A Working Definition of Moral Progress

Abstract

Essentially everyone agrees that the outlawing of slavery, or the beginning of women’s suffrage, or the defeat of Nazism constitute paradigmatic examples of moral progress in human history. But this consensus belies a deep division about the nature of moral progress more generally, a consequence of the foundational differences among and within normative traditions regarding the nature and scope of the ‘moral’ in moral progress. This essay proposes that philosophers might nonetheless converge on a working definition of moral progress by identifying a proxy property that reliably tracks moral progress, but which does not purport to be coextensive with the philosophically-relevant property. The aim of this essay is to identify this proxy property with emerging empirical measures of population welfare, and to show why this indicator of moral progress can garner overlapping consensus from a variety of normative traditions.

1 A Working Definition of Moral Progress

Perhaps because the state of the world often seems especially dismal from the perspective of the present, or simply because philosophers tend to emphasize the negative aspects of morality, contemporary philosophers have paid relatively little attention to the issue of moral progress in human history (for some notable exceptions, see Singer 1981; Wellman 1999; Moody-Adams, 1999; Rorty 2007). The situation has begun to change in recent years, spurred by recent claims from social scientists like Steven Pinker (2011) and Michael Shermer (2015) who point to particular socio-historical developments as indicative of a positive moral trajectory across human populations. The happy effect is that philosophers have returned attention to a seemingly central question in normative ethics, namely, what are the conditions under which a human community improves morally?¹ An answer to this question would not only constitute a significant philosophical achievement, but it would go some way toward strengthening the often tenuous relationship between the theoretical and practical spheres of ethical inquiry. After all, one of the central measures of good social or public policy, if not its sole measure, is the extent to which that policy contributes to a (morally) superior world. So to the extent we have a grasp on the indicators of moral progress, we have a practical tool for evaluating the merits of the social practices and institutions that concern applied ethicists and policy-makers. If we imagine the diversity of human populations as constituting a large-scale experiment in human social living, we might naturally want some normative measure of what appears to be working or failing. Ideally, we want to be able to assess the cultural differences between populations to see what kinds of social, economic, and political conditions are implicated with moral development. Having a theoretically-neutral, working definition of moral progress would be a small step in that direction.

But if the history of philosophy is any indication, the likelihood of converging on a shared conception of moral progress is vanishingly small, and not just because the problems is hard, which it is, but because different normative traditions do not even agree about the nature and scope of the moral. For example, we can imagine a broadly consequentialist conception of moral progress conceived as the trajectory toward a human community with a greater distribution of intrinsic goods (welfare, pleasure, etc.). Alternatively, we can imagine a deontological conception of moral progress as the trajectory toward a community where agents more reliably discharge their duties, or a character-based conception of moral progress as the advancement toward a world populated with progressively virtuous characters. Of course, these are just some of the ways of conceiving moral progress; moral progress could consist in a wide range of increases in the relevant right-making and/or good-making properties. Given the trenchant theoretical divides regarding the nature and scope of the moral, both within and among normative traditions, we have little reason to hold out hope for a theoretical consensus on a definition of moral progress.

Fortunately, we do not need an analysis of moral progress. For many of our purposes, it would be sufficient to specify a working definition of moral progress by identifying a ‘proxy property,’ a reliable correlate of moral progress that could garner overlapping consensus by most of the major normative traditions. This proxy property – which I identify with empirical measures of population welfare – would not purport to be coextensive with or identical to the philosophically-relevant property. It would nonetheless serve as a reliable predictor of that property in the actual world. If the correlation is sufficiently reliable, claims involving the proxy property can provide the basis of knowledge for claims concerning moral progress.

Consider an analogous strategy in the philosophy of mind. Philosophers routinely disagree about whether mental states, like pain, can be analyzed solely in terms of some underlying neural states. Nonetheless, physicalists and dualists overwhelmingly agree that, in the actual world, the pain property is very reliably correlated with certain neural states. They may continue to disagree about whether those neural states necessarily constitute pain, while agreeing that wherever we observe those neural states in the actual world, we are extremely likely to find the pain property there, as well. In short, neural states can and do function as reliable predictors for the metaphysically-relevant property, particularly in practical domains like medicine, which is one reason why dualists are not skeptics about anesthesia. Similarly, normative theorists might well agree that rising mean population welfare is a reliable indicator of moral progress in the actual world without committing to the idea that the presence of the relevant indicator constitutes moral progress.

For the purpose of this project, which aims at concordance, we will understand (global,² population-level) moral progress in the most expansive terms, as the trajectory toward a human community that is more like the kind we ought to bring about. This reflects the fact that philosophical interest in moral progress is tied closely to the Enlightenment ideal that saw ethical theory as providing a foundation for constructing a better world for human beings to live together. The hope then is to identity some reliable, theoretically-neutral indicator for when populations are moving in the right direction.

