 lives and respectful R is respecting Jerry's right not to be beaten up.A+ and R+ are respective small improvements.
responses.Roughly: you are required to be indifferent between φ and ∼φ iff φ and ∼φ are equally good, and you have the moral option to prefer A+ over B or B over A+ iff A+ and B are on a par.See, e.g., Chang (2017: 12-13) and Rabinowicz  (2008: 38-39).
Altruistic A+ is saving those same 5 lives and an additional one.Respectful R+ is respecting both Jerry's right not to be beaten up and his right not to be insulted.Now consider: Chang's Small Improvement Argument (2002: §1, 2017: 4) P1: A+ is better than A. P2: A+ isn't better than R. P3: If A+ is better than A but not better than R, then A and R are not equally good.[Substitutability of Equality] C1: Therefore, A and R are not equally good.[P1, P2, P3] P4: Neither A nor R is better than the other.C2: Therefore, A is not better than, worse than, or equally good as R. [C1, P4] Chang's Small Improvement Argument has four premises.The third premise is just an application of the Substitutability of Equality.If A and R are equally good, then A+ can't be better than A without being better than R.That would be like saying that 10 can be greater than 5 without being greater than the square root of 25.
The other three premises-P1, P2, and P4-follow from the intuitive deontic verdicts and Required iff Better.The first premise concerns the pairwise choice between A+ and A. You are required to choose A+ over A, so Required iff Better tells us that P1 is true, that A+ is better than A. The second premise concerns the pairwise choice between A+ and R.You have the moral option to choose A+ or R, so Required iff Better tells us that P2 is true, that A+ is not better than R. The fourth premise concerns the pairwise choice between A and R.You have the moral option to choose A or R, so Required iff Better tells us that P4 is true, that neither A nor R is better than the other.
The conclusion of the Small Improvement Argument tells us that A and R aren't comparable by any of the three traditional comparatives.A is not better than, worse than, or equally good as R. Chang argues that they are nonetheless comparable, and so there must be a fourth comparative.Chang concludes, more specifically, that A and R are on a par. 6hen we combine Chang's Small Improvement Argument with the above defense of each premise, we get a straightforward explanation of the normative significance of small improvements.Required iff Better explains pro tanto maximization.You are required to choose A+ over A, because A+ is better than A. Parity explains the stability of moral options.The choice between A and R is a moral option and the moral option remains in A vs R+, because A is on a par with both R and R+.
The Improvements Would Make Requirements argument ( §2.1) and Chang's Small Improvement Argument ( §2.2) reveal that equality cannot explain the stability of moral options.Parity can explain what equality cannot, because goods on a par are not always substitutable.Equality is so precise of a comparison that if A and R are equal, then they always compare in the same way to other values (Substitutability of Equality).If A is equally good as R, then A+ can't be better than A without being better than R.But parity is like an imprecise version of equal to, and this imprecision will introduce failures of substitution.If A is on a par with R, then A+ can be better than A without being better than R.For R can be on a par with both A and A+.
This failure of substitution entails that parity is intransitive (Chang 2017: 15).A is on a par with R and R is on a par with A+, but A is not on a par with A+.For A+ is better than A. This intransitivity is what allows us to combine Chang's explanation of pro tanto maximization with her explanation of stable moral options.It allows R to be on a par with both A+ and A even though A+ is better than A.
We are assuming that (im)permissibility is determined by the weights of reasons, but, so far, Chang hasn't told us anything about the weights of reasons.She holds that goods and reasons are so closely linked that they can be used interchangeably, at least when it comes to how they explain deontic status (2016: 214-5).There is, indeed, a natural way to link value and weight relations that is friendly to Weight Monism: better indicates weightier reason, worse indicates less weighty reason, equal indicates equally weighty reason, and goods on a par indicate weights on a par. 7ince A+ is better than A, the reason for A+ is weightier than the reason for R. Since A and R are on a par, the reason for A is on a par with the reason for R. In effect, Chang's explanation of the normative significance of small improvements suggests that Three Comparatives should be replaced by: Four Comparatives: every reason is either weightier than, less weighty than, equal to, or on a par with every other reason; This one change to Simple Weight Monism gives us Parity (Weight) Monism, which is the conjunction of Weight Monism (reasons have only one weight value), Four Comparatives, Monist Permissibility (permissible just when the reason for it is not less weighty), and Monist Requirement (required just when there is weightier reason for it).In addition to Chang (2017), Cullity  (2018: 431) also seems sympathetic to Parity Monism.We've seen that this position explains the normative significance of small improvements.What else can it do?

