Moral and Political Foundations: From Political Psychology to Political Realism

1 Introduction

‘Erst kommt das Fressen, dann kommt die Moral’

[First comes the grub, then comes morality]

—Bertolt Brecht, Die Dreigroschenoper

In a recent article entitled ‘Ideology Justifies Morality: Political Beliefs Predict Moral Foundations’, Hatemi, Crabtree, and Smith (2019) (henceforth HCS) explore the causal pathways between political and moral beliefs with the help of structural and predictive modelling. They challenge the orthodoxy of the moral foundations theory (MFT) as discussed in, among others, Haidt (2008, 2012, accusing MFT of predicting what is already intrinsic to the theory, namely that moral beliefs influence political beliefs. HCS argue that, first, political psychology must start from a position which treats political and moral beliefs as prima facie equals to avoid self-justificatory theorising, and second, that an analysis based on a level-playing field between moral and political beliefs provides stronger evidence that political attitudes (towards gay rights, abortion, stem cell experiments, etc.) and ideology (liberal–conservative) at time one predict moral attitudes at a later time (time two and even time three) than vice versa.^[1] HCS’s findings, I argue, have ramifications for methodological choices in normative political theory: in particular, they supply evidence for the viability of the political realist approach.

Political realism is sometimes (though neither exhaustively nor exclusively) understood as promoting the metanormative discontinuity of politics and morality.^[2] There are sources of normativity—that is, phenomena that generate normative force—and some of those are genuinely political sources of normativity, distinct from moral sources of normativity.^[3] Neither the realists nor the moralists (the realists’ adversaries) have thus far proposed an intelligible account of the structure of metanormative reality.^[4] When one tries to disambiguate realist avowals, such as political theory not being applicable to ethics, in a non-metanormative way, we briskly come to the conclusion that any firm distinction between politics and ethics must eventually break down. As I think Baderin (2021, 174) argues successfully, on the only sensible non-metanormative interpretation of many of the realist’s most elusive catchphrases, ‘contemporary political theory’s failures of realism reflect [the] political theorists’ tendency to replicate mistakes to be found within particular traditions of ethics’. Baderin concludes that realism’s aim can ‘no longer [be] to separate political theory from ethics but only from bad ethics’ (2021, 174); collapsing realism into moralism. The failure to account for a clear distinction between political theory and ethics suggests that the debate has become largely unwieldly. To remedy this situation, I argue that the realists should ‘downplay’ the (thus far) inconclusive debate over realism’s metanormative standing in favour of a non-metanormative inquiry exemplified by HCS’s empirical analysis of moral and political beliefs. Using HCS’s clinical studies of the causal pathways between moral and political beliefs,^[5] realists may better position themselves to defend political realism’s central claim: the rejection of the priority of the moral over the political.

2 Moral Foundations Theory

People hold beliefs, convictions and opinions for reasons that are often opaque to both themselves and others. As early as 1908, the Fabian socialist Graham Wallas (1908) argued against the technocratic assumption that political opinions are determined by relational calculation. As one of the first serious examples of social and political psychology, Wallas’ Human Nature in Politics expressed deep concern about democracy’s vulnerability to technocratic manipulation of political opinion. The political opinion of the masses, Wallas contended, could best be controlled by exploiting the masses’ subconscious.

The early Frankfurt School picked up on Wallas’ observation. Adorno et al. merged social, individual and political levels of psychological analysis in The Authoritarian Personality. One of the central insights of this heavily criticised study—criticised on both ideological and methodological grounds—is that specific ‘ideologies have for different individuals, different degrees of appeal, a matter that depends upon the individual’s needs and the degree to which these needs are being satisfied or frustrated’ (Adorno et al. [1950] 2019, 2). Individuals, Adorno et al. argued, hold political beliefs and appeal to underlying (moral) values in political argument because they are expressions of deep-seated psychological needs for order, security, cohesion, and self-esteem. Ideologies, as the Frankfurt School referred to amalgams of political convictions, serve psychological functions. Those functions may or may not be fully transparent to both the individuals themselves and others. From the perspective of the political psychologist (or the ideology critic), however, those functions and their reference points in the human psyche can help make sense of how actual politics shapes the matrix of our attitudes.^[6]