But to be clear, a working definition of moral progress will not necessarily put us in a position to label some specific historical trend as an instance of moral progress or moral decline. It is certainly possible for measures of population welfare to increase for reasons that are unrelated to moral progress (e.g., a disease that wipes out the frail), or through unsavory means (e.g., killing all the unhappy people). Conversely, a society may develop morally without it necessarily being reflected in our best empirical measures of human welfare. For example, race and gender equality has improved in the United States over the last several decades while happiness levels have remained relatively stable by some measures.³ In some cases, there may be too many variables to make a judgment about specific populations over relatively short time periods, though I will suggest that some of these concerns can be addressed by the development of multi-criteria measures of population welfare that can provide a more fine-grained picture of the correlates of moral progress. In any case, these concerns fade when we turn our attention to patterns of rising population welfare across human populations. When we look across populations, we observe a reliable, positive relationship between, for example, social/political equality and rising mean welfare. These patterns of rising welfare across human populations can serve as reliable indicators of the kinds of social and political institutions that conduce to an upward moral trajectory.

The strategy in this essay will proceed as follows: first, I will identify the proxy property by arguing for a correlation between moral progress and a sufficiently enduring⁴ increase in mean population welfare. Second, I will argue that this indirect measure of moral progress is amenable to a wide range of normative traditions despite its apparent consequentialist leanings. Finally, I consider some objections and conclude with some brief thoughts about the applicability of a working definition of moral progress for an interdisciplinary research program that connects applied ethics and public policy.

2 Measuring Population Welfare

Since the 1950s, psychologists have aimed to measure a central feature of human mental life, namely our perception of how well our lives are going for us. Typically called ‘subjective well-being’ (SWB) or somewhat misleadingly, ‘happiness’, the construct is meant to measure our cognitive and affective appraisals of our own welfare (for reviews of the literature, see Diener et al. 1999; Kahneman et al. 1999).⁵ Increasingly, theorists are interested in aggregating this data to provide a picture of population welfare. Population welfare is typically described in terms of the mean individual welfare in a population, where the population is understood straightforwardly as the group of individuals living and interbreeding within a geographically restricted area.⁶

Population-level measures of human welfare are rapidly becoming one of the most influential psychological constructs in contemporary public policy, having been adopted by an increasing number of major government and public policy organizations. For example, population welfare measures are now a standard part of the United Kingdom’s and Australia’s national assessments, and they are central evaluative tools used by the World Health Organization and the United Nations in assessing international welfare. Even the US military has begun to measure welfare in some of its populations, as have some multinational corporations, like IBM, who have begun to include such measures in their assessment of corporate welfare. The relatively rapid adoption of welfare measures by a wide variety of international organizations is a testament to a growing consensus that these indicators tell us something extremely important about a human community that is relevant to making good public policy (Forgeard et al. 2011; Oishi and Schimmack 2010).

2.1 Three Approaches to Measuring Population Welfare

There are three broad frameworks currently used to measure population welfare: a cognitive approach that focuses on individuals’ reflective judgments about the state of their life as a whole, an affect-oriented approach that focuses on individuals’ emotional appraisals of their life experiences, and a eudaimonic approach that measures individuals’ capacities to fulfill certain natural functions (e.g., their access to education or sufficient levels of personal autonomy).

The first two approaches involve measures of ‘subjective well-being’ that measure how individuals subjectively assess their own lives as a whole and/or the quality of their experiences over time. The cognitive camp relies on individuals’ reflective, deliberative judgments of their lives holistically or within major life domains, while the affective camp asks individuals to evaluate their experiential valence in terms of their emotions and moods. The affect-oriented theorists measure subjective well-being by aggregating emotional evaluations throughout a given day (Csikszentmihalyi 1990; Stone and Shiffman 1994), or over a completed day (Kahneman et al. 2004), or in recollection of the past two weeks (Diener et al. 2010a, b). The cognitive camp measures subjective well-being by soliciting a global assessment of one’s overall life circumstances, as in the influential ‘Satisfaction with Life Scale’ (SWLS) developed by Ed Diener and his colleagues (Diener et al. 1985). This scale asks individuals to rate five statements concerning their present satisfaction with life on a seven-point scale (e.g., “so far I have gotten the important things I want in life.”) A closely related measure is the ‘Cantril Ladder’ (Cantril 1965) in which people are asked to place themselves on a scale of 10 to 0, where the top of the ladder represents the best possible life for you and the bottom represents the worst possible life for you.

