1 Introduction

“Justice” claims to provide a central moral standard for judging not only individual behavior but also the basic structure of societies and their institutions (Cohen 1986, 1; Rawls 1999, 3). However, the scientific and philosophical debate about justice has not yet revealed a clear paradigm, with several lines of argument competing here (Cohen 1986; Sandel 2009; Miller 2017):

  • Grand theories of justice, usually connected to a famous proponent, for example Virtue – Aristotle, Utilitarianism – Bentham, Autonomy (Deontology) – Kant, Egalitarianism – Rawls, Libertarianism – Nozick (Sandel 2009).

  • Basic principles that would (alone or in combination) govern the human quest for justice, among them virtue, happiness, desert, merit, sufficiency, priority, need, equality, and liberty (Tyler et al. 1997, 56ff.).

  • Different conceptual approaches, for example naturalism (tracing justice back to natural phenomena) or contractarianism (justice being a social agreement), but there are many more (cf. Olsaretti 2018).

We find major conceptual contrasts (tensions, dichotomies). Among these are:

  • Substantive versus procedural justice (Miller 2017): Can justice strive for certain desired outcomes, or must justice restrict itself to just procedures, accepting any outcome? More generally, social psychologists are concerned with how procedural questions (non-outcome factors) affect perceptions of justice (Lind 2020).

  • Cognitivist versus decisionist approaches (Quante 2013, 40ff.): Is justice based on principles to be discovered or (only) subject to contingent agreement? A similar contrast is rationalist versus empiricist (Binmore 2005, 38): Can we deduct moral principles from reason alone, or do we have to consult data from the real world?

  • The role of intuitions (Gigerenzer 2007; Haidt 2013): Are intuitions (gut feelings) the benchmark for “genuine” justice or merely heuristics for rational reasoning?

This article focuses on distributive justice, i.e., the allocation of benefits and burdens in societies, the most important allocation being income for work. It does not discuss non-outcome factors like procedural fairness because in the end material outcomes are decisive (Harris 1979; but see also Miller 2017, 13). In markets, merit-based allocations (income proportional to economic contribution) enjoy practical prominence and widespread approval (Miller 1999; Mulligan 2018; Adriaans et al. 2019). But concerns arise about increasing inequality (Piketty 2014), and merit as a normative basis is widely questioned (Rawls 1971; Frank 2016; Markovits 2019; Sandel 2020).

The proportionality of inputs and outcomes (i.e., merit) as the hallmark of an equitable relationship was also the core proposition of Equity Theory. Equity Theory was the leading paradigm of social psychology during the 1960s and 1970s and led to hopes of becoming a general theory of social interaction (Adams 1963, 1965; Berkowitz and Walster 1976). Both inputs and outcomes could be positive or negative, and their relation was presented in a complex, somewhat counterintuitive mathematical formula (Walster et al. 1973). Note that the term equity has two meanings (Leventhal 1980, 29): Narrowly it refers to Equity Theory’s contributions rule which says that inputs must balance outputs. In everyday language it refers rather to justice or fairness in general (see also Young 1994). Throughout this article we will use the narrow definition of equity and prefer the term merit to the term equity.

While Equity Theory was considered plausible for most economic transactions, many societies fulfill the basic needs of their weak members, and often people are treated equally, economically or politically. Altruism is widespread: People donate blood or organs, get committed to refugees or environmental protection without expecting anything in return. The application of the quid pro quo of Equity Theory to intimate relationships was deemed especially disturbing (Walster et al. 1978, ch. 6). Equity Theory was criticized for its “unidimensionality” and other conceptual shortcomings (Deutsch 1975, 1985; Leventhal 1976b, 1980; Schwinger 1980; Folger 1986), and gradually the “principle triad” of merit, need, and equality emerged from the social-psychological debate. This approach quickly and consistently found the support of many scholars (see Lerner 1977; Mikula 1980; Reis 1986; Kabanoff 1991). Today it represents one paradigm within the justice discourse of social psychology (Lind 2020), and social justice research frequently refers to these three principles (e.g., Sabbagh and Schmitt 2016; Van Hootegem et al. 2020; Narisada et al. 2021).

In philosophy, desert-based theories of justice have a similarly bad reputation. The act of doing something for others can constitute a moral claim: people deserve certain benefits in light of their actions, and justice requires getting them (Feldman and Skow 2016; Lamont and Favor 2007; Mulligan 2018). This approach is traced back to Aristotle who argued in his Nicomachean Ethics (1131a) that “all men agree that what is just in distribution must be according to merit.” But desert has no philosophical lobby today. Rawls (1971, 1999) is considered being most influential in killing the concept: no one deserves anything at all, and, despite lacking empirical support, mainstream philosophers prefer variants of either egalitarianism (equality) or libertarianism (a purely procedural approach) (Miller 1999; Mulligan 2018). Mulligan (2018) recently published an elaborate defense of meritocracy and argued that, empirically and across ideological and cultural lines, people want rewards to reflect merit. Miller (1999, 2017) occupies a special position among philosophers: He endorses merit but only in combination with need and equality, i.e., the principle triad.

The term “principle of justice,” however, is used with two meanings. The social-psychological distinction between allocation goals and allocation rules (or distribution rules) is helpful in this respect (cf. Leventhal 1976a; Mikula 1980b; Greenberg 1982). Allocation goals are the preferences of the allocator. They are individual benchmarks of the allocator for maximizing his or her utility and can be implicit or explicit (e.g., keeping group productivity high or preventing a person from protesting against a certain distribution decision). Allocation goals are beyond the scope of the theory. Allocation rules are the principles of justice (in their plural meaning), they are practical rules for “dividing the cake” and hence often contested. Nevertheless, they must be consistent with a primary or foundational principle of justice (in its singular meaning), provided that you are not a “pluralist” who would deny that one such principle exists at all (for example, Walzer 1983; Miller 1999; Deutsch 1985). Social psychologists tend to use the term principles of justice in the sense of allocation rules, while philosophers tend to use it in the sense of primary principles, a normative foundation.

Whether independently or as a triad, the three allocation principles merit, need, and equality occupy center stage in the general debate about distributive justice, theoretically and empirically (Scott et al. 2001), compared to other principles like entitlement, happiness, virtue, or priority. Yet their relation remains diffuse, and current theory does not inform political practice (Honneth 2008). As long as merit, need, and equality are seen as competing principles, the dualism of other-regarding preferences vs. “selfish” motives will persist (see Sect. 2.4), and social policy will continue to struggle with redistribution (see Sect. 6).

Here, we aim to develop a coherent picture with an interdisciplinary analysis of when, why, and how people in modern societies allocate benefits and burdens according to merit, need, and equality, setting it in the evolutionary context of reciprocity. In the terms of Leventhal et al. (1980), this article is a theory of allocation preference (more precisely: a theory of allocation rule preference), combined with a theory of a foundational principle of justice. It defends the thesis that such a foundational principle exists: a balance in social exchange, called reciprocity. This translates into one fundamental allocation rule Meritocratic Principle, complemented by several other auxiliary allocation rules, most prominently need and equality. While merit matches reciprocity naturally, need (“altruistic” acts) and equality do not: what could fill the gaps between seemingly higher costs and lower benefits for one of the partners when choosing need or equality as distribution rule? We will show how classical applications of need and equality can be reconciled with reciprocity without ceasing to be distinct (auxiliary) principles in practice. “Fundamental principle” does not mean “only principle” but a hierarchy of principles (allocation rules) according to context and constraints.