Political psychology later went in search for the ‘holy grail’ of the most simplistic theory of political ideology. A study by Jost et al. (2003, 2012 is exemplary in illustrating the quest for maximum simplicity. Theirs is a theory of political ideology as motivated social cognition. As they see it, individuals with strong psychological needs for cohesion, order, and certainty are attracted to conservatism and repulsed by liberalism. The needs for cohesion, order, and certainty, they argue, are present in all individuals to varying degrees. Contemporary political psychology calls those innate (but socially formed) needs ‘moral foundations’. Those ‘foundations’ can be triggered and temporarily activated by, for instance, propaganda.^[7]

Haidt (2012) and Graham (2013), both pioneers of the moral foundations theory (MFT), have criticized Jost’s theory of ideology as motivated social cognition for its implicit devaluing of conservative moral principles^[8]—principles that liberals, according to MF theorists, don’t even ‘acknowledge to be moral principles, such as unconditional loyalty to one’s group, respect for one’s superiors, and the avoidance of carnal pleasures’ (Jost et al. 2013, 101). A suitably differentiated account of political psychology, both Haidt and Graham (2007) argue, would uncover the liberal’s impoverished sense of morality. Liberals, they say, don’t recognize the reduction of fear, anxiety, and uncertainty as being of moral normative importance. A liberal’s moral arsenal knows only issues of fairness and harm-avoidance. Conservatives, on the other hand, value loyalty, obedience, and purity in addition to fairness and harm-avoidance.

This insight on the liberal’s impoverished view of morality forms the backbone of MFT, which proceeds by codifying a sample set of normative moral attitudes into psychological modules—that is, moral foundations—which are then tested as to how they relate to other normative domains, such as, in our case, the political domain. In other words, there are basic moral foundations, five in the case of Haidt and Graham (2007), and those moral foundations make up an individual’s ‘moral palette’. The hypothesis: differences in the weighting between fairness, harm-avoidance, loyalty, obedience, and purity have an effect on, and make predictable, an individual’s political attitudes and beliefs.

MFT rests on four pillars. The idea that evolution encourages moral pluralism is the first pillar. The quest for stability, order, and cohesion is not the only psychological determinate. According to Haidt and Graham, there is no reason to expect nature to be parsimonious with respect to our moral foundations. There are five moral foundations at a minimum, each expressing a binary set of negative and positive moral attitudes: (1) care–harm, (2) fairness–cheating, (3) authority–agitation, (4) loyalty–betrayal, (5) disgust–purity. Those binary sets are supposed to measure an individual’s sensitivity to (1) the suffering of others, (2) exploitation, (3) social hierarchy, (4) social rivalry, and (5) taboos. Graham et al. (2011) further distinguish between the categories of individual morality (1, 2) and group morality (3, 4, 5).

The second pillar can be referred to as nativism. Marcus (2004, 39f.) describes it as ‘nature [providing] a first draft, which experience then revises’. Reminiscent of Chomsky’s idea of universal grammar, MF theorists think that genes write the first draft of a moral set-up into the child’s neural tissue, experience (i.e., cultural learning) does the editing of this draft. The first draft cannot be inferred from the moral set-up of an individual (or one culture), but only by examining the shared moral nexus of individuals from different cultural backgrounds. From this, MF theorists infer the common starting point of all moral psychology. As Graham et al. (2013, 61) say, ‘morality is innate and highly dependent on environmental influences.’

Cultural learning is the third pillar. If there were no common starting point of moral psychology, ‘then groups would be free to invent utopian moralities … and they’d be able to pass them on to their children because all moral ideas would be equally learnable. This is clearly not the case’, say Graham et al. (2013, 61).^[9] Without cultural learning, moral attitudes would have to be homogenous across cultures, and accordingly, there couldn’t be any differences in political ideology across polities, which is evidently not the case (see Haidt, Koller, and Dias 1993). MFT claims that ‘the universal (and incomplete) first draft of the moral mind gets filled in and revised so that the child can successfully navigate the moral “matrix” he or she actually experiences’ (Graham et al. 2013, 61). There are thus, metaphorically speaking, moral foundations on top of which additional and more complex moral structures may be formed.

The final pillar concerns the possibility of recognizing moral foundations, which Bargh and Chartrand (1999, 475) think reveal themselves through relatively immediate, automatic intuitive means. The social intuition model defines moral intuitions, which serve as the MF theorist’s access point to an individual’s moral matrix, as ‘the sudden appearance in consciousness, or at the fringe of consciousness, of an evaluative feeling (like–dislike, good–bad) about the character or action of a person, without any conscious awareness of having gone through steps of search, weighing evidence, or inferring a conclusion’ (Haidt 2001, 818).^[10] Moral reasoning is the post-hoc rationalization of those ad-hoc moral intuitions, and consequently more likely to be ideologically distorted.^[11] Studies in political psychology may therefore only be based on those ad hoc moral intuitions.

