Introduction

It is often argued that relational egalitarianism has a fatal problem with intergenerational justice. Since the lives of people that belong to different generations do not overlap, they cannot be in egalitarian social relationships (Quong 2018, pp. 317–318; Lippert-Rasmussen 2018, pp. 123–124). Because of this, relational egalitarianism is considered to have a disadvantage with respect to intergenerational justice in comparison to other normative theories such as utilitarianism, prioritarianism, and luck egalitarianism. However, in a recent article Timothy Sommers argues that there is no such comparative disadvantage for relational egalitarianism vis-à-vis intergenerational justice (Sommers 2023).

Sommer’s argument is quite modest. Rather than trying to show that relational egalitarianism can deal with the non-identity problem in its application to intergenerational justice better—that is, in a more plausible way—than competing normative theories, he only aims to reject the claim that there can be no way to extend relational egalitarianism to intergenerational justice (2023, p. 473). I will call this the ‘No Comparative Disadvantage Thesis’. In this paper, I want to challenge the No Comparative Disadvantage Thesis in two ways. First, I will show that Sommers fails to provide a reasonable constraint on causal efficacy, which is crucial to his argument for our relation (not relationship) to future generations. Second, I will show that the chain-relatedness problem casts a shadow over Sommers’s argument for the No Comparative Disadvantage Thesis. Based on this, I conclude that the No Comparative Disadvantage Thesis does not hold much promise.

The No Comparative Disadvantage Thesis

Sommers’s argument in support of the claim that relational egalitarianism can be extended to intergenerational justice (at least to the extent that other normative theories can so be extended) involves two steps.

First, he proposes to understand the concept of ‘overlap’ in the following way:

[W]e can be socially related in causally efficacious ways with future generations, for example, by saving resources that allow them to follow through on measures we initially plan, and they then develop further, to slow climate change. So, our lives could hypothetically overlap more with the lives of people that we do not overlap with temporally, than with the lives of the people that we do overlap with temporally, but do not interact. (2023, p. 475; emphasis original)

Sommers suggests here that the causal efficacy of our (possible) actions can create a greater overlap of our lives with the lives of other people, even if we and they do not interact directly. In other words, direct interaction is not a requirement. Next, he introduces a distinction between relationships and relations. Relationships are exemplified by direct and reciprocal interactions between specific individuals, such as close interactions within in-groups.Footnote 1 In contrast, relations do not require reciprocity and specificity in this sense. Social relations are typically based on enduring social connections without direct interaction, and are often unidirectional (and thus not reciprocal), as is commonly the case in contemporary society. A pay-as-you-go pension system is an example of a kind of institution that constitutes a social relation between generations by virtue of a (unidirectional) causal interaction between them (2023, pp. 474–476).

In the second step, Sommers argues that relational egalitarianism can be understood as a theory of justice for nonhierarchical, ongoing social institutions. He appeals to the neo-republican argument against hierarchical domination by a generous master (in which, due to institutionalization of the hierarchical structure, hierarchical social relations survive the death of the master) to show that, while hierarchical intergenerational social relations are possible, it is more plausible to describe better intergenerational relations as nonhierarchical and egalitarian. Thus, relational egalitarianism, including its support for egalitarian social relations with future others, is likely to succeed, at least to the same extent that other normative theories are likely to succeed (2023, pp. 477–479), and consequently, Sommers concludes that the No Comparative Disadvantage Thesis is tenable.

The Importance of a Reasonable Constraint on Causal Efficacy

Let me first show that Sommers fails to provide a reasonable constraint on causal efficacy. This is crucial to his argument so that our relation to future generations can be regarded as a social one. As seen above, the relation holds in terms of causal efficacy. According to Sommers (2023, p. 474; emphasis original), only ‘less overlap is required for social relations, than for relationships’ and the forms of overlapping socially ‘do depend on a causal relationship’. In the case of family relationships, for example, there is an obvious causal connection between parents and their children, and thus, causal efficacy is evident in such a relationship. Social institutions are causally related to individuals in society in some way, but even though there are no direct interactions between institutions and individuals, this causal relation is not negligible. Hence, Sommers’s claim seems plausible—in the same way that there are certain causal relations in contemporary society, there could also be some causal relations between generations; existing people can causally influence future generations. For example, our policy choices could causally contribute to a certain future climate (perhaps in a conditional probabilistic sense).

However, such influenceability on future generations within hypothetically constructed causal relations involves an absurdly large number of possible future worlds. This is problematic for the conception of relation at issue because most (if not all) groupings could then be seen as creating relations by identifying causal connections. Even groups with no social salience could be considered to constitute a relation; for example, any contingent contact with extraterrestrials in outer space would be considered to represent a relation between us and the extraterrestrials.Footnote 2 Hence, some constraint is needed to prevent the inclusion of such possibilities (possible future worlds), but what kind of constraint should it be? In philosophical debates about counterfactual dependence, a commonly used constraint is the ‘nearest’ clause: only the nearest possible worlds matter. Such a constraint may exclude possible worlds in which people can make choices that would disrupt a law of nature, for example (Lewis 1986, pp. 32–52). However, this is not an appropriate constraint to limit the number of relevant possible future worlds in Sommers’s argument, because it does not identify possible future generations that could be affected by our actions and policies in a relationally significant way. The causal relation between us and them must be sufficiently similar (though not necessarily identical) to the causal relations between social institutions and people in the present society to be able to claim that our present lives overlap with the lives of future people.