Both of the subjective approaches to human welfare rely heavily on self-reports, a notably imperfect strategy since we know that individuals can be prone to a variety of mood effects, recollection biases, situational pressures, and adaptation effects (Oishi et al. 2011). Still, the thrust of several decades of happiness research suggests that these effects are relatively minor. Most subjective instruments show impressive levels of internal consistency, temporal stability, convergence with non-self-report measures (such as facial expressions and physiological measures), and a sufficient degree of criterion validity (Lucas and Diener 2010 pg. 473).⁷ Self-reported measures of happiness also align surprisingly well with expert assessments of individuals, and the assessment of one’s close friends and associates, who sometimes know us better than ourselves (see Sandvik et al. 1993).⁸ Moreover, happiness levels fluctuate in precisely the ways we would expect: decreasing with experiences of sickness, isolation, and fear, and increasing with sociality, engagement, and autonomy (Headey and Wearing 1989; Dijkers 1997; Putzke et al. 2002; Lucas et al. 2004; Lucas 2007). Finally, when the object of scrutiny is partly defined by its subjective character (as is arguably the case with well-being), self-report measures may be a particularly appropriate strategy. Similarly, the best way to assess how much pain people are in is simply to ask them how much pain they are in. It is not a perfect system, but it is reliable enough that these self-reports are precisely what doctors use to assess the proper dosage of pain medication.

Life-satisfaction surveys are some of the most widely-used assessment tools by policy professionals, given the ease with which they can be administered and quantified, but they have also attracted the brunt of criticism from philosophers and psychologists. Dan Haybron (2007, 2011) is an example of the former who has expressed skepticism about global satisfaction surveys on the basis that such assessments are too cognitively demanding and too easily biased by current circumstances, such as the presence of a disabled person (Schwarz and Strack 1999). But subsequent research in this area suggests that individuals do often rely on chronically accessible elements of their life circumstances when assessing global satisfaction (Schimmack, Diener, and Oishi 2002), which helps explain the fact that global life satisfaction assessments remain relatively constant over an individual’s life, typically fluctuating only with genuinely significant changes in one’s life circumstances, and not, for example, with changes in the weather (Lucas and Lawless 2013). Nevertheless, such worries point to the merits of combining life-satisfaction measures with other ways of measuring population welfare. For example, the Gallup World Poll, which collects data from over 100,000 respondents across 105 nations uses a combination of life satisfaction surveys and affect-based assessments.

Some thinkers like Dan Hausman (2015) are skeptical of subjective welfare measurements altogether, arguing that well-being is too personal and idiosyncratic to be measurable. On his view, we cannot rely on subjective evaluations because the impact of the quality of subjective experience for an individual’s well-being is a deeply personal matter. For example, a broken ankle may be subjectively good or bad for my well-being depending on my personal circumstances and preferences. If I am a couch potato, a broken ankle may give me a reason to do precisely what I want, but if I am an athlete, it thwarts my deeply held goals. But Hausman’s views controversially assumes that there are no objective indicators of human well-being (such as the ability to engage in physical activity) that do not depend on how an individual subjectively assesses their value. We will return to objective measures of human welfare below. Moreover, Hausman’s critiques merely suggests that subjective measures of well-being are difficult in individual cases (which no sensible person would deny), but it does not rule out the possibility that subjective measures give us reliable information about the general case, which is what policy measures aim to do. On the whole, broken ankles do not tend to contribute to human welfare, and such generalized principles are the kinds that underlie good policy (see Alexandrova 2016 for a further discussion of Hausman’s skepticism).

2.2 Objective Indicators of Population Welfare

For some of the reasons described above, some theorists prefer objective measures of human welfare that do not depend on the subjective assessments of individuals. This third way of measuring population welfare typically takes a ‘eudaimonic’ approach in that it involves assessing the capabilities of individuals to fulfill basic human functions. The most influential eudaimonic measure is based on Amartya Sen’s and Martha Nussbaum’s ‘capability approach’ (Sen 1999; Nussbaum 2000) that focuses on identifying the particular set of needs required for complete human functioning. The capability approach proposes that the capacity to achieve human happiness is determined by what people are able to do to be, and thus the kind of lives they are actually able to lead. This research was the inspiration behind the Human Development Index (HDI) developed by the United Nations, which assesses well-being along three dimensions: longevity and health, access to knowledge, and a decent standard of living.⁹

The eudaimonic camp is often critical of the subjective well-being measures above for relying on what they consider unreliable information from respondents. Of particular concern are ‘adaptation effects’ which purport to show that individuals readily adapt to life circumstances, such as poverty, in ways that seem to misrepresent the impact of changes to one’s welfare conditions (Frederick and Loewenstein 1999). Subjective assessments may thus misconstrue the extent to which individuals have become inured to misery or exploitation. But subsequent research has suggested that adaptation effects, while significant, are not as strong as once thought. Moreover, affective experiences may show stronger adaptation effects than life satisfaction measures (see, e.g., Lucas 2007). In fact, a recent study of 54,000 Germans showed little evidence of adaptation within a spell of poverty (Clark et al. 2015). Instead, they found that poverty is bad, and remains bad, with regard to reported life satisfaction. Nonetheless, adaptation concerns highlight the merits of combining subjective assessments of population welfare with the kind of objective measures noted above (for a recent discussion about the relationship between objective and subjective measures of population welfare, see Jayawickreme and Pawelski (2013).