In sum, MFT predicts that moral foundations determine political attitudes. Crudely put, whether an individual is an alt-right fascist or a volunteer at refugee support groups is thus a matter of the individuals’ moral foundations. According to MFT, knowing about individuals’ moral foundations makes predictable their political attitudes. If we want to translate this position into the language of normative political theory, moral principles must be the starting and pivotal point of political argument and analysis. This, critics say, is where MFT displays an unwarranted bias towards the priority of moral attitudes. A bias that eventually, and unwarrantedly, gets codified into the theory’s conclusions.

3 Problems of the Moral Foundations Theory

MFT posits that substantive political attitudes—such as being a liberal, a socialist, or a conservative—differ with respect to the moral foundations adhered to in political judgments.^[12] Liberals, for example, say Graham et al. (2013), seek foundations for their political attitudes more often on the negative side of individual morality (that is, harm and cheating) than do conservatives. This points towards the first of the objections to MFT that I want to discuss: the liberal–conservative divide is a self-imposed limitation on MFT, one that is perhaps a hereditary defect derived from the scaling models introduced in The Authoritarian Personality and its infamous f-scale. This means that MFT does not test on robust political attitudes but only on the relatively narrow dimension of liberal versus conservative attitudes. Put differently, MFT works on simplified models at the cost of losing out on expressiveness.

Another worry is the problem of missing contestation. A given political concept, say freedom, ‘means one thing to liberals and another to socialists’ (Prinz and Rossi 2017, 355). MFT generally lacks dimensions that a more complex analysis, one that makes more detailed models of the relations between political attitudes and moral attitudes, would cover.

MFT’s main flaw, however, is the circular self-justification of its predictions. In most of its applications, the causal priority of moral attitudes to political attitudes is simply being asserted rather than argued for, and this is not a matter of chance, but a matter of theory design. Recall that MFT is based on a sample set of normative moral attitudes codified into psychological modules, the moral foundations, which are then tested on their command over other normative domains, such as, in our case, the political domain. Thus, by the very design of MFT, moral attitudes are being tested for their causal impact on political attitudes, but never vice versa. The possibility of the reverse direction showing even more significant empirical results, however, is exactly what is at stake here: Do moral beliefs determine political beliefs, do political beliefs determine moral beliefs, or is it a reciprocal relationship?

MFT thus presupposes (rather than proves) that moral beliefs determine political beliefs. This is of course problematic on its own, but also and especially, say HSC, before the background of traditional studies on moral reasoning, such as Kohlberg’s (1969) and the infamous Milgram experiment (1974), which suggest that ‘moral judgments are made in order to justify pre-existing social and political beliefs’ (HCS 2019, 769), and hence that social and political attitudes are in some sense prior to moral judgements.

Traditional studies indicate that there is a causal path running from political to moral attitudes. In Milgram’s behaviorism, for instance, political attitudes and ideology have greater influence on individual and group moral attitudes than moral attitudes have on political attitudes—situational attributes replace presumed dispositional attributes. Milgram’s core thesis in his Obedience to Authority—which rests on Milgram’s controversial electroshock experiments and was supposed to disprove the ‘Germans-are-different’ hypothesis^[13]—is supported by decades of research on authoritarian values and group identity. Those studies posit that ideological values largely guide our moral attitudes.^[14]

In order not to presuppose what is being argued for, any trustworthy study on the causal relation between moral and political beliefs must precede from a level-playing field, one that includes not only a sample set of normative moral attitudes codified into psychological modules, the so-called moral foundations, but also a sample set of normative political attitudes codified into psychological modules, which we can call political foundations. To presuppose that there are no political foundations that could possibly influence our moral attitudes is to beg the question.^[15]

4 Political Foundations

Hatami, Crabtree, and Smith (2019) add a sample set of normative political attitudes codified into psychological modules—political foundations—to their study of the causal links between political and moral attitudes.^[16] They rely on several external datasets arguing that political attitudes can, just like moral traits, be ‘anchored in rapid, implicit evaluations that spring from deep psychological and dispositional mechanisms’ (790). In other words, if there is such a thing as moral foundations, for which there has been no prior argument, why not also presuppose the existence of political foundations so as to create a level-playing field?