The methodology chosen is that of interdisciplinary explanatory coherence (Thagard 2000). This approach favors a broader coherence (compatibility) between different fields at the expense of narrow consistency within one field with its specific argumentative toolbox. A Theory of Justice must address (and balance) both normative and explanatory questions. When social practices are used overwhelmingly, timeless and culture-invariant, and enhance cooperation (as are reciprocal practices), they are not arbitrary choices but reveal a deeper social meaning that has to be explained. It is certainly a naturalistic fallacy to infer directly from natural facts to normative judgements (Moore 1903). The “moderate naturalism” chosen here means that no direct conclusions are drawn from natural facts, but that natural facts are relevant, in the sense of contributing to a coherent overall picture (cf. Vollmer 1995; Racine 2008).

In Sect. 2 we depict the evolutionary roots of reciprocity as “ideal justice” (foundational principle), describing the mechanism as well as the term’s broader (and contested) usage. The section also presents the corresponding social norm Meritocratic Principle. Section 3 explains why merit as a principle of justice (allocation rule) is not enough. Justice is rather a social bargain and an optimization problem than a clear-cut principle, and we have to explore the trade-offs and communicative difficulties of this process. Sections 4 and 5 show how several classical applications of need and equality can be reconciled with reciprocity. They are auxiliary allocation principles when merit is not effective, not efficient, or not communicable (or simply coincides with equality), but their scope is limited by merit. An important topic is implementation (Sect. 6): How is a Theory of Justice “applied,” especially regarding modern (anonymous) societies? Focusing on “non-merit” will open new and more effective policy options. Section 7 draws some conclusions.

Because there has been a shift from the outcome-focused to the relationship-focused paradigm in social psychology, beginning in the middle of the 1970s (Lind 2020), many of the works concerning distributive justice cited here date back to the outcome-focused period.

2 Reciprocity: from evolutionary roots to a social norm

2.1 Selective forces

Evolutionary selection has shaped a process in which individuals compete for resources to replicate their genes by reproduction, and humans make no exception. Besides competition, fitnessFootnote 1 can also be enhanced by cooperation, but according to the selection rules every cooperating individual must benefit from it in the long run. Today, the existence of a genetically based altruism can be ruled out – this would contradict any evolutionary functional logic (Voland 2013, 65).

Among animals, mutualism (behavior that provides direct benefits to every mutualist) can be observed regularly, often between species (Leigh 2010). But only higher primates seem to have evolved the cognitive abilities required for reciprocity, i.e., accepting costs for the benefit of others which are rewarded later, either by the beneficiaries themselves (direct reciprocity) or by others (indirect reciprocity) (Trivers 1971, 2006). Reciprocity is dependent on repeated interactions, and it is endangered by cheating, so the cognitive abilities required include individual recognition, temporal discounting, and memory (Stevens et al. 2005). In parallel, mechanisms have evolved for so-called altruistic punishment of cheaters, but also for building up reputation and the assignment of social approval (prestige) (Voland 2013, 74ff.).

2.2 Forms of reciprocity

Reciprocity among humans has been widely discussed for different ages and cultures in anthropology and sociology (see Adloff and Mau 2005), but only recently in psychology (Kurzban et al. 2015). Note that some authors (especially in economics) restrict the term reciprocity to a personal relation and motives of social approval (cf. Konow and Schwettmann 2016), but reciprocity implies an equivalence in all exchange relations: to repay in kind what another has done for us, materially or socially. Voluntary exchange will only occur when the benefits exceed the costs for both partners (costs and benefits in a wide sense, Blau 1968; Becker 1976). Due to different individual levels of marginal utilities and marginal costs, both partners can benefit from exchange. This proposition lies at the heart of Social Exchange Theory (Thibaut and Kelley 1959; Homans 1974; Blau 1964), and Neoclassical Economics uses the concept of consumer and producer surplus to characterize such situations.

The condition of equivalence must be met at least in the long run (Kurzban et al. 2015). Sahlins (1965) introduced the concept of generalized reciprocity: transactions that are seemingly “altruistic” but can be expected to be returned, not necessarily here and now, not necessarily by the beneficiaries themselves, undetermined in time, quantity, and value (for a compact overview, see Holcombe 2020, ch. 5). These transactions might be better described as investments. The modern welfare state, for example, can be regarded as an institutionally mediated, reciprocal arrangement, and several typologies of reciprocal expectations can be construed (Lessenich and Mau 2005; Fong et al. 2005). A decisive point is that free-riding is effectively prevented or punished. That free-riding is a social problem remains unintelligible without assuming a violation of reciprocal expectations (Fehr and Gächter 1998; Panchanathan and Boyd 2004).

Reciprocity therefore has two meaningsFootnote 2: (1) It is a mechanism (“reciprocal accounting”) that can stabilize mutually beneficial cooperation. (2) Along the range altruism \(\sim\) reciprocity \(\sim\) egoism, the term can be used to describe the overall fitness consequences of a social exchange, i.e., its individual advantageousness or disadvantageousness (cf. Sect. 2.4).

2.3 Ideal justice

From the evolutionary point of view (natural selection), equivalence is the only stable (and symmetrical) solution to the problem of how two individuals would engage in voluntary social exchange. Voluntariness is a strong normative argument, based on systematic considerations (and a contested term itself, see below). The idea of equivalence can be readily extended to the negative range: The proportional relation between desert and reward (using moral terms) also applies to guilt and atonement (Fig. 1). The purpose of punishment is to impose enough costs on the factual or potential defector to offset the temptation to violate the equivalence principle (Boyd and Richerson 1992; Clutton-Brock and Parker 1995). A general formulation of such violations of reciprocity (injustices) could be “advantages at the expense of others,” or “not bearing the costs of one’s own choices.” Durkheim (1933, 88) emphasized the intended exact balance between the severity of a crime and its punishment (see also Buchanan and Mathieu 1986, 13). People would even “construct” this balance if a person’s fate is not perceived as just, especially by blaming victims for their “own fault” (“Just World Belief,” Lerner 1980). Austin and Walster (“Equity with the World,” 1975) showed that subjects try to “heal” inequities across several relationships, if necessary.

Fig. 1
figure 1

The Ideal of Justice: an exact proportionality between personal benefits and (opportunity) costs (axes labeled with moral terms). Walster (1975) presented a similar depiction of an equitable relationship but see Sect. 3.4 regarding the notion of inputs and outcomes

One of the most striking arguments in favor of reciprocity as the foundational principle of justice is the existence of property, money, credit, and prices as such (Holcombe 2020, ch. 5). The natural rights definition of property rights is based on the idea of an achievement resulting from one’s own labor (Locke 1967; Hume 1874; Smith 1981; Marx 1906), or in a proverb: “as you sow, so shall you reap.” These social practices reflect a basic collective consensus about the conditions of social exchange. An argument often neglected is the proportionality between the amount of money and quantity of goods. Philosophers of justice regularly argue about the legitimate level of hourly rates (unequal pay, iustum pretium, Koch 1995), but they have a tacit consent that hourly rates as such are legitimate, and so are any other prices. The equation \(amount = quantity \times price\) states that two hours of work will cost two times the amount of one hour (e.g., Tyler et al. 1997, 46), reflecting the linear relation of Fig. 1.

