Essentially Comparative Prioritarianism

Introduction: A Problem for Prioritarianism

A prevalent version of prioritarianism states that benefitting people matters more the worse off those people are, the more of those people there are, and the greater the benefits in question1[1]. This note advances a noble version of prioritarianism by addressing an underexplored challenge: how would prioritarianism handle the phenomenon that Ruth Chang calls “parity”?2[2].

Let me present the challenge3. Imagine the doppelgangers Jerry and Jerry+, whose lives so far have been as good. They are now suffering almost identically severe pains although Jerry+’s pains are slightly more intense than Jerry’s. Arguably, Jerry+ is worse off than Jerry. This is due to three facts

• The degree to which one is in pain partly determines how good one’s life is,
• Jerry+’s pains are more severe than Jerry’s, and,
• All other things are equal. Suppose a distributor has only one indivisible dose of a Essay painkiller available, which would have an identical effect on the Jerrys. Then, on the face of it, prioritarianism implies that benefitting Jerry+ matters more than benefitting Jerry. Or, as I shall sometime put it, prioritarianism resolves “the {Jerry, Jerry+} dilemma” by prioritizing Jerry+.

Enter the “Parity Pattern”. Bernard is an amputee who chronically suffers from severe phantom pain. His pains are far worse than the pains from which the Jerrys suffer. However, despite his condition, he is an accomplished and wealthy Olympic sprinter and his life is far more interesting and rich than the lives of the Jerrys. According to Chang, we can assume that the following is true:

A Parity Pattern (for “…better off…”): (1.) Bernard is neither better off nor worse off than Jerry (2.) Bernard is neither better off nor worse off than Jerry+. (3.) Jerry+ is worse off than Jerry. (4.) The pairs < Bernard, Jerry > and < Bernard, Jerry+ > are not border line cases of “…better off …”.

Who, then, should get the indivisible dose of the pain- killer, Bernard or Jerry? Bernard or Jerry+? That is, how would prioritarianism resolve the {Bernard, Jerry} and the {Bernard, Jerry+} dilemmas? On the face of it, prioritarianism says of the {Bernard, Jerry} dilemma that neither should be prioritized; the same is true of the {Jerry+, Bernard} dilemma.

But, if no-priority is the resolution of both the {Jerry+, Bernard} and the {Bernard, Jerry} dilemmas, then, on the face of it, it follows that no-priority should be the resolution of the {Jerry+, Jerry} dilemma, contrary to prioritarianism’s initial verdict. Call this, the “Parity Puzzle”.

This essay presents a model—hereinafter, the “Parity Model”—that suggests an explanation for why, even if no- priority is the resolution of both the {Jerry+, Bernard} and the {Bernard, Jerry} dilemmas, no-priority is not the resolution of the {Jerry+, Jerry} dilemma. The essay proceeds as follows. Section 2 sketches the Parity Model, which includes a metaphysics and a semantics of complex comparatives, to which I appeal in Section 3 in advancing a version prioritarianism that resolves the Parity Puzzle.

The Parity Model: the Metaphysics and the Semantics of Parity

The Parity Model offers a metaphysics and a semantics of complex comparatives, which I will illustrate by the comparative “…balder than…”. (I will present only the basic ideas of the model; in a previous paper I offered the account on which I rely; a similar model was offered by Andreou)4[3, 4].

Imagine an individual, Larry, who lost one hair an hour ago; call Larry of today “Larry+”, and consider the true judgement “Larry+ is precisely one hair balder than Larry” (that is “Larry is now precisely one hair balder than he was yesterday”). The first intuition that the Model aims to explain is that such true precise baldness comparisons are rare. Consider, for example, the Barry/Larry comparison. Barry is doing better than Larry with respect to one underlying dimension that constitutes baldness i.e., Barry has 100 hairs more on his scalp; Larry is doing better than Barry with respect to another dimension, i.e., the width of the totally exposed patches on Larry’s scalp is smaller. The first intuition is that, the baldness of Larry and the baldness of Barry cannot be compared by number of hairs. Although Larry has 100 hairs less than Barry, he is not 100 hairs balder than Barry. The second related linguistic intuition that the Parity Model aims to capture is that the truth of “Larry+ is precisely one hair balder than Larry” partly follows from the fact the baldness of Larry and the baldness of Larry+ do not differ in any other underlying dimension. That is, the precise comparison is true since all other things are equal. Most comparisons do not proceed by hairs counting since mostly, other things are not equal. The Parity Model is based on a third linguistic intuition: the advantages that Larry and

