Distributive justice and the epistemological argument against desert

Desert-sensitive theories of distributive justice seem to be making a comeback, after what was generally regarded as a devastating critique in John Rawls’s extremely influential A Theory of Justice. (1973[1971/1999]) They are presented as having “a promising future” (Moriarty 2018: 153) and as likely to make significant contributions to debates on distributive justice “for some time to come”. (Moriarty 2018: 171)1 This unexpected success, however, will seem less surprising if we take into consideration the fact that Rawls’s critique of desert-sensitive theories was actually less devastating than it initially seemed. What it questioned was the practicability of the notion, (Young 1992, Moriarty 2002) not, as some commentators suggested, the very notion of desert or its reality. (See Baiasu UM for a discussion of the debate.) In other words, the claim was not that the notion is contradictory or empty, with no real-world instance to refer to, but that we cannot precisely ascertain who is deserving and to what extent. Moreover, more recently it has been suggested that even the practicability problem could be addressed. (Moriarty 2005) This paper focuses on one aspect of the problem of practicability. The paper argues that, in virtue of an issue related to the epistemological argument against desert (to be presented


Introduction
Desert-sensitive theories of distributive justice seem to be making a comeback, after what was generally regarded as a devastating critique in John Rawls's extremely influential A Theory of Justice.(1973[1971/1999]) They are presented as having "a promising future" (Moriarty 2018: 153) and as likely to make significant contributions to debates on distributive justice "for some time to come".(Moriarty 2018: 171) 1This unexpected success, however, will seem less surprising if we take into consideration the fact that Rawls's critique of desert-sensitive theories was actually less devastating than it initially seemed.What it questioned was the practicability of the notion, (Young 1992, Moriarty 2002) not, as some commentators suggested, the very notion of desert or its reality.(See Baiasu UM for a discussion of the debate.)In other words, the claim was not that the notion is contradictory or empty, with no real-world instance to refer to, but that we cannot precisely ascertain who is deserving and to what extent.Moreover, more recently it has been suggested that even the practicability problem could be addressed.(Moriarty 2005) This paper focuses on one aspect of the problem of practicability.The paper argues that, in virtue of an issue related to the epistemological argument against desert (to be presented in Section 2), a Rawlsian skepticism against desert persists (Section 3).I think it is unlikely desert-sensitive theories will be able to respond to this scepticism satisfactorily.Yet, I claim (Section 4) that this should not be taken to imply that desert-sensitive theories should not be applied at all in our societies, but only that their application should be limited to certain distributive issues.

The epistemological argument against desert
The notion of desert is standardly considered to refer to a three-place relationship between a subject of desert (usually a deserving person), a deserved treatment (in the case of distributive justice, a reward) and a basis of desert (that in virtue of which a person deserves the corresponding treatment).(See, for instance, Baiasu 2007: esp.180-2.)For instance, Dan may deserve a high income in virtue of his performance at work.
Assuming a basis for a deserved reward is something for which the deserving person can claim credit, we have the following epistemological problem of desert (see Moriarty 2005: 206): natural and social factors influence a person's actions and features; hence, she will not be able to claim complete credit for anything.If we cannot isolate that portion of a person's achievements for which she can claim credit, we cannot ascertain the person's just deserts and we cannot reward them.The epistemological problem of desert does not deny the coherence of the notion of desert and it does not even deny its reality; it only challenges the possibility of its accurate application.
One solution to the epistemological problem of desert has been offered by reference to John Romer's theory of equal opportunities.(See esp.Roemer 1998 and1993) Consider the following example offered by Moriarty (2005: 216-7): if the deserved treatment is admission to college and the basis for admission is achievement in high-school, then we need to determine first the natural and social factors which affect achievement in high-school.Let us assume that the latest research on this question yields the following factors: teacher quality, classroom size, native cognitive abilities and perceived parental expectations.
Roemer's theory of equal opportunities offers the following ingenious solution. 2In order to isolate the portion of a student's achievement in high-school for which she can claim credit, we compare her achievement with those of the students who are similarly affected by the four factors mentioned above.Hence, students will be grouped into types, where a type includes the students who are equally well (or badly) off in terms of the four factors.For instance, we may have a type with poor teachers, large classrooms, high IQs and low parental expectations.Since they are affected in more or less the same way within a type, students will be more (or less) deserving compared to the other members of the same type depending on how well (or badly) they performed relative to each other.
For instance, if P scores 80 and Q, 85, and if P and Q are within the same type, then Q is more deserving than P. If P and Q belong to different types, then we can first compare their scores with those of students within their types and then obtain their percentile ranks.The student with the higher percentile mark will be more deserving.For example, say P's score of 80 places her in the 70th percentile rank for her type, whereas Q's 85 puts him in the 60th percentile rank.P is then more deserving than Q despite Q's higher overall score.This is because the percentile ranks are reflections, across types, of what P and Q can claim credit for.
I have now presented the epistemological argument against desert.This is the argument that individual deserts cannot be ascertained, since we cannot isolate those factors which are not genuine desert bases (natural and social factors).I have also outlined a solution to this problem.The suggestion has been that the difference made by genuine desert bases can be identified by placing in the same group individuals affected by the same natural and social circumstances.Those doing better would therefore be more deserving and cross-group comparisons can be made.In the next section, I will focus on an objection to this solution and, more exactly, I will argue against a particular answer to this objection.