2.3 Multi-Criteria Indicators of Population Welfare

Some theorists and policy organizations are already taking a multi-criteria approach that combines objective and subjective measures of population welfare. One composite measure that shows significant promise is the ‘dashboard approach’ pioneered by Juliet Michaelson et al. (2009) with the New Economics Foundation. Their welfare measure is decomposed into nine components that include positive feelings, absence of negative feelings, life satisfaction, vitality, resilience, self-esteem, positive functioning, supportive relationships, as well as trust and belonging. One theoretical virtue of this multi-criteria approach is that it strikes a balance between many of the competing philosophical traditions that motivate the other frameworks. For example, it strikes a balance between the subjectivist and objectivist traditions by soliciting both subjective assessments of experiential valence in addition to objective measures of welfare, such as the existence of supportive relationships. It also strikes a balance between the hedonic and eudaimonic traditions by including subjective assessments of experiential valence (positive and negative emotions) in addition to objective measures of human functioning. Of course, one drawback to the multi-criteria approach (as opposed to a single criterion) is that it makes it somewhat more difficult to make simple comparisons between populations or time periods. But these problems are not insurmountable; we can still identify quantifiable changes in core areas that determine human welfare (see Huppert et al. 2009, for such an attempt).¹⁰

These measures are not perfect, but a multi-criteria approach represents one of the most promising avenues for developing a reliable measure of population welfare.¹¹ Even now, a concerted interdisciplinary effort is underway that includes philosophers, psychologists and economists aiming to refine current welfare measures. Importantly then, I intend my thesis regarding the correlation between moral progress in population welfare to involve our best empirical measures as they emerge from this interdisciplinary process. However, that is not to point to some distant or final science, but rather to a rapidly maturing research program with wide cross-disciplinary consensus and nearly a half-century of observational and correlational data. In the end, it should not be this surprising to learn that when the people in a society are doing increasingly well, it is a mark of a society that is doing well by its people. And despite the skepticism that is natural to philosophers, the truth is that we are getting remarkably good at measuring changes in human welfare in a population. These emerging empirical indicators mark a place where we might generate overlapping consensus among normative traditions on what constitutes genuine social progress, and we can use that consensus to build a stronger bridge between normative theory and the quantitative constructs that inform contemporary public policy.

3 SWB and Moral Progress

Why should anyone, save a certain sort of welfare utilitarian, believe that increasing population welfare tracks moral progress? After all, presumably a population could get happier without becoming more ethical – perhaps specifically by becoming less ethical. On the Freudian view, for example, the demands of morality serve as an ever-present constraint on individuals’ happiness (Freud and Strachey 1930/2005).

The response is simply that the empirical evidence does not bear out that view of human psychology. In fact, this is a rare occasion on which psychology offers genuinely optimistic news about our nature. So while there is a popular conception of happiness as a largely selfish state, the evidence suggests that nearly the opposite is the case. Summarizing several decades of research, Kesebir and Diener note that happiness “appears to bring out the best in people, making them more social, more cooperative, and even more ethical” (Kesebir and Diener 2008 p.121). Individuals with chronically high or experimentally increased subjective well-being evaluate others more positively, (Diener and Seligman 2004), exhibit more generalized trust (Brehm and Rahn 1997; Inglehart, 1999), are more generous (Konow and Earley 2008), more kind (Otake et al. 2006), more helpful¹² (De Neve et al. 2013), more apt to volunteer (Thoits and Hewitt 2001), and even respond more appropriately to ethical situations (James and Chymis 2004). Moreover, this link between happiness and ethical behavior appears to be replicated at the scale of populations. For example, happier nations score higher on generalized trust, volunteerism, and support for democracy, while exhibiting less intolerance of immigrants and racial groups (Diener and Tov 2007). Happier populations are also lower in human rights violations (Diener et al. 1995) and corruption, while higher in social equality, political freedom and access to education (Veenhoven 2005). The particular relationship between happiness and social trust is worth highlighting. After all, democratic institutions rely crucially on the ability of citizens to trust each other not to violate their civil and political liberties, and social trust makes individuals far more likely to cooperate. Given the link between happiness and trust, Diener and Tov (2007) have argued extensively that happiness may be a critical component for creating a culture of peace.