4 See, Yitzhak Benbaji, “Parity, Intransitivity, and a Context-Sensitive Analysis of Gradability” Australasian Journal of Philosophy 87 (2009) 313- 338, and Chrisoula Andreou, “Parity, Comparability and Choice” Journal of Philosophy CXII (2015): 5-22.

Barry have over each other can be cancelled out by each other. Indeed, Larry and Barry might be (roughly/crudely) as bald as one another.

In explaining these three intuitions the model advances one basic idea: different baldness comparisons might proceed along different, more or less, precise scales. In the Jerry/Jerry+ case the precise comparison is appropriate since their baldness is equal in every other contributing factor. In contrast, while they cannot be compared by a number of hairs degrees, the baldness of Larry and the baldness of Barry can be compared by comprehensive degrees, determined by the number of hairs they have on their scalp, their distribution, and, supposedly, many other factors, including the relative weight of each of them. Thus, according to the Parity Model, a fully articulated baldness comparison would specify on which scale the comparison in question proceeds. Assume that:

Bald(Larry, d1)” means that the degree to which Larry is bald is d1.

Then, according to the Parity Model,

“Larry+ is one hair balder than Larry” means that there are degrees, d1 and d2, on the numbers of hairs scale that measures baldness, such that Bald(Larry, d1) & Bald(Larry+ ,‏ d2) & d1-d2=1.

And “Larry is as bald as Barry” means that there are degrees (D1 and D2) on the comprehensive scale that measures baldness, such that Bald(Larry, D1) & Bald(Barry ,‏ D2) & D1=D2.

As its name suggests, the model is designed to capture a further fourth intuition, namely that “…balder than…” can satisfy the Parity Pattern. In order to allow a parity pattern, the Parity Model attributes to comprehensive degrees of baldness two features. Namely: comprehensive degrees are “categorical” and “crude”. These characterizations come from David Papineau’s analysis of our color concepts and his reinterpretation of the fact that “humans are capable of well over a million different conscious visual responses to colored surfaces.” Papineau’s crucial observation is that “the detection of color differences between adjacent surfaces [colored by, say, orange28 and orange29] does not derive from prior [independent] responses to each surface”5[5]. Instead, he suggests that our conscious color experience

5 See, David Papineau, “Can We Really See a Million Colors?”, in in Paul Coates and Sam Coleman, (eds.), Phenomenal Qualities (Oxford University Press, 2015), 274-297. The analogy I advance is similar to Andreou’s. I heavily rely on her presentation of Papineau’s argument in her “Parity, Comparability and Choice” p. 11-13.

is the joint product of two different cognitive capacities. We have (first) a relatively modest number of categorical concepts of colors: …very light orange,…orange….very light red, … light red… red, … deep red. These concepts allow us to categorize a surface as of a certain color. These color concepts are crude in a straightforward sense. Orange28 and orange29 are both instances of very light orange. Indeed, viewing them separately, a human observer would truly say of orange28 and orange29 that they are identical, since both are instances of very light orange6[4]. Still, when she views them simultaneously, a human observer can distinguish between orange28 and orange29, by what Papineau calls, “relational” color concept.