Parity Monism and Dorsey's Car Case
At first glance, Parity Monism nicely resolves Dorsey's version of the Justification Intransitivity Paradox.The paradox resulted from combining Justification Transitivity with the three deontic verdicts in Dorsey's Car Case: Car Justifies Against 5Lives, 5Lives Justifies Against Respect, and yet Car doesn't justify against Respect.For Respect is Required Against Car.When two reasons are on a par, neither is less weighty than the other.Hence, Simple Weight Monism's commitment to Monist Permission entails that each reason makes it permissible to act against the other.Since parity is intransitive, it follows that justification (the makes-it-permissibleto-act-against relation) is also intransitive.Dorsey's Car Case illustrates this failure.The proponent of Parity Monism will suggest that Car is on a par with 5Lives, 5Lives is on a par with Respect, and yet Car is not on a par with Respect.For Respect is weightier than Car.When combined with Parity Monism, these comparisons entail the relevant deontic verdicts.Car makes it permissible to act against 5Lives, because the two reasons are on a par.5Lives makes it permissible to act 7 Other ways of linking goods and reasons rule out Weight Monism.The Weight Pluralist might endorse axiological parity (one value's being on a par with another) but reject weight parity (one reason's weight's being on a par with another reason's weight).Such a Pluralist could cash out axiological parity and its normative significance by appealing, not to weight parity, but to justifying and requiring weight values that are fully commensurable (always comparable by >, <, and =).This seems to be Gert's strategy (2004 and maybe also his 2015).I am officially neutral on the existence of axiological (and weight) parity.
against Respect, because the two reasons are on a par.It is impermissible to act on Car rather than Respect, because Respect is weightier than Car.
Parity Monism, then, apparently provides an attractive explanation of why Justification Transitivity fails and why the three deontic verdicts in Dorsey's Car Case are true.In the previous subsection, we saw that Parity Monism explains the normative significance of small improvements.Why doesn't the paper end here? 8he primary motivation for Parity Monism and the existence of weight parity is their ability to explain the normative significance of small improvements.In § §5-6, I show that, despite appearances, Weight Pluralism can explain the same phenomena without appealing to incommensurability or parity.This undermines the standard case for Parity Monism (and the existence of parity), leaving us no reason to prefer it over my Pluralist alternative.
Furthermore, even if we allow that weights can be incommensurable and on a par, these features of weight cannot explain Dorsey's Car Case after all.Acting on 5Lives (aka A) and acting on Respect (aka R) are both good states of affairs.Arguably, their axiological value is on a par.(Terminology note: 'A' and 'R' refer to options and '5Lives' and 'Respect' respectively refer to the reasons for those options.)It is plausible, then, that the weights of 5Lives and Respect are also on a par.
Yet compare acting on Car and acting on 5Lives.Axiologically, these options are not on a par.It is far better to save five lives than to get the self-interested benefits of the car (cf.Hurka and Shubert 2012: 1; Muñoz 2021: 713, nt 12).I sincerely believe that you are special.You should be proud of yourself.Still, it wildly exaggerates your importance to say that your self-interested benefits from the car are on a par with the lives of five people.Since acting on 5Lives is better than acting on Car, it doesn't seem particularly plausible that the weights for Car and 5Lives are on a par.Hence, Parity Monism seems poorly suited to explain why Car Justifies Against 5Lives.
The same case provides a counterexample to Required iff Better.Acting on 5Lives is better than acting on Car, but you aren't required to act on 5Lives.Since Required iff Better is false, a key premise in the argument for weight parity is also false.
Maybe you are even more impressive than I give you credit for and your self-interested benefits really are on a par with the lives of five people.The Parity Monist is no better off.It is supererogatory to act on 5Lives rather than Car.In other words, it is morally optional and morally better to act on 5Lives over Car.If acting on 5Lives is on a par with acting on Car, then we can explain why acting on 5Lives is morally optional.But then it is mysterious how it is morally better to act on 5Lives rather than Car. 9 In other words, parity can't explain both features of supererogation at the same time.If two options are on a par and are thereby morally optional, then neither is morally better than the other.If one option is morally better than the other, then Parity Monism has no explanation of how the choice is a moral option rather than a requirement. 8Hurka and Shubert would answer that commonsense morality "isn't thinking about the [imprecise] comparability of reasons; it has no view about so recherché a topic" (4).That answer seems uncharitable.Parity Monists needn't assume that morality has an eccentric interest in imprecise comparability.They assume that φ is permissible only if it isn't worse than the alternative and parity is simply one way that φ might fail to be worse. 9The Weight Pluralist can explain the moral betterness of the altruistic act in two different ways: by appealing to the distinction between justifying and requiring weight (Muñoz & Pummer forthcoming)  I am open to the idea that weights can be incommensurable and on a par; however, we need a different mechanism to explain Dorsey's Car Case.We need a way to explain how it can be permissible to act on Car over 5Lives even though it is far better to act on 5Lives.
For the rest of the paper, I assume Three Comparatives, that all normative weights are comparable by weightier than, less weighty than, or equally weighty as.Since each of these comparatives is transitive, this assumption guarantees that all weight comparatives are transitive.If nothing else, it will be interesting to see whether we must abandon the transitivity of weight comparatives to explain Dorsey's Car Case and the normative significance of small improvements.Spoiler: we don't.

2
Weight Pluralism and Dorsey's Car Case

Pluralist Permissibility
Recall the three things that Simple Weight Monism added to Weight Monism.Perhaps the primary problem with Simple Weight Monism has nothing to do with its claim about comparatives.Perhaps the primary problem is with its assignment of deontic statuses in terms of a single weight value of reasons.The rest of the paper explores this possibility.It develops and defends a Weight Pluralist approach to explaining moral options and the normative significance of small improvements.This approach assigns deontic status in terms of two weight values of reasons, justifying and requiring weight.
A reason's justifying weight for φ (JW φ ) is how good the reason is at making an act permissible, and so how hard it pushes φ toward permissibility.A reason's requiring weight for φ (RW φ ) is how good the reason is at making a permissible act required, and so how hard it pushes the alternative toward impermissibility.A requirement to φ is just a compound deontic verdict of permissible to φ and impermissible to do anything else.Some reason requires you to φ just when its justifying weight makes it permissible for you to φ and its requiring weight makes it impermissible to do anything else.
When we consider the requiring weight for ∼φ (RW ∼φ ), the alternative pushed toward impermissibility is φ itself.Consequently, justifying weight for φ and requiring weight for ∼φ are opposing weights, or forces.They push the same thing in opposite directions.The former pushes φ toward permissibility and the latter pushes it toward impermissibility.According to (my version of) Weight Pluralism, it is this competition that determines whether φ is (all-in) permissible: This simple version of Pluralist Permissibility can be applied only when there are exactly two options.That's enough to work with for now.We'll generalize Pluralist Permissibility in §6. 11