2.4 The contested meaning of reciprocity

Therefore, the minimum condition of any kind of voluntary social exchange is a reciprocal balance. This means (a) a ratio of benefits to (opportunity) costs of unity and (b) a ratio that is not too unequal for both, however this is assessed by the individuals – and their collectives (cf. Sandel 2009, 145f.: “When consent is not enough”). Justice is a social bargain between individuals but witnessed by society.

It is important not to confuse two overlapping, but distinct assessments of social exchange: (1) “Is it just?” This concerns the “public” cost-benefit estimation, i.e., the visible (communicable) aspects of reciprocity that are depicted in Fig. 1. (2) “Is it worth the effort?” This concerns the individual, “private” cost-benefit estimation. Beyond the obvious, it includes costs and benefits that are not well communicable or only indirectly related to the transaction and hence unsure: when do benefits cease to be related to a specific social interaction and hence make for “genuine altruism,” like donations, loyalty, or civic engagement? The difference becomes salient for example when Nesse (2001b, 5) objected to “[t]he tendency to use reciprocity to stand for all cooperative relationships” and listed several “other” mechanisms why humans would cooperate (e.g., mutualism, coercion, commitment).

The terms altruism, reciprocity, and egoism (or selfishness), and their corresponding adjectives in particular, have long been a source of confusion, by mixing up descriptive and normative aspects (e.g., Trivers 1971; Wilson 1992; and especially the “social semantics controversy:” West et al. 2007, 2008; Wilson 2008). There seems to be a strong desire among some scientists to prove that humans are capable of “genuine” altruism, untainted by reciprocal considerations (e.g., Fehr and Fischbacher 2003; Gintis et al. 2005; Holmes et al. 2002).

Altruistic acts are usually defined as behavior which is costly to the actor, beneficial to the (unrelated) recipient, and cannot be expected to being returned. Evolutionary theory argues that such behavior could never have been selected for and assumes (in a kind of backward reasoning) that some long term benefit must occur – hence, such behavior is not altruistic or unconditional in the strong sense of the word (West et al. 2007). For common sense, however, it suffices that no returns are expected, that this behavior is not aiming at individual advantages. In the end, both views coincide, the difference being not actual outcomes but time horizon, causality, and – very important – intention (cf. the question of the moral motive, Sect. 6.1). But altruism is never unconditional, even normally understood. People always have reasons for their altruistic behavior. Unconditional in a strong sense would mean that giving money to people who don’t deserve or need it would be a socially meaningful action.

We argue that “altruistic” forms of other-regarding behavior can be plausibly viewed as an investment into the cohesion of society (see Sect. 3.1), hence do not contradict maximization of utility (and evolutionary theory). Instead of introducing a new adjective (like strong reciprocity, Gintis et al. 2005), with reference to Sahlins (1965) we will refer to this non-instrumental but targeted reciprocity as “very generalized reciprocity” (see Fig. 2).

Fig. 2
figure 2

Social exchange with regard to fitness consequences. Due to evolutionary bias, altruism is negligible and egoism substantial. But prevalent are shades of reciprocity

2.5 The meritocratic principle

2.5.1 Definition

There is only one fundamental normative standard: reciprocity. This translates into the social norm Meritocratic Principle, a widely accepted practical allocation rule, often formulated as comparison: “Those who do more than others shall receive more than others” (Bolte 1979, 26, translated; see also Harris 1976). Likewise note that a “reversed equity script” has been reported as well: “those who have more must have performed better” (Bierhoff et al. 1986, 174). Deserts require a personal desert basis (Feinberg 1970), and merit serves as the desert basis for economic benefits (Mulligan 2018). A meritocracy is a society that creates the legal framework and political instruments for linking live chances to “merit” and that discursively agrees upon what this means and how it is determined (Verheyen 2018, 65).

The Meritocratic Principle is fundamental in being the only allocation rule that is truly reciprocal (cf. the linear relation of Fig. 1). A formal (and lengthy) definition of merit can be found in Neckel, Dröge, and Somm (2008, 46) but in our view the best definition (especially for modern societies) is the legendary equation of Young (1958) in an extended version (Kariya and Dore 2006, 138),

$${{merit} = {talent}\,+\,{effort}\,+\,{luck}}$$

More easily than any other definition it can reveal the wrangling over the role of merit. This equation also matches economic theory when interpreting its main terms as “personal production factors:” the natural endowment of individuals with talents (capital) and their efforts (labor), complemented by the contingencies of the economic process and the sometimes hardly predictable assessment by others.

Regarding the relation between merit and desert, merit focuses on “doing” and visible results (achievements). Young’s formula is not about character and high ideals but about usefulness. Desert as a philosophical notion is consequential as well (you have to act in certain ways to deserve something) but, compared to merit, it emphasizes aspects of moral excellence and virtuous life (see, for example, Wagstaff 1994). This aspect of communicability and visibility is probably the most salient difference between merit and desert. Systematically, they are two terms for the same thing: If you do something good, you deserve something in return, be it money or social approval.

2.5.2 Tugging at the terms of Young’s equation

Luck is the least contested term in Young’s equation, for it has basically nothing to do with desert. Nevertheless, the term is important for structural and discursive reasons (cf. Dröge et al. 2008, sec. 1.7). First, it is unavoidable, second, people argue about what counts as luck, and third, it is often impossible to clearly separate it from the rest. Luck egalitarianism (e.g., Dworkin 2000) is effectively a debate about luck vs. choice, i.e., responsibility (cf. White 2007, ch. 4).

At the least, luck must not be too large. Today, a widespread economic contingency is one of the main acceptancy problems of the meritocratic principle. Neckel (2001, sec. 4) emphasized that merit as a foundation of distribution becomes absurd when jobs themselves have become scarcities (“economic culture of randomness”). On the other hand, Saunders (2006, 192) argued that luck cannot be systematically exploited, and everyone can have it, so it would not question the legitimacy of meritocracy.

Talent (or ability) is ambiguous. Exemplary for many, Rawls (1999, 86) emphasized that “inequalities of birth and natural endowment are undeserved”, results of a “natural lottery” (p. 64). At the most, achievements based on talent could entitle legitimate expectations from society, given that appropriate rules had been established in advance: “But this sense of desert is that of entitlement” (p. 88f.). Rather like in a tug-of-war, Rawls tried to “tug” talent toward the term luck and to turn a moral desert into a question of contingent entitlement. Libertarians argued the other way round (cf. Sandel 2009, 69; Rosenberg 1987): Talent requires a lot of effort to develop (“knowledge”), and who else but the talented should be allowed to benefit from his or her talent? Libertarians tried to tug talent toward the term effort and turn a moral desert into a question of property rights.