The Parity Model analogizes between the categorical, crude color concepts and the comprehensive degrees of baldness (and other complex comparatives). Larry and Barry are equally bald in virtue of the fact that they are both (say) very bald, rather than completely bald on the one hand, or mildly bald, on the other. That is, they are equally bald in virtue of the fact that they are bald to the same categorical degree. In contrast, the baldness of Larry‏+ and the baldness of Larry resemble two adjacent surfaces colored by orange28 and orange29 viewed simultaneously. While there is no categorical baldness concept by which the baldness of Larry and the baldness of Larry+ can to be distinguished, (crudely, Larry is as bald as Larry+; they are both, say, very bald) the fact that all other things are equal allows a relational comparison between them.

The relation between crude and precise degrees generates a vagueness free Parity Pattern: The Parity Pattern (for “…balder than”…”): (1) determinately, Larry‏+ is balder than Larry & (2) determinately, Barry is as bald as Larry & (3) determinately, Larry‏+ is as bald as Barry.

The first (Larry/Larry+) comparison proceeds by precise degrees while the next two (the Larry/Barry and the Larry+/Barry) proceed by categorical degrees: [(1.)]. “Larry+ is balder than Larry” means that there are degrees d1 and d2 on the number of hairs scale that measures baldness such that Bald(Larry, d1) & Bald(Larry+ ,‏ d2) & d2>d1.

[(2.) & (3.)] “Barry is as bald as Larry & Larry‏+ is as bald as Barry” means that there are degrees D1, D2, and D3 on the comprehensive scale that measures baldness such that Bald(Larry, D1) & Bald(Larry+, D2) & Bald(Barry ,‏D3) & D1=D2=D3.

6 Andreou, “Parity, Comparability and Choice”, p. 11.

Prioritarianism by Categorical, Crude Degrees vs. Essentially Comparative Prioritarianism Let us turn to the case by which I exemplified the Parity Puzzle. Who should get an indivisible painkiller, Jerry or Jerry+? Initially, prioritarianism seems to suggest that since other things are equal and Jerry+’s pains are slightly more intense than Jerry’s, Jerry+ is worse off than Jerry, and therefore benefitting Jerry+ is more urgent than benefitting Jerry. Consider, however {Bernard, Jerry} and the {Bernard, Jerry+} dilemmas, where Jerry is crudely as well off as Bernard, and Jerry+ is crudely as well off as Bernard. The Parity Puzzle stresses that, since no-priority is the resolution of both the {Jerry+, Bernard} and the {Bernard, Jerry} dilemmas, no- priority should be the resolution of the {Jerry+, Jerry} dilemma.

Based on two implications of it, the Parity Model suggests a simple solution to this puzzle. First, like baldness, wellbeing is a complex dimension. How good one’s life is, is determined by factors like, how painful, interesting, moral, social, etc., one’s life is. The Parity Model implies, then, that Jerry and Bernard are crudely, as well off, and so are Jerry+ and Bernard. It follows that Jerry and Jerry+ are crudely, as well off:

There are degrees D1, D2 and D3 on the comprehensive scale that measures wellbeing, such that Welfare(Jerry, D1) & Welfare(Jerry+, D2) & Welfare(Bernard, D3) & D1=D2=D3.

However, according to the model, there is no contradiction between this statement and the statement that Jerry+ is worse off than Jerry. This is because, the former comparison proceeds by categorical degrees while the latter proceeds by measuring the intensity of the pains that Jerry and Jerry+ suffer.

There are degrees d1 and d2 on the pain-intensity scale that measures wellbeing, such that Welfare(Jerry, d1) & Welfare(Jerry+ ,‏d2) & d1>d2.

A second implication of the Parity Model is that the prioritarianism formula—viz., benefitting people matters more the worse these people are—is underspecified. The Jerry/Jerry+ comparison might proceed by the intensity of their pains as well as by categorical, crude degrees. Hence, in addressing the {Jerry, Jerry+} dilemma, prioritarianism should answer the “priority by which degrees” question. According to one possible reading of prioritarianism— “crude prioritarianism”—benefitting one person (x) matters more than benefitting another (y) if and only if the categorical degree (D1) to which x’s life is good is smaller than the categorical degree (D2) to which y’s life is good. In other words.