How to Explain Dorsey's Car Case
To explain Dorsey's Car Case, we need to combine Pluralist Permissibility with an assignment of justifying and requiring weight to each reason.Consider: We'll review the rationale for Intransitivity Assignment in due course.For now, focus on how it would help explain Dorsey's Car Case given that its numbers are on the right track.Pluralist Permissibility says that φ is permissible if and only if the justifying weight for φ is at least as great as the requiring weight of the alternative.Now we just plug in the numbers: Car Justifies Against 5Lives Vindicated: it is permissible to get the car instead of saving 5 lives, because Car's 300 (units of) justifying weight is greater than 5Lives' 250 (units of) requiring weight.5Lives Justifies Against Respect Vindicated: it is permissible to save 5 lives by beating someone up, because 5Lives' 500 justifying weight is greater than Respect's 400 requiring weight.Respect is Required Against Car Vindicated: it is impermissible to buy the car if you have to beat up someone to get it, because Car's 300 justifying weight is less than Respect's 400 requiring weight.
In short, vindicating the three verdicts of Dorsey's Car Case is just a matter of plugging the right numbers into Pluralist Permissibility.But is it really plausible that Intransitivity Assignment's numbers are right?
To accept Intransitivity Assignment, you must be willing to tolerate at least three different kinds of reasons.A traditional view is that all reasons are balanced, i.e., their justifying weight equals their requiring weight (e.g., Kagan 1985: 381).Intransitivity Assignment says that Respect is a balanced requiring reason (400 justifying and 400 requiring weight).Such reasons are demanding, because they are just as good at making actions required as they are at making them permissible.If all reasons were balanced, there would be very few moral options. 11Gert has discussed incommensurability and parity in several places (2003: 21-26, 2004, 2007: 549, 2015), but he only addresses the normative significance of small improvements in his 2015: 264-5 and especially the appendix (if you think my paper is dense, read Gert's appendix!).It is not clear how to translate what he says there into the language of justifying and requiring weight, but I think he endorses Required iff Better and rejects Pluralist Permissibility.If so, that seems to be two strikes against his view.The permissibility of acting on Car over 5Lives is a counterexample to Required iff Better ( §1.2), and the conceptual relationship of justifying weight for φ and requiring weight for ∼φ suggests Pluralist Permissibility ( §2.1).
In order to make sense of why so many choices seem to involve moral options, a trendy view holds that self-interested reasons (aka prerogatives) are merely justifying, i.e., they have justifying but no requiring weight.12Intransitivity Assignment provisionally follows this trend by making Car have 300 justifying weight but 0 requiring weight.On this view, self-interested reasons are great at making actions permissible but incapable of making them required.We'll revisit Car's requiring weight in §6.2.
Archer's (2016: 455-61) revisionary view holds that altruistic reasons are justifying heavy requiring reasons, i.e., reasons with both justifying and requiring weight, but more of the former than the latter. 13On the above assignment, the altruistic reason, 5Lives, is a justifying heavy requiring reason.It has more justifying weight (500) than requiring weight (250).So understood, altruistic reasons can make actions required, but they are twice as good at making actions permissible.As we'll see in §3, such reasons can explain moral options up to a point and then they start explaining moral requirements.I endorse Archer's revisionary view, but it hasn't gone over very well.Even Weight Pluralists criticize it (e.g., Portmore 293, nt 11; Muñoz 707).I supplement Archer's arguments for his revisionary view with two more, one in this sub-section and one in §3.I also respond to Portmore and Muñoz's objections in § §2.3 and 4, respectively.
My first argument for the revisionary view combines two points.First, since altruistic reasons are pro tanto maximizing, 5Lives must have requiring weight.If you are required to φ (e.g., costlessly save someone's life), then it is impermissible to do the alternative and only requiring weight for φ is in the business of making alternatives impermissible ( §2.1).
The second point is that, if the three verdicts of Dorsey's Car Case are true, then 5Lives' justifying weight is greater than its requiring weight.If we consider what 5Lives Justifies Against Respect, Respect is Required Against Car, and Car Justifies Against 5Lives tell us about the relative weights of reasons in that order, we get this ranking: 5Lives' JW ≥ Respect's RW > Car's JW ≥ 5Lives' RW. 14 Given that the three weight comparatives (>, <, = ) are transitive, this ranking is essential to resolving the Justification Intransitivity Paradox.It entails that 5Lives' justifying weight is weightier than its requiring weight.
Let's put the two points together.The pro tanto maximization of altruistic reasons tells us that 5Lives has requiring weight.The truth of Dorsey's Car Case tells us that 5Lives is permissive, that it has more justifying than requiring weight.It follows that 5Lives is a justifying heavy requiring reason, a reason that has both justifying and requiring weight but more of the former than the latter.15

2.3
The Failure of Justification Transitivity Recall that the Justification Intransitivity Paradox was generated by combining Dorsey's Car Case with: Justification Transitivity: If R1 makes permissible acting against R2 and R2 makes permissible acting against R3, then R1 makes permissible acting against R3.
Portmore claims that Justification Transitivity is "difficult to deny" (294). 16By denying it to resolve the Intransitivity Paradox, it may seem that I've just traded one paradox for another.The weight of reasons is (assumed to be) transitive.The permissibility of an act is determined solely by the relative weights of reasons.So why isn't the makes-it-permissible-to-act-against relation also transitive?Because this relation supervenes on two weight values (justifying weight for φ, requiring weight for ∼φ) and a reason's justifying and requiring weight aren't always equal.Consider a candidate for the worst possible card game, BestiWorst Poker.Each player is dealt two cards.Your hand beats your opponent's hand if and only if your best card is a rank at least as high as your opponent's worst card.(If the rank is tied, the hands count as beating each other.)Rank comparatives are transitive (i.e., >, <, = are all transitive when comparing ranks). 2 is the worst possible rank and ace is the best possible.The beats relation supervenes on two ranks (the ranks of your best card and your opponent's worst card).The rank of a hand's best and worst card aren't always equal.As a result, the beats relation is intransitive.Consider three possible hands.
A beats B (6 > 5), B beats C (10 > 8), but C beats A (8 > 6).The intransitivity of the makes-permissible-to-act-against relation is no more mysterious than the intransitivity of BestiWorst's beats relation.Given that weight is transitive, the failure of Justification Transitivity isn't paradoxical.It is just another reminder that Weight Pluralism is true. 17e've seen that Simple Weight Monism is too simple to explain Dorsey's Car Case.We must pay the price of complication.We can complicate it by allowing for incommensurability/parity or we can complicate it by going Pluralist.These complications are compatible.In principle, Pluralist Permissibility could accommodate parity, such that φ is permissible when the justifying weight for φ is on a par with the requiring weight for ∼φ.Yet § §1-2 suggest that, if you must choose only one of these complications, then you should choose Pluralism.For we need Pluralism to explain Dorsey's Car Case.
Case that aren't motivated by the switch to life plans.The Justification Intransitivity Paradox reemerges for life plans-or what Portmore calls maximalism-if the five people you would save are strangers, and it would be supererogatory (rather than impermissible) to save them in the more altruistic life plan (cf.296). 16More specifically, Portmore contends that Justification Transitivity follows from what it means for one reason to have more (moral) justifying weight than another (294, nt 15).This contention is false.He's arguably right that it's a conceptual truth that: if R1's JW ≥ R2's JW, then R1 makes it permissible to act against any opposing reason, R3, that R2 makes it permissible to act against (294, nt 15).But this conceptual truth is irrelevant to the intransitivity paradox.Car Justifies Against 5Lives doesn't tell us that Car's JW ≥ 5Lives' JW.It tells us that Car's JW ≥ 5Lives' RW.How to Explain Stable Moral Options Weight Pluralism can do something that neither Simple Weight Monism nor Parity Monism can do, namely explain Dorsey's Car Case.But can Weight Pluralism explain the normative significance of small improvements?Over the next two sections, I'll develop and defend the framework we need to address this question.In this section, I explain how Weight Pluralism explains moral options and provide a second argument for the claim that altruistic reasons are justifying heavy requiring reasons.