Since political attitudes are multifaceted, HCS break them down into three discrete ‘ideologies’: social ideology, defence ideology, and economic ideology.^[17] The first has a 7-item measure, the second a 3-item measure, and the third a 4-item measure. HCS are (again replicating the MFT setup) using a self-reported 7-point liberal–conservative scale (from 1 = extremely liberal to 7 = extremely conservative to avoid ‘cheap political agnosticism’, that is a middle position such as 5 on a 1–10 scale) in addition to the experimentally examined attitudinal measures concerning the discrete ideologies.

The three discrete ideologies have subfactors: social ideology comprises opinions on homosexual marriage, stem cell experiments, evolutionary theory, euthanasia, the legalization of marijuana, pre-marriage sexual contact, and attitudes on X-rated movies. Defence ideology is measured on the basis of opinions on the war in Afghanistan, military spending, and the Iraq War. The factors of economic ideology correspond to opinions on Medicare for all, the need for labour unions, and education spending.^[18]

For the set of moral foundations, HCS chose the orthodox MFT’s set of harm avoidance, fairness, loyalty, obedience, and purity. The study evaluates them by asking, first, about the relevance of each item to the participant’s perception of his or her own moral compass, and second, about how much the participants agree with the relevance of a certain item in the context of a given moral domain (e.g. ‘Is purity relevant when choosing how to distribute exclusive resources?’).

HCS test the causal relationship between political attitudes and moral attitudes by using a cross-lagged panel (CLP). A CLP consists of two variables, X and Y, which are measured at two difference time stamps. From that panel we get four measurements: X1, X2, and Y1, Y2. X1 is a political attitude at time one, X2 the same political attitude at time two; Y1 is a moral attitude at time one, Y2 is the same moral attitude at time two. From those four measurements six correlations can be determined. Most significant for the point in case are the stability relations between X1 and X2, and Y1 and Y2, which answer the question, ‘Do political and moral attitudes remain stable over time?’ as well as the cross-lagged correlations between X1 and Y1, and X2 and Y2, which answer the question, ‘If political attitudes justify (i.e. cause) moral attitudes at time 1, do they also have the same causal effect at time 2?’. The evaluation follows multiple regressions. If the coefficient on Y1 or X1 is statistically sufficient, not only is that evidence for covariation (i.e. evidence that X causes Y), but also for temporal precedence. This is an appropriate CLP. We can not only see if there is an immediate causal path, but also test on long-term causal effects, for what is measured is the impact of variable X1 on variables X2, Y1, and Y2, as well as the impact of variable Y1 on Y2, X1, and X2.

The findings of HCS’s cross-lagged panel suggest that ‘ideology, no matter how measured, was more stable than any moral foundations’ measure within each study’ (HCS 2019, 795). The correlations for political attitudes causing moral attitudes are 0.71 for social ideology, 0.41 for defence ideology, and 0.58 for economic ideology. This is a significant difference compared to the impact of the moral categories (1) individual morality = 0.35 and (2) group morality = 0.38.^[19] This implies that, first, political attitudes are more stable (over time) than moral attitudes and second, that moral foundations have (almost) no causal impact on political attitudes.

Do political attitudes determine moral attitudes? In four out of seven models (those concerning authority, loyalty, purity, and obedience) the parameter ‘political attitude at time 1’ had a statistically significant impact on the parameter ‘moral attitude at time 1’. In contrast, only in two models has the ‘moral attitude at time 1’ been statistically significant for political attitudes at time 1 (see HCS 2019, Figure 5). The models of HCS’s study suggest that political attitudes determine moral attitudes.^[20]

5 Moral and Political Normativity

HCS’s study supplies evidence principally for the political realist approach in normative political theory. I want to suggest that HCS’s findings may replace established ways of validating the realist approach.

According to political realists, most anglophone political philosophy from Rawls, to Dworkin, to Cohen have misconstrued the political domain as being under the guidance of morality. Realism is the counter-project. The political is considered an autonomous domain with its own distinctive concepts and its own distinctive source of normativity. The ‘ethics-first approach’ must be abandoned, says the realist, and politics properly theorized through the concept of political legitimacy.^[21] This—the rejection of the universal priority of the moral over the political in political theory—I consider the core thesis of political realism, as espoused in the work of prominent realists like Bernard Williams (2005) and Raymond Geuss (2005, 2010.^[22]

About a decade ago, a heated debate erupted, taking its cue from Bernard Williams’ theory of legitimacy. Williams claims that his theory of legitimacy does not contain a morality that is external to politics.^[23] This is what makes a theory of legitimacy a realist theory. The ongoing dispute between realists and moralists about the standing of political normativity vis-à-vis moral normativity has its origins in Williams’ rather elusive notion of ‘making sense’. It plays a vital role in the evaluation of claims to legitimacy. Let me explain.