The relationship between happiness and human virtue seems consistent with lived experience, as well. I suspect that when we take a moment to consider the very happiest people we know (in the holistic sense), we are also calling to mind some of the very best people we know. Anecdotally, we observe that those who do well in life are also those who do well by others. We can speculate on the direction of causality, but one likely (partial) explanation involves the fact that happier individuals tend to have higher levels of positive affect and lower levels of negative affect, and these affective tendencies influence how individuals perceive and relate to one another. A large body of psychological research suggests that positive emotions predispose individuals to trust and cooperate with others and to engage in prosocial behaviors (for review, see Lyobomirsky et al. 2005). On the other hand, negative emotions (such as fear and anxiety) often motivate some of humanity’s worst tendencies, since we tend to divert attention toward the self when stressed (Mor and Winquist 2002). Unsurprisingly, when people are angry and fearful, they are more punitive (Rucker et al. 2004) and less politically tolerant (Skitka et al. 2004). Of course, there are almost certainly feedback effects here, as well; ethical individuals build better relationships, and we know that strong networks of authentic personal relationships are a major determinant of human welfare (Diener and Seligman 2002).

So Freud had it wrong; the ethical life does not typically come at the cost of happiness. Instead, the truth is closer to the Aristotelian conception of human nature on which the virtuous life and the happy life are coextensive (more on this point later). But to be clear, the claim is not that happy people are ethical; the claim is that making people happier is very likely to make them more ethical. The move may ultimately be from unethical to less unethical, but that is progress nonetheless.

4 Increasing SWB in a Population

To show why increasing population welfare is a predictor of moral progress, I want to highlight two points about how welfare often increases in a contemporary human community. The first point is that the level of happiness in a population is, to some significant extent, a function of its social, political, and economic institutions. So while natural events like earthquakes can, of course, alter access to resources in ways that impact welfare, much of what influences long-term welfare is a result of the institutions involving distributive justice, political process, social organization, etc. We infer this from the fact that the most significant changes in population welfare typically correspond to institutional developments, such as the breakdown of governments, increasing/decreasing inequality, and civil unrest (cf. Helliwell and Wang 2012).

If this is right – and this is the second point – then increases in population welfare will often involve changes to the enabling institutions that most directly impact and contribute to individual welfare. So what are the typical contributors to individual human welfare? There is no final consensus yet, but an overview of the literature suggests that the main contributors likely include: the satisfaction of basic needs, a favorable ratio of positive to negative emotions, personal autonomy, engagement in absorbing activities, strong networks of authentic personal relationships, and meaningful life pursuits (see Seligman 2012 for a review).¹³ Seligman notes, and it is worth emphasizing, that the empirical work on human happiness coheres surprisingly well with ancient wisdom. A life well-lived that involves meaning, solidarity, friendship, engagement, and pleasant experiences are common tropes across the world’s virtue traditions, and in the Aristotelian tradition in particular. The upshot is that increasing population welfare will often involve institutions that better enable individuals to live the kinds of lives that philosophers have historically promoted as the good ones, a holistic type of well-being that favor virtuous actions. Unsurprisingly, those nations that historically score in the top tier of international welfare assessments, such as Denmark, Norway, Sweden, and the Netherlands are those nations with institutions that better enable these welfare conditions. Happier populations tend to have strong social safety nets that ensure basic needs are met, a high degree of social trust, relatively high levels of personal autonomy, and strong social support systems (Helliwell and Wang 2012). On the view proposed here, this gives us some reason to think that promoting these enabling institutions is a reliable way of promoting moral progress.

5 Population Welfare and the Major Normative Traditions

Recall that the goal is to identify a reliable correlate of moral progress that can be endorsed by a wide range of normative traditions. The aim of this section is to show why most normative traditions should concede a strong, if imperfect correlation between rising mean population welfare (PW) and moral progress (MP). But it should be noted at the outset that not every conceivable normative framework will be amenable to the proposed correlation. The kind of ethical egoist (e.g., Rand 1964) who holds that right actions are only those actions that benefit the self may resist identifying moral progress with increasing mean welfare in one’s community. That said, this may be unfair to the more nuanced egoist who observes that the rising tide of human welfare tends to raise all ships (our own included). Similarly, the proposed correlation may also be resisted by the kind of libertarian ethicist who holds that our only obligations involve negative duties, essentially duties not to restrict the liberty of other agents.¹⁴ But perhaps the most trenchant opponents are environmental activists, such as those in the Voluntary Human Extinction Movement who, as their name suggests, believe that a better world would be one without any human beings.¹⁵ Nonetheless, I hope to suggest that the correlation proposed here between PW and MP can gain overlapping consensus from many of the widely embraced normative traditions.