Mistaken desert ascription and progress
The solution to the epistemological problem of desert presented above relies on the assumption of a correct identification of the factors which affect how a person becomes deserving of the reward.In the example discussed, we identified four factors, which, given the latest research, seem to affect achievement in high-school.However, as research in the area continues to be produced, we may discover that we have misidentified the social and natural factors affecting achievement.
In reply to this, the answer I want to consider is that, in fact, the more factors we correct for, the more likely we are to identify the person who is more deserving.(Moriarty 2005: esp.218-9 3 ) This assumption, however, seems unwarranted.While I agree that by discounting some factors, we can improve our ability to ascertain desert, I would like to show that there is no guarantee that identifying more factors will continue to improve our desert-ascertaining capacity.
To see this, consider the following example: P, Q and R are competing for admission to college.Imagine we are at a stage in the development of empirical studies in education and we are in the context of a society, where IQ is considered by far the most important natural/social factor affecting achievement in high-school.Let us assume that P's, Q's and R's scores are, respectively, 90, 80 and 70.Now, since P has, for instance, a high IQ, Q a moderate one, and R, low, they end up in the following percentiles, respectively: 60th, 70th and 80th.As a result, Q and R are admitted in college, whereas P is rejected.
We can now imagine that empirical studies develop and it turns out that, for the society in which P, Q and R live, the quality of the teacher is also a significant factor affecting achievement.Looking back at the situation for P, Q and R, we discover that P had one of the worst teachers, whereas Q and R, some of the best.Their percentiles now, from the perspective of the new theory, would be, respectively 70th, 60th and 50th.With these new results, P and Q are admitted to college, whereas R, rejected.Now, after another iteration in the development of empirical studies, apart from IQ and teacher quality, perceived parental expectation becomes also a significant factor.Given that P had very encouraging parents, whereas Q very discouraging (with R's parents being relatively indifferent), the new percentile ranks are, respectively 50th, 80th and 60th.Let us assume, finally, that this is the just outcome -Q and R are admitted, whereas P is rejected.This suggests that correctly identifying more natural/social factors does not necessarily improve our capacity to ascertain desert.In some cases, it may even make it worse, as in the example just presented and schematically represented in Figure 1.

Conclusion
We do not seem to have an end in sight for the development of our empirical disciplines, including in education.In principle, research relevant for the discussion above may not only identify new achievement-affecting factors, but may also show that some of the factors already identified have only a negligible impact on achievement.However, what I take my objection above to have shown is that, even when empirical studies correctly identify natural/social factors which may affect desert bases and even when the degree with which they affect such factors are accurate, there is no guarantee that our capacity to ascertain desert will continue to them as possible.This is what the pragmatic theory does.Although it does not ensure total accuracy, it ensures greater accuracy."(Moriarty 2005: 219)

Academia Letters, May 2021
Corresponding Author: Sorin Baiasu, s.baiasu@keele.ac.ukCitation: Baiasu, S. (2021).Distributive justice and the epistemological argument against desert.Academia Letters, Article 127.https://doi.org/10.20935/AL127.be improved.The conclusion, however, is not that we should abandon desert as a criterion of distributive justice; what I would suggest is that we should make sure that desert is not used as a criterion of justice when at stake are goods and rewards (such as healthcare resources), which are crucial for our lives.
Let me make explicit some of the reasons which I take to justify this conclusion. 4Recall that a desert-based distribution determines distributive shares by taking into consideration that part of what individuals have done or are doing, for which they can claim credit. 5This view of distribution offers a perfect response to what seems to be a definitional aspect of justice, namely, that it is unjust to reward or punish a person in virtue of aspects for which she is not accountable.For instance, we find it unjust to reward a person for something for which it turns out she cannot claim any credit. 6he capacity of a desert-based distribution to satisfy this condition of justice makes desert a valuable criterion of justice for contexts where distribution in accordance with accountability is required. 7Moreover, it can be argued that there is heuristic and symbolic value in employing this criterion of distribution despite the epistemic uncertainty I argued for above.Thus, by attempting to ascertain individual deserts, we have at least a chance to discover a more promising approach to the epistemological argument against desert.This applies despite the fact that there is no guarantee of progress in this respect.Moreover, by employing this criterion of distribution, we reassert (at least symbolically) the value of responsibility in distribution, a value which would be ignored by accounts which employ other criteria of distribution (e.g., equality, need or efficiency).Now, if the argument of this paper is correct, we cannot reliably ascertain individual deserts, since the process of ascertaining individual deserts depends on the results of empirical research concerning the natural and social factors affecting specific performance.Such an investigation is in continuous development and may upset outcomes, which rely on the conclusions of previous research.It follows that, through desert-based distributions, we risk introducing further injustices in the world.We have, therefore, (heuristic and symbolic) reasons to employ desert as a criterion of distribution and, at the same time, given potential unjust situations generated, reasons not to employ desert as a distributive criterion.Hence my proposed solution: using desert as a distributive criterion in relevant situations, where the goods to be distributed are not crucial for our lives.In this way, we may pursue the values of desert-based distributions without risking to commit significant injustices. 8those factors which may unfairly advantage or disadvantage students.
8 Bogdan Olaru raises the important question of how we can separate between goods which are crucial to our lives and those which are not.Would education, a place to a college, for instance, not be a significant good, which should perhaps not be distributed on the basis of desert?I certainly agree -the example of admission to college, which I use in the paper, is not meant as an example of an insignificant good, but as an example of a context where desert is an important criterion of distribution.I tend to agree that education is crucial for our lives.I do not propose a distinction here between crucial and insignificant goods.All my argument needs is the acceptance of a weak claim, according to which at least some goods are not extremely significant for our lives.Rejecting this would reflect a kind of moral enthusiasm, where everything is normatively extremely significant for us, and this seems to me implausible.