A History Lesson
Why is it so popular to hold that self-interested reasons have more justifying than requiring weight?Because it explains why moral options involving self-interested reasons tend to be so stable.Consider balanced requiring reasons.They are demanding.They are just as good at making acts required as they are at making acts permissible, they pro tanto require just as much as they pro tanto justify.Hence, they can't explain the stability of moral options for essentially the same reason that Simple Weight Monism can't explain the stability of moral options.
Car Justifies Against 5Lives is part of a moral option: it is permissible to act on Car and it is permissible to act on 5Lives.Suppose that both reasons are balanced (each reason's justifying weight is equal to its requiring weight).To get a moral option, the two reasons need to tie: Car's justifying and requiring weight must equal 5Lives' justifying and requiring weight.If they don't tie-say, we make the self-interested benefits of the car a little weightier-then Car's requiring weight is weightier than 5Lives justifying weight.Given Pluralist Permissibility, it would be impermissible to act on 5Lives.Balanced requiring reasons make moral options fragile.Small increases/decreases to one of the opposing reasons breaks the tie and so makes one of the two alternatives required. 18ntuitively, the trade-off between self-interest and altruism involves very stable moral options.You can significantly increase/decrease either reason and the moral option persists.You can double the weight of the self-interested reasons and the moral option persists.It also persists if we can 18 This paragraph is essentially a compressed version of Improvements Would Make Requirements ( §2.1), as it applies to the claim that all reasons are balanced.Let Car+ refer to the improved self-interested benefits and then we can see that improvements convert moral options into requirements: Improvements Would Make Requirements (for Balanced Requiring Reasons) P1: Car's justifying and requiring weight and 5Lives' justifying and requiring weight are equally weighty.[if all reasons are balanced, then Pluralist Permissibility allows moral options only in the case of ties] P2: Car+'s justifying and requiring weight are weightier than Car's justifying and requiring weight.In what follows, I deny P1.If reasons have more justifying than requiring weight, we can get moral options when the weights of the opposing reasons don't tie.
save 6 lives or 4 (or 3 or 2 or 1).In other words, moral options concerning the tradeoff between self-interest and altruism are very stable.
The more stable a moral option, the more that at least one of the reasons' justifying weight must outstrip its requiring weight.Intransitivity Assignment assumes that Car has 300 justifying weight and 0 requiring weight.This big gap explains why the moral option to choose Car over the altruistic option persists regardless of whether 1, 2, 3, 4, 5, or 6 lives are at stake.In short, if self-interested reasons have a lot more justifying than requiring weight, then we can explain why we have very stable moral options to choose between self-interest and altruism. 19here is a limit to the stability, however.These limits are explained by requiring weight.We are required to forgo the car to save 100 lives.The requirement kicks in when altruism's requiring weight is greater than self-interest's justifying weight (cf.Gert 2003: 24).If self-interested reasons ever require (e.g., require you to choose your life over sparing someone a mild burn), then it is possible for self-interest's requiring weight to be greater than altruism's justifying weight.(To represent this possibility, we would need to revise Intransitivity Assignment and give Car a little requiring weight.) This history lesson has two morals.First, to explain the stability of a moral option to choose between two kinds of reasons, at least one reason's justifying weight must be greater than its requiring weight.Second, to explain why the stability is finite-why the moral option can become a requirement to act on one of the reasons-you must appeal to the requiring weight of the reason that generates the requirement.

History Repeats Itself
History repeats itself when we consider the tradeoff between altruism and respecting rights.5Lives Justifies Against Respect is part of a moral option: it is permissible to save 5Lives and permissible to respect Jerry's rights.This moral option is relatively stable.It persists even if beating up Jerry saves 4 lives or 6.To explain this stability, at least one of altruism or respect's justifying weight must outstrip its requiring weight.Suppose that 5Lives and Respect are both balanced requiring reasons.Balanced requiring reasons make moral options fragile.If it is permissible to act on either reason when they are both balanced, then the reasons tie: 5Lives' justifying and requiring weight must equal Respect's justifying and requiring weight.Presumably, saving 6 lives has more justifying and requiring weight than saving 5, and so 6Lives' requiring weight is greater than Respect's justifying weight.Thus, it would be impermissible to respect Jerry's rights over saving 6 lives.20But that's implausible.Moral options involving a tradeoff between altruism and respect are relatively stable.
The first moral tells us that, to explain the relative stability of the moral option to choose between altruism and respect, we must allow at least one of the reasons to have more justifying than requiring weight.I take altruistic reasons to be more permissive than reasons of respect, as suggested by Intransitivity Assignment ( §1.1; cf.Archer 2017: 457).If altruistic reasons have twice 19 To explain moral options involving the tradeoff of self-interest and altruism, Portmore (2011: 128) makes self-interested reasons (i) non-moral reasons (ii) that are merely justifying (have moral justifying but not moral requiring weight).(i) is a red herring.Portmore stresses it because that's the only way he can tolerate (ii) which does the real work in explaining the stability of moral options.
as much justifying as requiring weight, then we can explain the relatively stable moral option to choose between altruism and respect.At 100 justifying weight and 50 requiring weight per life, 6Lives has 600 justifying weight and 300 requiring weight.It is permissible to act on 6Lives over Respect, because 6Lives' 600 justifying weight is weightier than Respect's 400 requiring weight.It is permissible to act on Respect over 6Lives, because Respect's 400 justifying weight is weightier than 6Lives' 300 requiring weight.
The second moral tells us that the limits of the stability (when the option becomes a requirement) are to be explained by the reason that generates the requirement.You are required to choose altruism over respect whenever altruism's requiring weight is greater than Respect's justifying weight.On the toy Intransitivity Assignment, we would be required to choose altruism whenever we could save more than 8 lives by beating up Jerry.But the details are negotiable.As long as there is some number of lives that would require you to beat up Jerry, altruism has some requiring weight.
The first moral tells us that, to explain the stability of the moral option to choose between altruism and respect, altruistic reasons have more justifying than requiring weight.The second moral tells us that, to explain why you can be required to choose altruism over respect, altruistic reasons have some requiring weight.The two morals together reveal that altruistic reasons are justifying heavy requiring reasons.