The legitimation story, a narrative on the grounds of which a political actor demands legitimacy, must make sense to those subject to its power in order for that political actor’s authority to count as properly legitimate. The notion of ‘making sense’ is unique in that it is ‘evaluative when applied to other [political] contexts though it is normative when applied to our own [political context]’, says Sleat (2010, 488). The question later generations of realists have been grappling with is, ‘What is this mysterious political normativity, and where does it come from?’ In their attempts to answer this question, some realists went so far as to posit that Williams’ theory of legitimacy is ‘morality-free’.^[24] Whereas Williams claimed not to be working with a ‘morality external to politics’, subsequent realist scholars framed their position, in one way or other, as void of all moral import. Note that these are two completely different claims.

The transition from ‘external to’ to ‘free from’ was realisms’ first sleight of hand. The second followed shortly thereafter. Thinking about ways to amplify the notion of ‘morality-free’ into something like a substantial claim for realism, subsequent commentators asserted the existence of a genuine political source of normativity.^[25] The demands of legitimacy, some realists argued, are generated within the real practice of politics, and so is politics’ own normativity.^[26] A pertinent example of this metanormative realist position is Rossi and Sleat (2014, 1).^[27] Political realism, they say, is defined ‘on the basis of its attempt to give varying degrees of autonomy to politics as a sphere of human activity, in large part through its exploration of the sources of normativity appropriate for the political’. Jubb and Rossi (2015, 1) argue that ‘a non-moral distinction between politics and sheer domination can give us a distinctively political normativity’.

Ever since that metanormative turn in the literature, realists themselves have been introducing metanormative terms without ever going down the rabbit hole of actual metanormative inquiry to show what must be shown, namely, first, that political normativity exists, and second, that it has precedence over moral normativity. As Leader-Maynard and Worsnip (2018, 764) correctly identified, the recent discussion is reminiscent of metaethical ruminations about whether prudential normativity is a kind of moral normativity, or whether epistemic normativity is a kind of instrumental normativity’. The question of realism has become a disagreement about the ontology of the normative sphere, a Quinean controversy about existential quantifiers. It used to be a question of logical priority.^[28]

We can read this metanormative thesis in a number of ways. On a strong reading there is absolute discontinuity between the moral and political sources of normativity.^[29] I call this, the discontinuity thesis: there is a genuine and autonomous political source of normativity. On this strong reading of the discontinuity thesis, the moral and the political sources of normativity are diametrically opposed, exhaustive, and exclusive. In a debate concerning the normative predominance of either the political or the moral source of normativity in political theory, an argument for either side is ipso facto an argument against the other. ^[30]

For the ‘strongish’ realist, on the other hand, there may not be a genuine political source of normativity distinct from the moral source of normativity. All that matters is that politics (or political theory, for that matter) is guided by a non-moral source of normativity. ‘Most contemporary theorists of normativity accept that not all normativity is moral normativity. Other candidate kinds of normativity include epistemic normativity, prudential normativity, “aim-given” normativity, and aesthetic normativity,’ say Leader-Maynard and Worsnip (2018, 756).^[31] Rossi (2019), to mention just one example, advocates an epistemic (and therefore non-moral) realist programme that he calls radical realism. Still, there is an assumed discontinuity between the realist’s preferred source of normativity (epistemic normativity, in the case of Rossi) and the moral source of normativity.

On a weak reading, ‘morality may have a role to play in providing a source of political normativity, yet it remains important to appreciate the manner in which politics remains a distinct sphere of human activity’ (Rossi 2014, 2).^[32] Note that this doesn’t suggest any distinct metanormative position whatsoever.^[33] Establishing ‘continuity’ or ‘coexistence’, rather than ‘priority’ or ‘autonomy’, might, after all, be the realist’s most sensible and promising objective. Baderin (2021), for instance, argues for a continuity approach to the study of ethics and political theory. She says a continuity approach better realises the realist’s own aspiration for greater sensitivity to empirical detail in normative political theory.^[34]

It is, however, and regrettably so, the strong reading of the realist position which has assumed an (almost) hegemonial position in the literature. A partial reason for the strong view’s dominance over, say, a continuity approach, might be that advocates of moralism tend to target the strong reading. Perhaps inevitably, realists tend to defend the strong thesis in turn, digging a rabbit hole. Take Leader-Maynard and Worsnip (2018), self-proclaimed moralists, who consider themselves in opposition to a ‘slew of recent theorists contending that political normativity is its own distinctive kind of normativity, independent of moral normativity’ (2018, 756). Leader-Maynard and Worsnip take turns against strong realism.