5.1 Utilitarianism

Utilitarianism is the view that right actions are those that maximize human happiness, and as such, it is a natural ally to a view which correlates rising happiness levels with moral progress. Because utilitarians conceive happiness as the ultimate or primary good, raising levels of happiness just is moral progress. Given the centrality of subjective well-being measures to PW, the correlation between MP and PW should be amenable to hedonic utilitarians who are concerned primarily with rising levels of positive affect. But the view should also be amenable to to welfare utilitarians (e.g., Goodin 1991) who are interested in the propagation of a much richer conception of human happiness that included things like meaningful life pursuits and authentic personal relationships. Because our best measures of PW include include hedonic and welfare components, all of the standard constituents of utility are strongly correlated with it.

5.2 Aristotelian Ethics

The other natural ally to the view proposed here includes neo-Aristotelian ethicists who support a virtue-based moral framework primarily focused on identifying the traits that conduce to human well-being (Eudaimonia). Aristotelian ethicists typically follow Aristotle in recognizing that ethical theory must adapt to the messiness and imprecision of its subject matter in a way that a reliance on absolute normative principles often fails to do, but that a working definition, by its very nature, can embrace. Moreover, and for partly that reason, Aristotelian ethicists are often responsive to the best empirical evidence of what human flourishing involves. As such, the general methodology proposed here should be amenable to the Aristotelian project, as should the correlation between PW and MP in particular. After all, a central tenet of Aristotelian ethics holds that human flourishing involves rational activity in accordance with virtue, and thus the good life for human beings is, at once, an ethical life. In particular, well-functioning humans live in accordance with virtues such as bravery, temperance, and generosity, in a network of authentic personal relationships that supports their own virtue and the virtue of the polis. The purpose of politics, for Aristotle, is to set the stage for the development of the character traits that conduce to flourishing and to right action.¹⁶ And thus Aristotelians might naturally conceive of moral progress as a trajectory toward a polis where individuals are better able to fulfill their natural function, which is to say, a community populated with more flourishing individuals whose well-functioning inclines them toward right action.

Arguably, that is precisely what we observe in populations with higher welfare. Recall that increasing mean welfare typically requires institutions that enable its constituents to pursue more meaningful, socially-connected, engaged, and pleasant lives that are conducive to an array of prosocial behaviors – all of which are core components of eudaimonia in the Aristotelian tradition. Just as Aristotle would have predicted, happier individuals tend to be more generous, more cooperative, more willing to help others – in short, more inclined to virtue. In fact, an increasing number of thinkers have suggested that our best measures of human welfare and eudaimonia are, while not identical, very closely associated (Sumner 1999; Kesebir and Diener 2008). So for Aristotelians who conceive moral progress as a trajectory toward a polis where individuals are more capable of flourishing and expressing their ethical natures, then increasing mean welfare in a population should serve as a relatively reliable indicator of that phenomenon.

5.3 Deontology/Kantian Ethics

Deontology refers to a wide range of normative theories on which right actions are identified as those that are consistent with a certain set of rules or principles. A traditional deontic system, such as the one W.D. Ross proposes in The Right and the Good (1930/2002), is simply a list of duties, including the duty to keep promises, to repay favors, to not harm others, etc. These traditional forms of deontology should prove amenable to a correlation between PW and MP, since as we have already noted, happier individuals are significantly more inclined to discharge an array of duties (e.g., increased helping behavior, generosity, volunteerism) and they are less inclined to a range of moral failures (e.g., corruption, prejudice). Moreover, Ross also includes a general duty of beneficence in his list of obligations, which is a prima facie duty to contribute to human welfare. As such, a Rossian deontologist should be happy to concede a correlation between increasing population welfare and moral progress.

The issue becomes trickier with Kantian deontology, the normative framework in which all duties are reducible to the obligation to treat rational agents as ends in themselves (1785/2002), which is to say, to treat them with the dignity that is due to self-legislating beings. On the Kantian view, moral progress involves a trajectory toward the Kingdom of Ends, his term for a moral community that manifests mutual respect for human dignity in both action and intention. The natural worry is that Kantians cannot use happiness as a reliable correlate of MP, since Kant is at great pains to show that happiness is not intrinsically valuable.¹⁷ Kant holds that for an action to have unqualified moral worth, it is neither necessary nor sufficient that it create happiness. Instead, to have complete moral worth an action must be consistent with a principle that manifests respect for all rational agents as ends in themselves, and it must be motivated by the duty to respect that moral law. Kant argues that we have a duty to restrain our wills in precisely this way regardless of the consequences for human happiness. And so the natural objections is that increasing PW is, by itself, morally neutral in the Kantian tradition.