Generally, it is contested whether talents are mainly genetically determined, mainly a product of education and “background,” or mainly acquired by practice (Rosenberg 1987). Discussing meritocracy, Dear (2019) criticized that political philosophers yet have missed to develop a clear concept of talent and its implications. They would stick to a static nativist theory whose straight causal relation between talent and success simply would not exist (p. 162). Modern talent research shows that expert performance is not possible without intense training and long practice, indicating that it is decisive how early persons begin with focusing on certain areas of interest, how much they invest, and how much support and encouragement they receive (cf. Ericsson et al. 2018).

The term effort is indeed for many the allegory of a deserved income. Everywhere in the world hard working people are acknowledged as creators of economic value – at least if this effort is targeted at a socially desirable outcome and achieves a result. Throughout the debate the tension between effort and achievement is a recurring topic: effort without achievement is a pity, achievement without effort has no merit (Heckhausen 1974; Braun 1977; Bolte 1979; Deutsch 1985; Neckel 2001; Dröge et al. 2008). Neither entitles reward, and it could even be argued that the equation should rather read merit = talent × effort + luck, so that if either talent or effort are zero, their product would be zero as well, the remainder being luck. Yet the basic problem of “performance-related pay” is that while its normative basis can only be merit, its transparent (communicable) basis can only be achievements (Brei-sig 2003, sec. 2.2).

Rawls (1999, sec. 12) went even further and declared that the ability to make an effort was dependent upon a happy childhood and social circumstances (cf. also Sandel 2009, 158). In fact, Rawls tugged any term of Young’s equation toward luck, apparently to be free for his contractarian approach to justice, i.e., contingent agreement (cf. Miller 1999, pp. 53–59).

Karl Marx concentrated on effort and called this, following his classical antecessors, a labor theory of value. In his writings Marx spoke of the equality of human work, of the time elapsed as a measure of effort (Heinrich 2005, 71ff.), but also that a higher qualification would cause higher costs, and hence justify a higher value of the working time (hourly rate, p. 91). He largely ignored talent and fully disregarded the contribution of a whole class, namely capital as the merit of the entrepreneurs, which he assumed to be mere organization plus supervision of exploitation and therefore zero (p. 157).

Talent and effort have yet another dimension: Societal functionality. Early accounts of the social benefits resulting from functional stratification are Durkheim (1933) and Davis and Moore (1945). The importance of not discouraging the talented is another recurring topic of the debate about merit, tightly connected to secure property rights (cf. Acemoglu and Robinson 2012).

2.5.3 Conclusion

Any objective definition of merit is impossible since it depends on context and circumstances – or the other way round: Merit is uncontroversial only when defined in extremely narrow terms (cf. Deutsch 1985, 133f.; Breisig 2003, sec. 3.2). What merit in individual cases means is discussed collectively and will always be controversial because of conflicting interests.Footnote 3 Social exchange is a relation that requires concessions on both sides. It is always the beneficiaries of an achievement who will assess its utility. This is a core principle of the division of labor and reflects the general tension between individual and collective perceptions of merit. In markets, the interplay of supply and demand and hence price formation is not the result of luck but of this decentralized reconciliation.

Contrary to Young’s apprehension and other objections, the meritocratic principle is a fundamental and widely accepted social norm: Meritocracy “resonates powerfully with deeply held ethical values about fairness, and these are broadly shared throughout the population” (Saunders 2006, 193, original emphasis). It “corresponds to the widespread belief that people deserve to enjoy unequal incomes depending on their abilities and how hard they work” (Miller 1999, 178). The meritocratic principle establishes a relationship between personal market value and contribution to productivity (Marris 2006, 159). Critics of meritocracy in fact often criticize that people do not get what they deserve, i.e., they actually defend merit (Mulligan 2018, 131).

Unlike many other theorists, Miller (1999) considered popular conceptions of justice (“what the people think”) and empirical data. He stated that desert and need criteria feature prominently here (p. 90), and criticized political philosophers, Rawls among them, for neglecting such empirical evidence, especially when it is in favor of desert (ch. 7). Though meritocracy is sometimes suspected to serve as a hierarchy-legitimizing ideology that justifies current societal inequality, experiments show that the principle of merit is endorsed independently of its actual implementation (Son Hing et al. 2011). Empirical support for merit as a base for distributive justice can be found for example in Saunders (2006), Neckel et al. (2008), Mulligan (2018), and Adriaans et al. (2019).

The results of the scientific and philosophical debates about the meritocratic principle can be summarized as follows:

  • It is defendable as the most fundamental (reciprocal) allocation rule.

  • It is functional in increasing societal wealth.

  • It is not possible to define merit objectively.

  • In practice, the principle often is not met (effortless incomes, underpaid work, the competing principles need and equality).

Most objections against the meritocratic principle, however, can be rebutted by concentrating on “non-merit” (Sect. 6).

3 Justice as a social bargain and optimization problem

Figure 1 depicts the ideal of justice, demanding to appreciate the circumstances (costs and benefits) of every single case. But neither our physical nor our social world are ideal. Among the real-world restrictions are (1) costs and time constraints for information retrieval, (2) limitations for communicating emotions, convictions, and intuitions, (3) general uncertainty about future events, (4) strategic behavior of our fellow humans.

Moral reasoning is a communicative strategy for resolving conflict, enhancing status, and maintaining cooperation (cf. Deutsch 1985, 26; Mikula and Wenzel 2000; Haidt 2013), making justice rather a social bargain. The ideal of justice is just one of many dimensions (cf. Reis 1984; Deutsch 1985, 99; Young 1994), and not always the strongest force. In practice, justice must in our view fulfill at least three additional requirements: effectiveness, efficiency,Footnote 4 and communicability. For each, an Aristotelian “mean between extremes” is optimal but since trade-offs exist between them, we actually have an optimization problem.

3.1 Effectiveness

Stylized range: functionalist \(\sim\) expedient \(\sim\) sticking to principles

Applying principles of justice has to increase individual or collective utility, this is one core proposition of utilitarianism. If justice degenerates into “pure principle” (dogmatism), creating problems rather than solving them, it misses the target (e.g., Kant’s murderer dilemma, Sandel 2009, 132). Mikula (1980b) discusses extensively the role of “allocation goals,” and Leventhal (1976a) and Greenberg (1982) are rich sources of pragmatic considerations when choosing an allocation rule. Effectiveness has priority when it comes to rationing (Elster 1992; Young 1994). On the other hand, justice must not be functionalist, this is one core proposition of deontology. If justice uncritically maximizes “outcome” (especially in the short run) it becomes arbitrary, and many individual and societal self-bindings remain unintelligible (cf. Sandel 2009, 32f.).

We view self-bindings as an important, yet underestimated aspect of such effectiveness considerations. Self-bindings (also called commitment) are self-imposed restrictions designed to help resist actions that are tempting now but detrimental in the long run (Schelling 1978; Elster 1979, 2003). For the topics discussed here we suggest the following definition:

A self-binding is the strict adherence to a contingent rule which aims at some “higher goal” that is difficult to achieve directly.