Hence, the constraint on causal efficacy must be relevant to the identification of social relations between the present generation and future generations. This raises the question: Does Sommers’s argument provide an explanation for a relevant constraint on causal efficacy? As I see it, it does not; Sommers simply takes the United States as an example of how social relations can obtain without social interactions (2023, p. 475). In other words, the constraint in question is merely assumed in an extrapolative manner. But this extrapolative constraint raises doubts about its relevance as implying a constraint on future possible worlds, because what is inferred from facts about some particular existing society may not necessarily be true about social relations in general in the future.Footnote 3

It could be argued that the nomological constraint on counterfactual dependence is sufficient for an appeal to causal efficacy such that current and future generations can be socially related. This is because one could say that future people are (better) able to imagine the better outcomes of certain past decisions in subsequent possible worlds where the laws of nature are given, even though they cannot themselves change the present or past. This, one might argue, can be taken as a case of social relations. However, since this strategy is bullet-biting and thus cannot rely on the plausibility of the proposed argument, it requires a full justification.Footnote 4 I would add that there is no hint of such a justification in Sommers’s argument.

The Chain-Relatedness Problem and Specific Principles for Duties to Future People

Now suppose, for the sake of argument, that Sommers’s argument includes a proper constraint on causal efficacy. Can we then say that relational egalitarianism can be expected to provide a particular principle for duties to future others based on the existence of social relations with future generations, in the same way that other normative theories can be expected to provide specific principles for their applications of intergenerational justice? My answer is no, and I will show why.

Sommers argues that, just as relations without reciprocal and specific interactions with others can persist in a relatively durable way over time, so too can relations of the same kind persist over generations. As seen above, his particular argument draws on the neo-republican argument against hierarchical domination to show that it is more plausible to conceive of intergenerational justice as involving non-hierarchical, ongoing social institutions than hierarchical ones. Sommers believes that this allows us to see the viability of the application of relational egalitarianism to intergenerational justice. In this way, relational egalitarianism, including its support for social egalitarian relations with future others, is likely to succeed, at least to the extent that other egalitarian theories are likely to succeed (2023, pp. 477–479). Sommers thus argues in support of the No Comparative Disadvantage Thesis.

This argument can be challenged, however. Even with the proper causal efficacy constraint, there must be a reasonable condition (along with the causal criteria) under which intergenerational relations can be regarded as social, which Sommers does not offer. This condition seems difficult to propose within the theory of relational egalitarianism in such a way as to find the theoretical plausibility of intergenerational justice that involves nonhierarchical, ongoing social institutions. Consider, for example, the case of people in a given population (Sommers takes the United States as an example), who are socially related to their descendants even though they died before their children were born. This chain extends not only back in time for 200,000 years, but also forward into an unknown future. The obvious problem with ‘chain-relatedness’ is that the importance of social relations of equality for justice diminishes as the chain becomes longer and looser.Footnote 5 More problematic for Sommers’s argument is that, under chain-relatedness, the plausibility of intergenerational justice as nonhierarchical can be questioned; at the very least, we are not sure that its plausibility extends promisingly into remote antiquity, such as the primitive age in which the earliest hunter-gatherers live, and into the distant future in which technology might fundamentally change the way of political life.

One might argue that the shared aim of establishing and maintaining institutions and mutual responses (in some indirect way through documents such as the constitution) can be taken as a reasonable constraint on social relations across generations. Indeed, Scheffler (2013, 2018) and his followers (e.g., Karnein 2022) support this line of argument in such a way that humanity (human flourishing) gives us a reason to engage in responsible activities in order to avoid human extinction after our death; the extinction in question frustrates our activities to generate benefits in the present and, thus, is intrinsically bad for us. However, this cannot be a reasonable constraint on intergenerational social relations because the shared aim of humanity (human flourishing) is too general, i.e., it is not specified in a way that is relevant to social relations of equality.

Let me clarify this point and then show why the No Comparative Disadvantage Thesis is untenable. Relational egalitarians appeal to particular principles, such as the guarantee of equal basic capabilities, that create and secure egalitarian relations among citizens (e.g., Anderson 1999). Notably, this relational egalitarian appeal to the sufficientarian principle is grounded not simply in humanity or for human flourishing, but rather in people’s current activities as citizens in civil society, including equal participation in governmental affairs and the economy (Anderson 1999, p. 317). It is then not clear whether relational egalitarians can legitimately apply this contextually justified principle across generations; the relevance of the sufficientarian principle has a strong bearing on our current practice. In contrast, other theories of justice, such as utilitarianism, prioritarianism, and luck egalitarianism, can avoid this problem. These theories can appeal to specific principles for duties to future others based simply on considerations of their (weighted) disutility or (the effects of) brute bad luck on them. These principles can play a role in valuing the well-being of future people.Footnote 6 Relational egalitarianism, on the other hand, cannot appeal to the principle of justice across generations in the way that these normative theories can do, because of the sensitivity of its sufficientarian principle to the current context and practice. This means that we cannot be sure that the sufficientarian principle of relational egalitarianism as a principle of justice is extendable to the future (as well as to the past). Sommers provides no argument for how relational egalitarians can overcome the problem in question; he says nothing about a reasonable condition for being social (other than the condition of causal efficacy).

Thus, Sommers’s argument fails to show convincingly that relational egalitarianism is likely to succeed as a theory of intergenerational justice, at least to the extent that other normative theories do. I conclude that the No Comparative Disadvantage Thesis is not promising.