But this feature of Kantian ethics is actually no obstacle to the view presented here, and this becomes obvious when we recall Kant’s espoused political views. The first thing to point out is that Kant explicitly describes a world that he believes would be a moral improvement over the current one in his well-known political essay, Perpetual Peace. Here Kant takes the opportunity to promote the abolishment of standing armies, and it is perfectly obvious that Kant views a world with that institutional arrangement as morally superior to our current world. So how do we square this very reasonable position with Kant’s view about the moral worth of actions? Well, Kant is pretty clear that for an action to have unqualified moral worth, it must be motivated by respect for the moral law. But there are other things in addition to actions that are morally evaluable, such as institutions, including the military institutions he decries in Perpetual Peace or the educational institutions he is keen to see improved in On Pedagogy (1803). Institutions can succeed or fail to manifest respect for individuals as ends in themselves and are thus morally evaluable in that regard. Indeed, it would be very strange if Kant held that the moral assessment of an institution derived solely from its intentional design, given that many institutions, including some involving sexism and racism, are not the product of individual intentions, but emerge from the unintended consequences of collective action. Such institutions can still have moral value in the Kantian system, which again, explains Kant’s moralizing in Perpetual Peace ¹⁸ (see Anderson-Gold 2001 for a thorough discussion of Kant’s views about historical moral development).

Thus, Kantians can maintain the plausible view that we find ourselves in a morally better world if our institutions better accord with the demand of the categorical imperative to manifest respect for individuals as ends in themselves. Insofar as happier populations are more inclined to maintain institutions that respect human rights, more supportive of democracy, and less inclined toward prejudice, then happier populations will tend to be more like the Kingdom of Ends. Moreover, the Kantian can maintain this view alongside a view of right action on which the expressions of our wills, in particular, derives unqualified moral worth solely from our intentions. And thus the Kantian should happily concede that increasing human welfare in a population is a reliable, if imperfect correlate of moral progress.

6 Objections

We do not have the luxury of addressing every normative tradition here, including various forms of contractualism, contractarianism or care ethics, though I see no particular reason to think these traditions would be opposed to the core proposal, given that happier people are prone to discharge an array of duties that exist in the majority of mainstream normative traditions. In this section, we turn to some objections. But before we do, it is important to reiterate that a correlation is not an identity, and so there will undoubtedly be instances in which PW and MP come apart.¹⁹As already noted, population welfare may increase for reasons that have nothing to do with moral progress (e.g., a disease that eliminates an unhappy subset of the population), just as a population may improve morally without it necessarily being reflected in its welfare indicators. Finally, and perhaps obviously, positing a correlation between PW and MP doesn’t entail that any way of generating moral progress is permissible or good. Some methods may be unacceptably paternalistic, for example. The working definition helps in identifying the enabling conditions of moral progress, but we still require normative ethics to judge how best to instantiate those conditions within a community. With these caveats in mind, let us consider several objections that aim to highlight cases in which the correlation between PW and MP might fail to obtain.

6.1 The 1 % Objection

One natural objection concerns the potential scenario in which an increase in mean PW is the result of rising happiness levels restricted only to the most advantaged in a population. This challenge can be made worse by imagining a situation in which the mean is raised, in part, at the expense of the least advantaged. Call it the 1 % objection.²⁰

Fortunately, such an outcome is very unlikely. The main reason follows straightforwardly from facts about diminishing marginal utility, which make it implausible that one could raise the mean in a relatively large population with increases solely near the top. In most cases then, raising mean welfare will likely be the result of increases at the lower tiers, and that is indeed what history demonstrates. Historically, when populations have increased welfare over time, it is typically as a result of acquiring institutions that better enable individuals to meet basic needs, which is typically paramount to reducing human suffering in that population (Helliwell and Wang 2012). Importantly then, for the majority of populations, increasing mean welfare will be paramount to reducing misery in that population. And we have reason to be wary of any normative theory that does not identify large-scale reductions in misery as instances of moral progress. Of course, none of this rules out the possibility that raising population welfare might be orchestrated by exploiting some subpopulation. I have only argued that it is very unlikely, and it is telling that populations with higher welfare tend to have fewer exploitative institutions, not more.²¹ Still, it is worth recalling that the thesis posits a reliable correlation between PW and MP, not an identity relation, so the view can weather the charge that these properties might come apart in some exceptional cases.