We prefer the term self-binding over commitment because the latter is too near to a (personal) promise and lacks the contingency aspect. A self-binding lies somewhere between “pure investment” where costs now will surely reap benefits later, and “pure superstition” where costs now definitely have no connection to assumed benefits later. “Higher goal” means that the connection between the rule (costs) and the desired outcome (benefits) is not transparent. Their probabilistic relation may be plausible in some cases but controversial in others (see examples below). Therefore, the rule is contingent. But the rule must be strictly adhered to for it to be communicable, fulfillable (clear criteria), and socially monitorable. Put it another way: A self-binding is a communicable point on the blurred Aristotelian continuum between the obvious extremes.

Hence, a self-binding could also be viewed as a “trust-led investment”Footnote 5 that only looks like non-consequentialism. Regarding motives, there is no need to restrict self-bindings to be intentional (conscious), as Elster (2003) did. The unconscious is our unconscious that pursues our goals and can make judgements on its own (Sumser 2016, 275, original emphasis).

The following list of examples is neither exhaustive nor strict in its categories (see also Nesse 2001, for an overview):

  • On the individual level, we find concepts like civility and good conduct, moral principles, loyalty, veganism or a “pro-life” stance, awareness-building, religious dietary laws, superstition, and political correctness. People believe rather than know that certain behavior would be favorable or adverse for themselves or for running a society, and they discuss passionately about it (Haidt 2013). Self-bindings can be part of a (political) identity and contribute to authenticity (e.g., Greenebaum 2012). The higher goal of credibility may sometimes require even “silly” investments (keeping a ruinous promise, carrying out a senseless threat, Nesse 2001b, 12).

  • On the level of societal subsystems, we have most prominently science and the judiciary. For the higher goal of finding “objective truth” in a society of strategic self-interested individuals, both have given their systems strict, transparent rules so as to avoid or correct any biases of their members and maintain a high standard of impartiality.

  • On the level of societies (nation states), we observe restrictions regarding the ownership of weapons or consumption of drugs. Constitutions are among the highest-level national self-bindings we know, “intricate pieces of machinery” (Elster 2003, 1779).

The tension between rigidity and flexibility of self-bindings is a recurrent topic. The “dogmatic” adherence to a self-binding is in fact its caricature (because the rule itself is contingent), but a “flexible” self-binding is neither reliable nor communicable or monitorable, so at least public collective self-bindings can be nothing but “rigid.”

3.2 Efficiency

Stylized range: sweeping \(\sim\) practicable \(\sim\) sophisticated

Since humans have evolved to maximize efficiency (Sanderson 2001, 148), justice should be rather straightforward and “easy,” at least practicable. It must be applicable to everyday situations, not only to philosophical thought experiments. Long theories, complex reasoning and the elaborate dissection of moral subtleties are as “unjust” as endless meters of tax legislation and its commentaries. But justice cannot be allowed to simply flatten relevant differences between the cases in question to push through one single principle. We cannot deviate too far from our ideal of justice to appreciate the individual circumstances of every single case.

According to Gigerenzer (2007, ch. 8 and references therein) our intuitive judgements are often based on only one good reason. This is contrary to the tenets of rational decision theory which holds that we must consider and weigh all relevant information. A reason is a cue or signal that is significant for a relevant fact, and humans use reasons to make decisions intuitively and efficiently. The Take-the-Best heuristic is an application of that principle: reasons are ordered by relevance and then checked successively (lexicographically) as to whether they enable a decision. Instead of weighing all options “until the end,” the most relevant reasons are checked first.

Rules of thumb play an important role in moral behavior, the paradigmatic example being the Golden Rule: “Treat others as you would like others to treat you.” Another prominent example for a rule of thumb is “tit for tat,” a simple recipe for successful cooperation without being exploited (Axelrod and Hamilton 1981): Start with cooperation and then replicate every opponent’s move (cooperation or defection). Both rules are strictly reciprocal. A rule of thumb especially important for equality – equality as default – is discussed in Sect. 5.1.

3.3 Communicability

Stylized range: populist \(\sim\) thoughtful \(\sim\) unworldly

Justice must be “reasonable.” A Theory of Justice must be consistent and withstand rational reflection. Any distribution rule must fulfill the rule of the good reason and “treat like cases alike.” Theories that contradict basic intuitions of justice and ignore empirical evidence are to be refuted (Miller 1999). On the other hand, justice cannot be merely a question of majority (vox populi) or strong convictions. A precondition of a shared sense of justice is comprehensibility and transparency of the criteria. Justice must be socially monitorable.

Intuitions (gut feelings) refer to a judgement (1) that appears quickly in our consciousness, (2) whose underlying reasons we are not fully aware of, and (3) that is strong enough to act upon (Gigerenzer 2007, 16ff.). But it can be difficult to defend one’s intuitions. Haidt (2001) presented the “Social Intuitionist Model” (SIM) which initiated a turn from rationalism to intuitionism in moral psychology (cf. also Haidt 2013). According to SIM, moral intuitions have a primacy, they allow an immediate and effortless assessment of moral situations. The post hoc reasoning has above all a communicative function of (social) justification. Intuitions are not immune to the reflection of “thoughtful and well-educated” people (Ross 1930, 40), but to change one’s mind they usually need to activate another intuition. For justice, especially significant is the tension between a strongly felt moral conviction and its bad communicability (“moral dumbfounding,” Haidt 2013, ch. 2).

Gigerenzer’s (2007, 199) statement is concise: “Simplicity is the ink with which effective moral systems are written.” Some very interesting observations regarding simplicity can be found in Schelling’s classic “The Strategy of Conflict” (1960). Focal points are outcomes where mutual expectations easily converge because they “enjoy prominence, uniqueness, simplicity, precedent [...]” (p. 70). Schelling points to the “strong magnetism in mathematical simplicity” in international bargains and emphasizes the “remarkable frequency with which long negotiations [...] converge ultimately on something as crudely simple as equal shares, shares proportionate to some common magnitude [...] or the shares agreed on in some previous but logically irrelevant negotiation” (p. 67). This makes clear that the communicability of the outcome can be decisive.

3.4 The communication problems of equity theory

The way Equity Theory once was presented (and defended) was probably infelicitous. Equity Theory would allow both signs (positive and negative) for inputs and outcomes, what made the concept confusing, and it was rightly criticized that the meaning of inputs and outcomes has been repeatedly stretched ad hoc only to achieve a balance, making the terms meaningless (e.g., Schwinger 1980, 98f.; 105f., Deutsch 1985, 30; Reis 1986, 189). The mathematical formula presented conveyed the idea of precision and applicability but could never deliver on its promise (cf. Folger 1986).

The basic problem we see is that the social notion of “inputs” and “outcomes” refers implicitly to the visible (communicable) preconditions and effects of social exchange.Footnote 6 It is especially salient in the deviations from equity that inputs and outcomes can be difficult to communicate. Compared with this, the notion of costs and benefits with their definite sign and holistic meaning is much clearer. There will never be a quantification of equity (just another word for full communicability), as demanded by Adams and Freedman (1976), when the terms of the underlying formula are difficult to quantify, strategically used, and tugged between conflicting interests. But there could be a theoretical agreement on its underlying principle.