Crude Prioritarianism: benefitting people matters more, the more categories of wellbeing there are above the wellbeing level at which these people are.

In effect, crude prioritarianism employs the Parity Model in order to deny that, in the {Jerry+, Jerry} dilemma, benefitting Jerry+ matters more than benefitting Jerry. Since their lives are crudely as good, benefitting Jerry+ is as urgent as benefitting Jerry. Indeed for crude prioritarianism, no priority is the resolution of the {Jerry+, Jerry} dilemma.

But crude prioritarianism seems implausible. Interestingly the Priority Model allows for a different answer to the priority by which degrees question. To use a phrase coined by Larry Temkin, this answer construes prioritarianism as essentially comparative view7[6]. It reads as follows:

Essentially Comparative Prioritarianism: In a {P1, P2… Pn} dilemma, benefitting Pi matters more than benefitting Pj if and only if (1.) the degree di to which the life of Pi is good is smaller than the degree dj to which the life of Pj is good and (2.) di and dj belong to the most precise scale by which the quality of the lives of all members in {P1, P2…. Pn} can be compared.

According to essentially comparative prioritarianism, priority in allocating a benefit is given based on the most precise comparison of the wellbeing of those who have a standing to claim the benefit in question. The scale of the most precise degrees is the appropriate tool for comparing the quality of the lives of Jerry+ and Jerry in the {Jerry, Jerry+} dilemma. Therefore, contrary to crude prioritarianism, essentially comparative prioritarianism says that benefitting Jerry+ matters more than benefitting Jerry, when they are the only claimants. It follows that, according to essentially comparative prioritarianism, the fact that no-priority is the resolution of the {Jerry, Bernard} and {Jerry+, Bernard} dilemmas doesn’t imply that Jerry+ has no priority in the {Jerry, Jerry+} dilemma. This is because the Jerry/Jerry+ comparison does not have to proceed by the crude degrees by which the Jerry/Bernard comparison proceeds.

Now, admittedly, both versions of prioritarianism overlap in most cases. Even according to essentially comparative prioritarianism, the vast majority of real world distributive

7 Larry Temkin “Intransitivity and the Mere Addition Paradox,” Philosophy & Public Affairs 16, (1987): 138-87.

dilemmas should be resolved based on crude comparisons of the wellbeing of potential beneficiaries. This is because, typically, the situation of the claimants differs in many factors and so the most precise degrees by which we can compare the wellbeing of the claimants would necessarily be categorical. The {Jerry+, Jerry, Bernard} dilemma is a simple example. The three of them are claimants of one dose of a very effective painkiller. The most precise comparison of the wellbeing of these individuals is crude, and crudely, their lives are as good. Then, according to both versions of prioritarianism, benefitting one of them is as urgent as benefitting the other.

Let me end by pointing to a result that might seem surprising to proponents of prioritarianism of all sorts. Consider Jerry++, and suppose that his life is categorically bad, while the lives of Jerry+, Jerry and Bernard are good. While Jerry++ is just below the good enough level, the other three are just above it. In the {Jerry++, Jerry+, Jerry, Bernard} dilemma, both versions of prioritarianism elaborated here say that benefitting Jerry++ matters more than benefitting Jerry+, Jerry and Bernard. This is so, even if Jerry++ is worse off than Jerry+ to the extent to which Jerry+ is worse off than Jerry.

Conclusion

I have shown in this essay that if the phenomenon of parity is to be understood through the Parity Model, the priority formula—benefitting people matters more the worse off these people are—is undefined. Two different interpretations to this formula suggest themselves. The most attractive one implies that prioritarianism is an essentially comparative view. It says that even if no-priority is the solution of the {Barnard, Jerry} and of the {Barnard, Jerry+} dilemmas, in the {Jerry+, Jerry} dilemma, benefitting Jerry+ matters more than benefitting Jerry.