6.2 The Happiness-Drug Objection

Another concern involves a scenario in which we develop some kind of happiness-enhancing drug that, when administered to the population, causes individuals to score higher on measures of individual welfare. Intuitively, such a change would fail to track moral progress. Firstly, it should be noted that we are nowhere near developing drugs that make individuals consistently satisfied with their lives; in fact, current evidence of drug users suggests that they are less happy than the average population (cf. Tuicomepee and Romano 2005). Secondly, we will almost certainly want a multi-criteria approach to population welfare that includes some non-subjective measures (such as individuals’ access to education and health outcomes) that might be difficult for a happiness-enhancing drug to impact. Nonetheless, let us suppose that we do ultimately develop technology that makes individuals feel genuinely happier, and this happiness translates to improved subjective and objective assessments. It should be noted that these pharmacologically-enhanced individuals would also tend to be more trusting and kind, and their societies would be more inclined to democracy and cooperation, and less inclined to rights violations and social injustice. So while I am skeptical of the plausibility of developing a drug that consistently raises multi-criteria measures of welfare, but the case is not clearly a counter-example to the correlation between PW and MP. Now of course we can imagine a more sinister scenario in which the happiness drug is forced on a population. But, as already noted, the correlation between PW and MP does not entail that every way of generating welfare is morally permissible.²² In any case, that is the beauty of a working definition; in such a scenario, we are free to adapt the theory to the facts on the ground, and thus we can mark it as an instance in which the proposed correlation does not hold.

6.3 The Wealth Objection

Another concern involves a scenario in which we raise population welfare by making a relatively wealthy society more wealthy.²³ Of course, this objection relies on the controversial claim that increases in wealth are positively correlated with increases in welfare, and currently the evidence for that view remains mixed (see Stevenson and Wolfers 2008 in support of a correlation and Easterlin Easterlin et al. 2010 against a correlation). Notably, it is income equality rather than inequality that appears to be positively correlated with higher welfare at the population level (Oishi et al. 2011). Nonetheless, let us suppose that there is a correlation between wealth and happiness even at the upper ends of the wealth spectrum, and that we find ourselves in a position to make a relatively wealthy population even more wealthy. I suspect the objection relies on the thought that increasing a population’s consumption capacity does not necessarily track any kind of moral improvement. But this may oversimplify the relationship between wealth and happiness. Arguably, it is not mere wealth that makes a high-income population happier, but wealth spent in the right kinds of ways. For example, it is often noted that wealth contributes most to our happiness when we favor experiences over consumer goods (Carter and Gilovich 2010). Some studies have also suggested that prosocial purchases generate more happiness than material consumption (Dunn et al. 2008). In fact, materialistic people widely report less happiness than those who prioritize non-materialistic goals (Kasser et al. 2004). So it is not implausible to think that a population that is better positioned to generate positive memories with their loved ones or to support meaningful prosocial causes is indeed a small step toward a morally better world, especially when we recall that these happier individuals will likely be more inclined to the range of ethical proclivities described earlier.²⁴

7 Conclusions and Applications

The aim of this essay has been to identify a working definition of moral progress that relies on a correlation between increasing population welfare (as measured by our best multi-criteria measures) and ethical advancement. The benefits of this strategy are two-fold. By identifying moral progress with a proxy property, the hope is to generate overlapping consensus on a reliable indicator of social progress that is feasible, quantifiable and relatively theory-neutral. By highlighting a point of overlapping consensus, a working definition of moral progress is a small step toward bridging the theoretical and practical spheres of ethical inquiry, and thus bringing together practitioners across philosophy, psychology and economics who espouse a shared goal of identifying the opportunities to create better human communities.

Secondly, a proxy property provides a shared normative framework where practical ethical claims can face some modest measure of empirical scrutiny. More generally, a method for identifying and measuring moral progress may prove useful in assessing the impact of the practices and institutions that concern applied ethicists and policy professionals. For example, there is currently a debate in the United States about how much economic inequality is permissible in the population. The fact that high levels of economic inequality are negatively correlated with population welfare can, I think, serve as the basis for a reasonable normative claim about what kind of society we are obliged to create. This appeal to empirical measures is meant to augment, not replace the traditional tools of normative ethics. Among other things, normative ethics must still identify the permissible ways of generating moral progress. But an empirical measure that serves as a proxy for moral development may open an avenue of rational discourse where a liberal society that includes very different ‘theories of the good’ can have a relatively objective discussion about what institutions and practices are good for us collectively.

Our best current measures of population welfare are not perfect, and they never will be. But despite the skepticism that is natural to philosophers, we should concede that these measures are better and more explanatory than perhaps they have any right to be. Nonetheless, philosophers have much to contribute to their future refinement, and in particular, to identifying the precise relationship between our best empirical measures of human welfare and the normative construct associated with well-being/flourishing. If such a project is successful, the future looks brighter than ever for the Enlightenment ideal of designing a society that better enables its citizens to live fulfilling, satisfying, and meaningful lives.