4 Need as auxiliary principle

This section discusses two topics where “giving something for nothing” (Gouldner 1973) seems to be a deviation from reciprocity: (1) Exchange in the family is governed by kinship, not need. (2) Institutionalized aid to non-kin (welfare) is dependent on many preconditions, the most important being “deservingness,” what can be best interpreted in terms of reciprocal expectations.

4.1 Need as justice in the family?

Frequently, families are stated as a counter example to reciprocity, since children would not be able to reciprocate (Gouldner 1973; Tyler et al. 1997, 51; Heidenreich 2011, 130), or families would obviously distribute mainly according to equality and/or need (Lerner et al. 1976, 153; Hochschild 1981, 107; Deutsch 1985,  29f., 42f.; Miller 1999, 26). Viewed sociobiologically, it is evident that kin selection rules family life, but a reciprocal vocabulary can also be used.

The basic “utility” of children lies in replicating their parents’ genes, and evolution could hardly select for anything else. Walster et al. (1978, ch. 6) discuss in depth the costs and rewards of the parent–child relationship in terms of Equity Theory, and they cite evidence that a kind of “immortality” is a strong source of parental satisfaction.Footnote 7 It is part of the parental dilemma that children accomplish this “achievement” simply by birth but their genetic interests differ from that of their parents. These genetic parent–child conflicts become evident for example in weaning conflicts and cry-babies (Voland 2013, sec. 4.6). If children do not “cooperate,” parents cannot withdraw their support without risking their parental investment, and parental love has evolved to mitigate this dilemma (Trivers 1972; Daly and Wilson 1999).

This “economic” or “biologistic” wording may sound repulsive (cf. Graeber 2011, 91ff., wrestling with this question), but sociobiology provides overwhelming evidence that it is the biological descendancy that is decisive for parental investment, not the “family” social unit (for the following cf. Voland 2013, ch. 3 & 4, with even more examples, or Daly and Wilson 1999):

  • Frustrated reproductive expectations are probably a main reason for divorces, and in many cultures any “constraints” on reproductive performance are accepted as reasons for divorce. Stepchildren, unsure biological fatherhood, and sick or disabled children are destabilizing factors for marriage.

  • Stepchildren or children with unsure biological fatherhood suffer massively higher risks of accident and death, have higher stress levels and are often materially disadvantaged, compared to biological children.

  • Parents try to massively intervene in decisions of their children, ignoring their “needs,” e.g., with mating prohibitions (“Romeo and Juliet”) or appeals to a “family solidarity.”

Although “Life Course Reciprocity” plays an important role for both directions of intergenerational exchange (Silverstein et al. 2002), parents do more for their children than vice versa during their lifespan, even when not considering childhood and youth phase (Hollstein 2005, 196f.). All this makes clear that in the family “need” is rather a small wave on a deep lake of (hardly quantifiable) reciprocity.

4.2 The role of deservingness for welfare

Modern societies have usually institutionalized aid to non-kin as “welfare,” with full social inclusion as one of the central aims of social policy (see van Oorschot and Roosma 2017, for an overview). This is not self-evident since there is “a social norm against living off other people and a corresponding normative pressure to earn one’s income from work” (Elster 1989, 101).

During the last decades there has been a growing body of literature on the role of “desert” in social welfare, because “the deservingness opinions of various social actors play a pivotal role in the social legitimacy of welfare schemes” (van Oor-schot and Roosma 2017, 4). Especially important is the repeated finding of a “universal dimension of support” for certain welfare schemes: “[T]he rank order of the average deservingness of the groups of ‘the elderly’, ‘the sick and disabled’, ‘the unemployed’ and ‘immigrants’ tends to be the same” in all European countries (cf. van Oorschot and Roosma 2017, 14, 20f., and references therein). The basic conditions for the legitimacy of welfare are that (1) aid as such is restricted to those who “deserve” it, and (2) the extent of aid is restricted to “need,” i.e., objective and legitimate necessities.

van Oorschot (2000) has developed a framework of five criteria (the “CARIN” scheme) that can largely explain differences of the perceived legitimacy of targeting aid to target groups. Even though reciprocity is among the criteria, this is “obvious” reciprocity. But all criteria can serve as proxies for justifying a “very generalized reciprocity” on the part of the donors, and they match with the ideal of justice. They exhibit the properties of good reasons and are well communicable.

  • Control: Fate or fault? If the recipients are personally responsible for their predicament, they are less deserving (if at all). Regarding unemployment, an important parameter is the availability of jobs: A high unemployment rate reduces the personal responsibility for being jobless (Fridberg and Ploug 2000; Jeene et al. 2014). Our interpretation: Violating the norm of personal responsibility reveals a lack of “reciprocal precaution” - or in other words: guilt, and the ideal of justice requires some atonement to prevent thoughtlessness becoming an attractive option.

  • Attitude: Eager or sloppy? Signs of compliance to welfare conditions and to expectations of the donors in general (e.g., gratitude) increase deservingness. Kootstra (2016) found that receivers of welfare who have a long work history and invest great efforts into finding a new job are considered more deserving. Our interpretation: These signs are proxies for social reliability, a will to engage in one’s own contributions and not exploit the situation.

  • Reciprocity: Gift or compensation? During the need situation, reciprocity cannot be expected, but already acquired merits (work history) as well as merits likely in the future increase deservingness, for obvious reasons.

  • Identity: Us or them? The “closer” the recipients are to the donors, the higher the perceived deservingness is. Our interpretation: The paradigmatic example of closeness with regard to welfare is “nation,” and Miller (1995) provided an account of how a “well understood” sense of nationality can contribute to adhering to a nation as an ethic community. A functioning nation is a good precondition for a high level of generalized reciprocity, as opposed to mere fairness, since generalized reciprocity affords a limited group (p. 70ff.).

  • Need: Basic or superfluous? Greater need means more deservingness. Our interpretation: It is always the donors who define what is considered as need, and they restrict aid to what is necessary, i.e., “true” or “objective” needs. Any kind of institutionalized welfare aid has relatively low upper limits (“socio-cultural breadline”), while the avoidance of misery specifies the lower limit of aid. Mulligan (2018, 155ff.) suggested to view aid for the “undeserving poor” not as a requirement of justice but as a minimization of opportunity costs. Extensive checks of available means of the needy are common. This can be understood as limiting the deviations from reciprocity to an acceptable extent.

The notion of deservingness seems to govern the domain of welfare, and desert is used to justify reciprocal expectations and deviations from it. Contrary to Rawls, it is primarily desert that is relevant to the design of at least some social institutions. Need is no desert, but it can (and should) be a case for entitlement. Need has been interpreted as a metaphor for a Mutual Insurance Society (Lucas 1972), the main fear being free-riding, so several precautions and limitations are necessary. Hence, acts of welfare can be interpreted as a “very generalized reciprocity,” a trust-led investment of people into the cohesion of society – their society.