Altruistic Reasons: Pro Tanto Maximizing but Permissive
Old habits die hard.You've never appealed to justifying heavy requiring reasons, and you don't want to start now.You will be delighted to hear that Muñoz (2021: 707) provides an apparently devastating counterexample to the claim that altruistic reasons are justifying heavy requiring reasons.
Save1 vs Save2: Two children are trapped in a burning building and you can easily carry both.There is, therefore, no greater cost to save both rather than one.
Intuitively, it is impermissible to choose Save1 over Save2.To do so is to allow the second child to die when you could have costlessly prevented it.My account of altruistic reasons seems incompatible with this intuition.On our toy assignment, altruistic reasons have twice as much justifying as requiring weight.100 justifying weight and 50 requiring weight per life.It seems, then, that there will be the same amount of justifying weight to save one as there is requiring weight to save two (100 justifying weight for Save1 = 100 requiring weight for Save2).Given Pluralist Permissibility, it follows that it is permissible to Save1 over Save2!My account seems doomed.But it's not.At least not if altruistic reasons are comparative. 21omparativism about reasons is a thesis about how reasons (better: grounds 22 ) are individuated: in a pairwise competition between φ and a given alternative, a reason for φ is a way that φ 21 For a defense of comparativism, see Snedegar (2017, 2018).Muñoz's (2021) discussion of the Justification Intransitivity Paradox also endorses it at least for self-interested reasons. 22The difference between grounds and reasons is important for Dancy-style holism (as defended in Dancy 2004).A ground R for φ is just a reason for φ in some context or another (cf.Bader 2016: 28ff).That φ would keep your promise to so-and-so is a ground for φ.It is normally also a reason for φ; however, if the promise was coerced, the ground may not be a reason, i.e., it may lack any weight whatsoever.If so, the coercion disables the ground's weight.
is better than that alterative.Consider a choice between Save0 and Save1.The reason to choose Save1 is not, strictly speaking, the child's life you save.It's something like that you save one more child in Save1 than Save0.Since reasons for φ are ways that φ is better than an alternative, you can change the reasons/grounds for φ simply by changing the alternative.In Save1 vs Save2, there is no way that Save1 is better than Save2.There is, therefore, no reason at all to choose Save1 over Save2.
Altruistic reasons are pro tanto maximizing, because they are comparative and have at least a little requiring weight.There is no reason at all to choose Save1 over Save2 (0 JW Save1 and 0 RW Save1 ).In contrast, there is a reason to choose Save2 over Save1, namely the additional life you would save (100 JW Save2 and 50 RW Save2 ).Pluralist Permissibility tells us that it is permissible to Save1 exactly when the justifying weight for Save1 is at least as great as the requiring weight for Save2.But it's not.0 justifying weight for Save1 is less than 50 requiring weight for Save2, so Save1 is impermissible after all.
In effect, we have identified and resolved another puzzle.Altruistic reasons are pro tanto maximizing.When the only relevant consideration is the number of lives saved, we are required to save the most lives that we can.We also seemed to show in § §2-3 that altruistic reasons are permissive, that they have more justifying than requiring weight.Pro tanto maximization and permissiveness seem to conflict, as illustrated by Muñoz's Save1 vs Save2 objection.The puzzle is to explain how some reason can be both pro tanto maximizing and permissive.My resolution appeals to comparativism.
Altruism can be pro tanto maximizing and permissive, because its permissiveness (the extent to which its justifying weight outstrips its requiring weight) matters only when you have to trade altruism off against some other morally relevant consideration.In such cases, you must weigh the altruistic reason's justifying weight against an opposing reason's requiring weight.Yet comparativism tells us that Save1 vs Save2 is not a case in which we weigh opposing reasons.Since there is no cost to saving both, there is no reason for Save1 that opposes the requiring reason for Save2.When the requiring reason for Save2 is unopposed, you are required to choose Save2.Altruistic reasons can be both pro tanto maximizing and permissive, because their permissiveness only comes into play when there is a cost, only when there is some opposing reason. 23on't misunderstand.Comparativism doesn't entail that a reason's justifying weight can vary as you vary the alternative.Comparativism doesn't entail that there is a single reason for Save1, such that the reason has more justifying weight against Save0 than Save2.What varies is not the weight of a given reason but which, if any, reason applies.When Save1 competes with Save0, there is a reason for Save1 (the additional life you save).When Save1 competes with Save2, there isn't a reason for Save1 at all.It is not the least bit surprising that there is more weight for Save1 when there is a reason for it (as against Save0) than when there is no reason for it at all (as against Save2).Comparativism entails that the weight for φ can vary as you vary the alternative only because 23 This claim leaves open that some altruistic reasons compete with one another.This openness allows my account to be neutral on (a certain understanding of) the normative distinction between persons.Given the normative distinction, it matters not only how many people die, but who dies.Suppose that you must choose between saving Oscar or saving both Bert and Ernie.The normative distinction between persons arguably entails that there is a cost to saving more lives, namely that Oscar dies.The extent to which altruism's justifying weight outstrips its requiring weight is now relevant.Hence, my account predicts that choices between saving different numbers of different people (e.g., Oscar vs Bert&Ernie) are more permissive than choices between saving different numbers of the same people (e.g., Save1 vs Save2).At least, that's the prediction given the (above understanding of) the normative distinction between persons.
which reason applies (the ways and extent to which φ is better than the alternative) can vary as you vary the alternative. 24omparativism is a controversial position.If you reject it, think of this paper as an opportunity to scout the competition.It is easy to underestimate how well Weight Pluralism and comparativism interact.Cullity and Rabinowicz claim that there is no way for Weight Pluralism to explain the normative significance of small improvements (without letting the same reason have different weight values in different contexts).By appealing to comparativism, I will show that there is a way after all.Maybe the noncomparativist about reasons can do even better than the comparativist.But comparativism gives us progress, and progress is progress.