5 Equality as auxiliary principle

This section discusses four topics where the equal treatment of people can be reconciled with reciprocity: (1) Equality can be efficient as a default (or fallback) when lacking good reasons to choose otherwise. (2) Equality of material rewards for individual achievements (income) is restricted to groups of manageable size where “invisible” costs and benefits may set straight the reciprocal calculation. (3) The access to a decision about the distribution of (dis)advantages is equal for all (Equality of access). But the decision itself is basically about merit, and equality of access can be best interpreted as a self-binding. (4) Political equality is primarily the lack of good reasons for inequality (and a self-binding).

5.1 Equality as default

Equality can be a fallback when the relevance of other criteria is not given or not known precisely enough. “With a good reason, accept inequality – with no good reason, strive for equality” is a widely cited rule of thumb (Wollheim and Berlin 1956, 305; Frankena 1962, 5; Benn and Peters 1965, 128; indirectly: Schelling 1960, 65; Mikula 1980b, 140; Buchanan and Mathieu 1986, 15; Elster 1992, 70; Young 1994, 8; Miller 1999, 233; White 2007, 11; Heidenreich 2011, 177; experimentally: Deutsch 1985, ch. 11; Konow 2003, sec. 5.1; dissenting: Kolm 1996, 37; for even more references see Gosepath 2011, sec. 2.4). Generally, justice theory aims at justifying deviations from equality (Elster 1992, 200).

When lacking a good reason, equality is always communicable and difficult to challenge, it “fills the vacuum of indeterminacy” (Schelling 1960, 73). An equal distribution has the additional advantage of not requiring any effort to assess individual contributions and to defend the final distribution (Mikula 1980b, 131) – equality is easy, and E. Walster et al. (1978, 213) provide some efficiency considerations regarding “When Equality? When Proportionality?”: time constraints, communication costs, the value in dispute, or the significance (precedence) for future decisions. But regarding communicability, even the proportionality rule is “equality,” albeit of units of claim (Young 1994, 80).

5.2 Equality in groups

Since the 1970s, and in contrast to Equity Theory, several social psychologists have developed multi-dimensional approaches to justice. They connected distributive principles with types of social relationships in groups and their specific goals, characterized by a tension between equity and equality (cf. Druckman and Wagner 2016, 389f.). For example, Deutsch (1985) contrasted an egalitarian, solidarity-oriented group (focus on social relations) with a meritocratic, economic-oriented system (focus on individual outcomes). Kabanoff (1991) observed equality combined with an emphasis on solidarity, while equity (proportionality) was combined with an emphasis on productivity.

Solidarity should not be confused with altruism. Solidarity is a reciprocal relation with a norm of mutual obligation to contribute one’s share to a common goal (Bayertz 1999). Group dynamics have shown the interdependence of group cohesiveness, performance, and task commitment (Mullen and Copper 1994; Porter and Lilly 1996).

5.2.1 Confluence of equity and equality

Distributions based on merit can coincide with those based on equality when everyone’s contribution (input) to a common output is of equal value (either actually or supposedly), which Leventhal (1976a, 115) called the “confluence of equity and equality norms.” There is a good reason to choose an equal distribution – equality is not a fallback here. But then the principle of distribution is still merit, not equality. Investigating “backward reasoning” from visible rewards to unknown performance, Bierhoff et al. (1986, 181) reported that unequal rewards would activate the equity script, whereas equal rewards would activate both the equality script and the equity script, not the equality script alone.

5.2.2 Allocators and the communicability of the chosen distribution

But it would be a mistake, however, to contrast these two allocation norms too sharply. Leventhal (1976a) discussed in detail the role of an allocator in groups or organizations. Allocators would usually choose a compromise satisfying both norms partially (p. 114; see also Scott et al. 2001). They must balance several dimensions, among them the productivity of the group, avoidance of conflict, and their own approval and acceptance by the group members (authority).

The chosen allocation rule is strongly contingent on communicability considerations: when allocators can conceal their allocation decision to the group members (secrecy), they are inclined to follow the equity norm, for reasons of justice, productivity, and reduced group conflict due to a lack of social comparison (Leventhal 1976a, 111). But when group members are strongly interdependent, when they perceive themselves as “similar,” when their success can only be credited to the whole group, and when group productivity can be fostered by keeping up a “solidary spirit” within the group, several authors reported a preference for an equal distribution (Leventhal 1976a, 108ff.; Greenberg 1982, 418ff.; Deutsch 1985, 147).

5.2.3 Group members and the communicability of costs and benefits

Social exchange includes tangible goods or services, but also less “concrete” resources like status or information (Resource Theory, Foa and Foa 2012). Less concrete does not mean less valuable or not targeted at material gains – status is a means to enhance fitness in various ways.

Long-term cooperation is contingent on many circumstances, so it can be advantageous for group members to forego immediate benefits. Mikula (1980b) and Schwinger (1980) reported that in groups a “politeness ritual” would prevail when the group members themselves were to propose a distribution rule: High performers tended toward an equality rule and low performers toward a proportionality rule, both seemingly at odds with their material interests. Yet Bierhoff et al. (1986) cited several studies where high performers tended toward more equal distributions if future interactions could be expected but to equity otherwise. Corgnet et al. (2011) reported that social motives to shift the allocation decision toward equality can be quite strong and even result in inefficient teams.

Choosing equality here can be regarded as a self-binding (investment) since reciprocity by definition depends on repeated interaction. Social harmony, less conflict, and a high degree of group cohesion are less visible benefits, justifying visible costs.

5.2.4 The case of Israeli kibbutzim

Kibbutzim are voluntary rural communities in Israel that share an egalitarian ideal. They are characterized by communal living with equal sharing of outcomes, no private property, and no cash exchange inside the kibbutz. But they face largely three threats, as predicted by economic theory and empirically verified (Abramitzky 2011): the exit of more productive members (brain-drain), the entry of less productive members (adverse selection), and shirking in the communal work (free-riding). In effect, also kibbutzim must set straight the reciprocal calculation for each member, be it a high or low performer.

5.3 Equality of access

Application processes are generally characterized by an equality of access. It would contradict the idea of an “application” when applicants with a prima facie claim on the good being allocated are treated differently before the decision process (cf. Miller 1999, ch. 5). But the decision is basically about merit.

Equality before the law primarily affects the ability to be indicted, the possibility to “apply” for a trial, no matter what status or “deserts” one has. During the trial, procedural justice requires all defendants to be treated impartially, but substantial justice requires the opposite. The exact purpose of a trial is to determine objectively the extent of guilt and to find out what the just deserts are.

Likewise with political offices. The goal of every election of representatives is to bring only those in offices who are suitable, not anyone. But we admit everyone to be elected, again for reasons of self-binding, and merits are assessed afterwards according to results. If in doubt the person will be unelected on the next occasion. But if there is a candidate of whom we know that she is the most suitable, she will certainly win the election, yet not according to criteria determined in advance, but based on a collective assessment.

The same holds true for all other applications, be it a new job or a place at university – equality is equality of opportunity, the rest is about merit, but this will be assessed afterwards. It is a common misunderstanding made in philosophical debates on justice when asking “Who deserves a job?” (cf. Miller 1999, ch. 8: “Deserving Jobs”): already the hiring is regarded as a reward. But the rewards only come later, as monthly payment for achievements or as high grades for good exams, i.e., according to merit. The successful application as a chance of a true test can be viewed as a reward in a factual sense, but not in the systematic sense discussed here.