5
The Normative Significance of Small Improvements

A Pluralist Explanation of the Normative Significance of Small Improvements
This sub-section section applies the lessons of § §3-4 to explain the normative significance of small improvements.I focus on small improvement cases that build off of Dorsey's Car Case (altruism vs respect in this sub-section and then altruism vs self-interest thereafter).This focus will highlight the continuity between my explanations of Dorsey's Car Case and the normative significance of small improvements.It will be easy enough to generalize my resolution to whatever small improvement case your heart desires.
To explain the normative significance of small improvements we must explain both the pro tanto maximization of certain reasons and the stability of moral options involving those reasons.The puzzle begins with two options that involve acting on different morally relevant considerations, such as altruistically saving five lives (A) and respecting Jerry's right's not to be beaten up (R).(Again, to prevent confusion about the nomenclature, 'A' and 'R' refer to options.5Lives is the reason to choose A and Respect is the reason to choose R.) Then consider modestly improved versions of these considerations, such as saving those five lives and an additional life (A+) and respecting both Jerry's right not to be beaten up and his right not to be insulted (R+).Pro tanto maximization is illustrated by the requirements to choose A+ over A (saving 6 rather than 5 lives) and R+ over R (respecting both rights instead of just one).When we aren't trading morally relevant considerations off against one another, small improvements tend to generate requirements.
Nonetheless, we have the moral option to choose between A and R, between saving 5 or respecting Jerry's right not to be beaten up.The stability of moral options is illustrated by the moral option persisting after we improve either reason.The moral option persists if we have to choose between A and R+, and it persists if we have to choose between A+ and R. When we are trading morally relevant considerations off against one another (e.g., altruism vs respect), small improvements tend 24 Muñoz falsely claims that his resolution of Dorsey's Car Case involves the "same sort of comparativity" as comparativism (708).His solution relies on selective weight (the very same reason, Car, has more justifying weight when it opposes altruistic reasons than respect reasons).That is a more controversial claim than comparativism (vary the alternative and you can vary which reason applies).To be sure, comparativism can resolve Kamm's and Parfit's version of the Justification Intransitivity Paradox (Muñoz 2021: 706-7; Tucker 2022:  §6).For, given comparativism, the relevant self-interested reasons in their cases vary as you vary the alternative.Dorsey's version is harder to resolve because the same self-interested reason, Car, applies in both Car Justifies Against 5Lives and Respect is Required Against Car.This feature of the case partly explains why we are forced to make 5Lives a justifying heavy requiring reason in order to explain the three verdicts of Dorsey's Car Case ( §2.2).
to not generate requirements.They tend to illustrate the stability of moral options.Whether the improvement, A+ and R+, generates a requirement seems to depend on whether we are trading off morally relevant considerations.This is puzzling, but it's a puzzle that we've already solved.
Per §4, we explain pro tanto maximization by appealing to comparativism and a little requiring weight.It is impermissible to choose A over A+ (save 5 rather than 6), because there is a requiring reason to save 6 that is unopposed by any other reason.Comparativism tells us that reasons for an option are ways that an option is better than some alternative.The reason to save 6 over 5 is the additional life you save.We stipulated that the reason has 100 justifying and 50 requiring weight.But there is no reason to choose saving 5 over saving 6, because there is no way that saving 5 is better than saving all 6.Pluralist Permissibility tells us that the im/permissibility of φ is determined by the competition between the justifying weight for φ and the requiring weight for the alternative.Since there is 0 justifying weight for saving 5 and 50 requiring weight for saving 6, it is impermissible to save 5. Parallel points explain why it is impermissible to choose R over R+, why it is impermissible to costlessly respect only one instead of both rights.
Per §3.2, we explain finitely stable moral options that involve tradeoffs between altruism and other morally relevant considerations by making altruistic reasons justifying heavy requiring reasons.In each of the relevant pairwise choices (A vs R, A vs R+, A+ vs R), there is a way that each option is better than the other.Comparativism tells us, then, that there is a reason to choose each option over the other.In contrast to A vs A+ and R vs R+, these altruism vs respect choices are cases in which we weigh opposing reasons.Consider the choice to act on 5Lives or Respect (the choice between A vs R).It is a moral option, because each reason's justifying weight is at least as weighty as the other's requiring weight.It is permissible to act on 5Lives rather than Respect, because 5Lives's 500 justifying weight outweighs Respect's 400 requiring weight.It is permissible to act on Respect rather than 5Lives, because Respect's 400 justifying weight outweighs 5Lives' 250 requiring weight.
The moral option persists regardless of which reason is improved, because each reason's justifying weight would still be at least as great as the other's requiring weight.Consider the choice between 6Lives and Respect (A+ and R).It is permissible to choose 6Lives over Respect, because 6Lives' 600 justifying weight outweighs Respect's 400 requiring weight.It is permissible to act on Respect over 6Lives, because Respect's 400 justifying weight outweighs 6Lives 300 requiring weight.Parallel points explain why the moral option persists in the choice between A and R+, why we can permissibly save 5 lives or permissibly respect both Jerry's right not to be beaten up and his right not to be insulted.
Altruism can be pro tanto maximizing and generate stable moral options, because its permissiveness (the extent to which its justifying weight outstrips its requiring weight) matters only when you have to trade altruism off against some other morally relevant consideration.In such cases, you must weigh the altruistic reason's justifying weight against an opposing reason's requiring weight.In the choice between saving 5 and 6, there is no reason that opposes saving 6 and, when the requiring reason for saving 6 is unopposed, you are required to save 6. Altruistic reasons can be both pro tanto maximizing and permissive, because their permissiveness only comes into play when there is a cost, only when there is some opposing reason.
That's it.That's how the Weight Pluralist can explain the normative significance of small improvements.We did not need parity or any other fourth comparative.Nor did we need to deny the transitivity of weight comparatives.We explained these phenomena by appealing to Pluralist Permissibility, comparativism about reasons, and the revisionary idea that altruistic reasons are justifying heavy requiring reasons.All that's left now is to tie up two loose ends.
Cullity and Rabinowicz argue that Weight Pluralism gets mired in contradiction when it tries to explain the normative significance of small improvements.The first loose end is to articulate their objection and explain where it goes wrong ( §5.3).The second is to generalize my explanation to cases in which we have more than two options at a time ( §6).I explained the normative significance given pairwise choices.But how do we explain it when we have all four options (e.g., A, A+, R and R+) at the same time?

No Contradiction, Just Ambiguity
Let A be the altruistic act of saving 5 lives and A+ saving those same five people and an additional person.Let S be the self-interested act of getting the car.Let S+ be getting the car with a free upgrade that you would enjoy (maybe it's those warming seats that you've always wanted for the cold morning drives or the better sound system that will make singing along all the more fun).
Intuitively, it is impermissible to choose A over A+ and S over S+, while it is permissible to choose A over S+ and S over A+.Cullity (2018: 431, nt 9) contends that "There is no consistent assignment of requiring and justifying strengths to the reasons favoring these options that can deliver this result". 25In a slightly different context, Rabinowicz (2008: 32-6, 2012: 142) gives essentially the same objection to Gert (2004). 26he objection can seem decisive.It seems that we can generate a contradiction in two simple inferences.The first inference combines the permissible tradeoff of A over S+ with the impermissible choice of S over S+ (that doesn't involve any tradeoff between different morally relevant considerations).

Inference 1
Tradeoff 1: The justifying weight for A is at least as weighty as the requiring weight for S+.The second inference combines the permissible tradeoff of S over A+ with the impermissible choice of A over A+. 25 Cullity does allow that the Pluralist could appeal to holism (the idea that a reason's weight can differ in different contexts) to provide a consistent weighting.But I show that the Pluralist can provide consistent weightings without holism.See nt 23 above for an explanation of holism. 26More precisely, Rabinowicz's target is Gert's (2004) account of betterness.Roughly, Gert (2004b: 505) holds that X is better than Y iff it is impermissible to choose Y over X or, equivalently on Gert's 2004 view, JW Y < RW X .Rabinowicz complains: It is easy to show that one cannot assign intervals [i.e., JW, RW] to the four items in such a way that this intuitive structure of betterness relationships is preserved.If we set up the intervals so as to make X + better than X [JW X < RW X+ ] and Y + better than Y [JW Y < RW Y+ ], then either X + will come out as better than Y or Y + will come out better than X. ( 2012 Implications 1 and 2 seem contradictory.We are assuming that weight comparatives are transitive, so the two inferences are valid.It seems, then, that the premises of the two inferences (and so the four intuitive deontic verdicts) are jointly incompatible.Pluralism seems to founder on the normative significance of small improvements.Why doesn't it?
Given comparativism, Implication 1 and 2 are compatible.The implications are contradictory only if the assignments of justifying weight for A and S remain constant through each of the relevant pairwise competitions.They don't.Comparativism makes reasons-and so the weight for a given option-relative to the alternative.There is no such thing as the justifying weight for A full stop.There is the justifying weight for A in A vs A+.There is also the justifying weight for A in A vs S+.Different pairwise competitions involve the same reasons only insofar as the same differences/comparisons apply.If you vary whether or how A is better than the alternative, then you vary the reasons (and weight) for A in those pairwise competitions.
Implication 1 is that the justifying weight for A is weightier than the justifying weight for S. Implication 2 is that the justifying weight for S is weightier than the justifying weight for A. Given the transitivity of weightier than, it follows that the justifying weight for A in Implication 1 is weightier than the justifying weight for A in Implication 2. How could that be?In Tradeoff 1 and Implication 1, 'the justifying weight for A' refers to the justifying weight of 5Lives, the reason you have to choose A over S+ (i.e., the reason you have to choose saving 5 people over the upgraded car).In No Tradeoff 2 and Implication 2, 'justifying weight for A' refers to the justifying weight of no reason at all to choose A over A+ (saving 5 people vs saving those same 5 people and 1 additional person).So, of course, 5Lives' 500 justifying weight for A in Implication 1 will be greater than no reason's 0 justifying weight for A in Implication 2. There is no contradiction in the two implications.There is just ambiguity.Cullity and Rabinowicz's objection fails to pose any threat to my Weight Pluralist explanation of the normative significance of small improvements.

The Generalization
One limitation of the account so far is that we've considered a simple version of Pluralist Permissibility that applies only to cases in which you have two options. 27If this is all that I could manage, I would be disappointed.You might see a normally composed, fully functioning adult break into tears.Thankfully, the hard work is behind us and we'll avoid any awkward emotional displays.The generalization is straightforward.
Recall that Simple Pairwise Permissibility holds that φ is permissible just when justifying weight for φ is at least as great as the requiring weight for ∼φ.In other words, when the only options are φ and ∼φ, φ is made permissible by winning one competition with one alternative.It's also true, though, that φ is made permissible by winning a tournament, a pairwise competition with every alternative.This latter idea is the key to generalizing simple Pluralist Permissibility to cases in which there are more than two options.This gives us the perfectly general: (Generalized) Pluralist Permissibility: φ is permissible if and only if, in each pairwise competition with each alternative A, JW φ ≥ RW A .φ is impermissible if and only if, for some alternative A, JW φ < RW A . 28 compare φ with each alternative, and if φ wins every comparison, then φ is permissible.The basic idea is that it is permissible to choose an option exactly when you can justify it over each alternative.
Above we considered the normative significance of small improvements one pairwise competition at a time.This process allowed us to determine the deontic verdict of each act when it competed with a single alternative.When we have all four options at once (A, A+, S, S+), deontic status still boils down to what happens in these same pairwise competitions.
To determine which options are permissible, we need to see which options win all of their pairwise competitions.Consider altruistic A (saving 5 lives).Whether A is permissible is determined by a tournament that includes three pairwise competitions: A vs S, A vs S+, and A vs A+.Our discussion in §5 revealed that A wins both its competition with S (getting the car) and its competition with S+ (getting the upgraded car).In other words, the justifying weight for saving 5 lives is greater than the requiring weight for getting the car or the upgraded car.Yet we saw that A loses its competition with A+ (saving the same 5 people and an additional person).The justifying weight for A is outweighed by the requiring weight for A+.Losing once is losing too much to be permissible.Since A fails to win its pairwise competition with A+, A is impermissible when we have all four options at the same time.
Self-interested S (getting the car) is also impermissible when we have all four options at the same time.S loses its competition with S+ (getting the upgraded car).The justifying weight for S is outweighed by the requiring weight for S+.
On the other hand, A+ and S+ are both permissible when you have all four verdicts at the same time.For they win their respective tournaments: in each pairwise competition, the justifying weight for A+ (S+) is at least as great as the requiring weight for each alternative.If you won't take my word for it, consult Small Improvement Assignments in §6.2.A+ features in competitions 1, 4, and 6.S+ features in competitions 2, 3, and 6.
When you have all four options at once, (Generalized) Pluralist Permissibility tells us that A+ and S+ are permissible but A and S are impermissible.I take it that these are the intuitive deontic verdicts.Hence, Weight Pluralism has no trouble explaining the normative significance of small improvements even when we have all four options at once.The rest of this section is like an appendix.It may not be useful to the average reader, and you should feel free to skip it.I plug in some numbers to illustrate in more detail how I resolve the normative significance of small improvements for altruism and self-interest.You will then be able to verify that, when we have all four options at the same time (A, A+, S, S+), only A+ and S+ win their respective tournaments.

6.2
The Full Details We are working with the idea that there is 100 justifying weight and 50 requiring weight per additional life saved.A is the act of saving 5 lives, and A+ is the act of saving those same 5 lives and 1 additional life.Self-interested S and S+ don't save any lives.Comparativism tells us that the altruistic reason for φ, if any, is the additional altruistic benefit of φ over a specific alternative.Here, then, are the relevant altruistic reasons and when they apply: Before addressing the self-interested reasons, we need to have a heart-to-heart.Typical Weight Pluralists assume that self-interested reasons are merely justifying (have justifying but no requiring weight).I provisionally went along with this assumption in §2.2, because its proponents are cool people and I wanted them to like me.Yet sometimes even the best of friends must tell each other the truth.The merely justifying reason view is correct insofar as self-interested reasons have less requiring weight than altruistic reasons.But they still have some.Otherwise, it would be morally permissible to choose trivial altruistic benefits (which have some justifying weight) at the cost of traumatizing yourself (which would have no requiring weight). 29Let's say that selfinterested reasons have 10 times less requiring than justifying weight.This little bit of requiring weight will be important for explaining why it is impermissible to choose S over S+.Ok, now to the self-interested reasons.
S is the act of getting the car and S+ is the act of getting the car with the free upgrade.The altruistic A and A+ don't involve any self-interested benefits.Here are the relevant self-interested reasons and when they apply: No Reason (0 JW, 0 RW): there is no reason for S over S+.Upgrade (10 JW, 1 RW): the self-interested benefits of the upgrade that you get in S+ but not S. Car (300 JW, 30 RW): the self-interested benefits of the car that you get in S but not A or A+.Car+ (310 JW, 31 RW): the self-interested benefits of the upgraded car that you get in S+ but not in A or A+.
To tell whether φ is permissible, Pluralist Permissibility tells us that we need to compare the justifying weight of the reason for φ with the requiring weight of the reason for the alternative.The following table identifies the relevant justifying and requiring weight values for each pairwise competition.I've shaded the part of the table that vindicates the deontic verdicts that comprise Cullity and Rabinowicz's objection in §5.The rest of the table contains all the other information that matters when we have all four options at the same time.