5.4 Political equality

Suffrage or membership in political bodies has become ever more inclusive because reasons for inequality that revealed as dysfunctional in a liberal society had to be dropped. In fact, no reason has remained except a (contested) minimum age, and the self-binding of equal dignity and respect for all humans is now widely accepted as a minimal standard (cf. Gosepath 2011, sec. 2.3). Most societies would profit if only “thoughtful and well-educated” people could vote (noocracy) yet try to operationalize that. Already the idea seems absurd, and especially when placing people in positions where they are to expend other people’s money (e.g., taxes for public expenditure) or decide over access to advantages we must be very cautious with conflicting interests. It could put the fox in charge of the henhouse, though, when giving the wealthy a higher share of votes, and the same holds true for all other reasons discussed. “One person, one vote” therefore might not be rooted in the idea of equality as such, but in an effective self-binding on a highly sensitive field.

6 How to implement the meritocratic principle

6.1 From moral will to institutions

How is a Theory of Justice successfully implemented? Traditional moral philosophy and social psychology (as well as common sense) focus on individual actions, assuming a duality of “genuinely moral motives” and “all other motives” (cf. Homann 2014, critically discussing this discourse; Montada and Maes 2016, regarding the “justice motive”; cf. also Trivers 1971, 51, and Sect. 2.4). Moral weakness is a weakness of moral will, the answer being “moral rearmament,” i.e., developing better moral reasoning to resist temptations. Ulrich (2016, 451), for example, contrasts economy and morality, and morality would require a breach of economic logic.

Homann (2003, 2014) is a critic of such a dualistic opposition: Moral behavior must be advantageous for the individual if he or she is not to be exploited by others who are less moral and would initiate a moral downward spiral. Any successful implementation of ethical rules therefore must be compatible with economic incentives. Homann (2003, 226) warned against regarding implementation “only” as a practical problem: implementability precedes moral validity. An ethic that presents unrealistic challenges to its participants is not impractical but unethical.

Brennan and Buchanan (1985) made a distinction between choices within rules (choosing actions) and choices of rules (designing constitution and laws). The choices of rules must be moral as to enable people to be guided in their actions only by economic considerations, since economic profits are a valuable indicator of sources of societal wealth (Homann 2014, 99). It is only this compatibility of sanctioned rules and economic incentives that makes the system inherently stable. The most important rule is to guarantee fair competition in market economies (Eucken 1992; Homann 2003; Mankiw and Taylor 2011). According to economic theory, competition would push income levels toward economic contribution, i.e., according to merit.

6.2 Concentrating on “non-merit”: tackling economic rents

There has been a long-standing wrangling over the meaning of merit. Participants in a study by Neckel et al. (2008, 45) found it easier to agree on what merit is not. This can make the exclusion of non-merit an effective implementation of the meritocratic principle, as suggested by Richters and Siemoneit (2019, 134) and Mulligan (2018, 131). From an Aristotelian perspective, we should not struggle with the blurred mean between extremes but concentrate on the extremes themselves.

In economics, undeserved incomes without a corresponding creation of wealth are called “economic rents.” Thus, justice in modern, democratic market societies would first and foremost mean to institutionally drain the wellsprings of such economic rents. Richters and Siemoneit (2021) extensively discuss what in their view are two important sources of economic rents: (1) income shares which are substantially based on resource-intensive technologies, and (2) land rents, where a value generated by society is sold on private account. Appropriate institutions are caps on the extraction or the import of non-renewable natural resources and a land value tax. Without economic rents, market economies could come closer to the goal of just self-regulation (Richters and Siemoneit 2019b). Further measures (meritocratic taxation and equal opportunities) are discussed in detail by Mulligan (2018).

7 Conclusions

7.1 Theoretical unification, semantic pluralism

We have argued that a reciprocal balance can be regarded as the normative ideal of justice, and that an overwhelming, though often inconspicuous evidence indicates that reciprocity indeed governs all social relations – with the caveat that the term reciprocity seems to be an inappropriate framing for a substantial number of social exchange situations. It should be possible to conceptualize justice in a single framework theoretically but communicability seems to require several distinct principles. The more generalized and the less obvious reciprocity is, the more difficult it is to argue that certain actions are still part of an individual maximization of utility in the long run and would lead to material benefits.

In practical terms, need and equality can still be treated as independent principles of justice (allocation rules). But we must take seriously the sociobiological account that altruism in its selfless interpretation could not have been selected for in the evolutionary process. As long as costs do not clearly exceed benefits we should interpret altruistic acts as “trust-led investments” with a questionable profitability – hence, assuming reciprocity. With reference to Sahlins (1965) we have suggested to view such acts as “very generalized reciprocity.” We do not need to refer to “genuine” altruism to explain one-sided contributions for others or for the common good. Those contributions are usually conditional and benefit the actors, at least on the average, in the long run.

Fig. 3
figure 3

The Practice of Social Justice: Achievements and income serve as proxies for desert and reward, but our fellow citizens should not fall below the socio-cultural breadline (left edge). When individual efforts in groups are roughly comparable, their members may not argue about subtleties but prefer equality (dotted circle)

In practice, justice is a social bargain and an optimization problem, with local deviations from proportionality as depicted in Fig. 3. If we assume that the central goal is utility maximization, then balancing reciprocal expectations becomes only one goal among others that affect costs and benefits. Often maximizing utility means minimizing (social, invisible) costs rather than maximizing (material, visible) benefits.

One could ask: When we still have several principles of justice in practice – what is the use of singling out merit as the foundational one? Our answer is: As long as need and equality are perceived as competing principles, the meritocratic principle lacks persuasiveness, which severely limits its political applicability. Economic rents are a broad field in economics (empirically and theoretically) and offer plenty of politically promising starting points for more justice. A proper distribution in the first place (fair incomes) should be preferred to redistribution, for minimizing social conflict and making better use of markets. In modern, anonymous societies, we do not need people with higher moral motives but better institutions to ensure this.

7.2 The egalitarian spirit of merit

As mentioned in Sect. 2.5, merit is criticized for justifying inequalities, and inequality is often identified with unfairness. But in practice people prefer fairness to equality (Starmans et al. 2017). Merit has a non-negotiable core, especially when time or quantities are the objective yardstick of achievement or when some people achieve results that others cannot, no matter how much effort they invest. They deserve higher wealth. But merit has also a deeper social meaning. Merit is the normative basis of the division of labor since nobody can achieve unboundedly (this is why we divide labor). Effort (i.e., hours spent on work) is an income-limiting factor, because only in cooperation with others can the talents of the few unfold and create specialists who deserve higher wealth. Their productivity is contingent on their social environment (team, firm, society, ...) – hence, how is “high performance” to be attributed? From the collective perspective, endless accumulation is (and always has been) a social misunderstanding (though not from the individual perspective of course). The meritocratic principle has a strong egalitarian element that is often not allowed for in the debate. If we focus politically on tackling economic rents, then the meritocratic principle could lead to a more egalitarian society than John Rawls could have ever imagined.