Scoring Recap
At first glance, it seemed that Parity Monism and Weight Pluralism were equally matched in their ability to explain moral options and that only Parity Monism could explain the normative significance of small improvements.Weight Pluralism's approach to small improvements seemed mired in contradiction.
The second glance was the heartbreaker for Parity Monism.You have the moral option to act on Car or 5Lives.Parity Monism would explain this moral option by claiming that acting on Car is on a par with acting on 5Lives.But that's false.The value of saving five lives is far greater than the value of getting the car's self-interested benefits.This is not an isolated phenomenon.Paradigmatic instances of supererogation often involve a moral option to choose worse self-interested benefits over better altruistic benefits.Parity (for value or weight) is useless in explaining why we have moral options in such cases.
Resistance on the relative value of the self-interested and altruistic benefits is futile.If the Parity Monist insists that acting on Car is on a par with acting on 5Lives, then they make it mysterious how acting on 5Lives could be supererogatory.Supererogatory acts aren't just morally optional; they are also better than some permissible alternative.Yet, if acting on Car is on a par with acting on 5Lives, then neither is better than the other.Neither would be supererogatory.
Weight Pluralists, then, have the advantage when it comes to explaining moral options.Since they can allow that self-interested reasons have more justifying than requiring weight, they have no problem explaining how you can have a moral option to choose worse self-interested benefits over better altruistic benefits.They also are at no disadvantage when it comes to explaining the normative significance of small improvements.
To explain the normative significance of small improvements, a normative theory must be able to combine two things.The first is an explanation of pro tanto maximization.When the only relevant consideration is how many lives you save-when there are no tradeoffs-it is impermissible to choose A over A+ (or Save1 over Save2).The second is an explanation of stable moral options when trading off morally relevant considerations against each other.There is a moral option to choose A over S and that moral option remains when we choose between A and S+.
This paper defended a revisionary Weight Pluralism that can be distilled into three theses.The first is a general account of how reasons determine deontic status: Pluralist Permissibility: φ is permissible if and only if, for each pairwise competition with each alternative A, JW φ ≥ RW A .
The paper's approach to weighting altruistic reasons is revisionary, because it bucks the near universal assumption that altruistic reasons are balanced (their justifying weight = their requiring weight).Instead, it holds: Revisionary Weighting: many reasons, including altruistic and self-interested ones, are justifying heavy requiring reasons, i.e., they have both justifying and requiring weight but more of the former than the latter.
These first two theses explain why altruistic reasons tend to generate stable moral options.We get a moral option because, e.g., 5Lives and Respect each have more justifying weight than the other has requiring weight.The moral option is stable because, when 4 or 6 lives are at stake, it is still true that the altruistic reason and respect reason each have more justifying weight than the other has requiring weight.
The permissive nature of altruistic reasons is compatible with pro tanto maximization, because: Comparativism about reasons: a reason for φ is way that φ is better than some specific alternative.
When you are choosing between A and A+, there is no way that A is better than A+, and so no reason to choose A over A+.As long as altruistic reasons have a little requiring weight, as per Revisionary Weighting, you will be required to choose the best.The permissive nature of altruistic reasons only matters when there is a cost, only when the altruistic reason must be weighed against some other opposing reason.These three theses are a terrific trio.Together they explain Dorsey's Car Case; the stability of moral options involving self-interested or altruistic reasons; how altruistic reasons are pro tanto maximizing; and the normative significance of small improvements.These achievements carried it to victory over Parity Monism, which had its own list of impressive accomplishments.But let's not get carried away in celebration.Beating one competitor (or two, if you count Simple Weight Monism) hardly makes this trio Olympic gold medalists.Yes, a new bar has been set.Any Weight Improvements Would Make Requirements (for Simple Weight Monism) P1: The reasons for A and R are equally weighty.P2: The reason for R+ is weightier than the reason for R. P3: Any reason that is weightier than R is weightier than A. [P1, substitutability principle to be introduced below] C1: Therefore, the reason for R+ is weightier than the reason for A. [P1, P2, P3] P4: If the reason for R+ is weightier than the reason for A, then you are required to choose R+. [Monist Requirement] C2: Therefore, you are required to choose R+ over A. [C1, P4] or by appealing to a third weight of reasons, commending weight (Little & Macnamara 2020; Horgan & Timmons 2010: §VI.B also appeal to a third value but use different terminology).
[premise] P3: Any reason with more justifying and requiring weight than Car also has more justifying and requiring weight than 5Lives.[P1, Substitutability of Equality] C1: Therefore, Car+'s justifying and requiring weight are weightier than 5Lives' justifying and requiring weight.[P1, P2, P3] P4: If Car+'s justifying and requiring weight is weightier than 5Lives's justifying and requiring weight, then you are required to choose Car+ over 5Lives.[Pluralist Permissibility] C2: Therefore, you are required to choose Car+ over 5Lives.[C1, P4] [It is permissible to choose A over S+ & Pluralist Permissibility.]No Tradeoff 1: The requiring weight for S+ is greater than the justifying weight for S. [It is impermissible to choose S over S+ & Pluralist Permissibility.]Implication 1: JW A > JW S , i.e., the justifying weight for A is greater than the justifying weight for S. [Tradeoff 1, No Tradeoff 1] The justifying weight for S is at least as great as the requiring weight for A+.[It is permissible to choose S over A+ & Pluralist Permissibility.]No Tradeoff 2: The requiring weight for A+ is greater than the justifying weight for A. [It is impermissible to choose A over A+ & Pluralist Permissibility.]Implication 2: JW S > JW A , i.e., the justifying weight for S is greater than the justifying weight for A. [Tradeoff 2 and No Tradeoff 2]