By comparing Williams’ realism with, for instance, Leader-Maynard and Worsnip’s rendition of the realist approach, we notice that the debate has transitioned from describing realist theories as void of a morality external to politics to framing them as normatively autonomous and prior to moral normativity. These are two completely different claims. As mentioned above, what used to be a question of logical priority has been refashioned as an ontological dispute. If Sleat and Hall are right in their interpretation of Williams, which suggests that he thought his theory of legitimacy works without moral principles external or prior to politics, then it would seem that the chief realist, or liberal-realist, source of inspiration—namely Williams—is a proponent only of the weak understanding of the autonomy of the two realms. The fact that the strong understanding is still (broadly) dominant in the debate^[35] can thus be accused of breaking with its own heritage. The strong realist’s talk about the metanormative autonomy of politics is a misreading. Williams’ central objective was to avoid moral principles that are not already internal to politics, not to over-inflate realism into a metanormative proposition.^[36]

The debate now hinges on whether the metanormative disagreement can be rendered substantive, or non-verbal, or indeed intelligible to begin with. What does it mean for two sources of normativity to be distinct in the relevant metanormative sense? The answer to this ontological question posed by realists, even how to go about answering it, is far from clear.^[37] As long as there is no clear assessment of metanormative reality, I suggest realists are better advised to turn to alternative ways of making their case. One avenue for doing this may be to explore HCS’s clinical studies on the causal paths between moral and political beliefs.

6 Attitudes and Normativity

Priority is a polymorphous concept.^[38] The notion of priority at play in Williams’ realist approach is logical priority. Later generation realists have attempted to make their case for ontological priority. Logical priority means that in deductive reasoning, you can’t get from your first principles to your conclusion without moving through some other proposition. The concept of ‘student’ is prior to the concept ‘undergraduate’ because you cannot answer ‘What is an undergraduate?’ without resource to the concept ‘student’. For Williams, the concept ‘politics’ is prior to the concept ‘morality’ because of his definition of politics. He thinks that moral issues, like issues of justice, cannot be addressed unless the question of whether the political actor is justified in addressing issues of justice has been resolved successfully. HCS’s study confirmed the directionality of reasoning from political to moral attitudes. I therefore suggest realists replace their established (logical, metaphysical) notions of priority with the causal sense of priority of political over moral attitudes.

One might think that there’s an objection to my proposal in the manner of an open-question argument, that it would be a fallacy to think that because political attitudes proportionately influence moral attitudes, political values should determine moral values. But what is the objection to a theory like realism—itself adept at producing normative verdicts—having its motivational or definitional grounds in factual statements? Realism is a normative theory (or at least it wants to be)^[39] but it’s definitional or motivational grounds themselves are not necessarily normative and can be factual. My proposal is not to ‘link’ the empirical to the normative, and thus try to bridge the is–ought gap. Rather, I take it that as long as there is no robust metanormative definitional ground for realism, realists would do better building their normative theories on an empirical footing.^[40]

The theoretical commitments of the position I sketch here are certainly not unheard of in metaethics. Consider metaethical subjectivism as an example. It is a function of metaethical subjectivism that moral statements are factual statements about the attitudes that normal human beings hold on particular moral issues. If metaethical subjectivism is not being accused of unwarrantedly trying to bridge the is–ought gap, why would we want to accuse a realism based on HCS’s political foundations of committing such a fallacy?

I want to encourage those realists (and moralists, for that matter) unconvinced of my proposal and intent on making their case for metanormative priority to ask themselves whether the metanormative debate is really going anywhere. What should concern these metanormativists is that the recent literature on metanormative realism bypasses the established literature on metanormativity almost entirely.^[41] Korsgaard’s famous text on the Sources of Normativity (1992), to take a prominent example, doesn’t concern ‘sources’ in a sense relevant to the debate between moralists and realists. Metanormative discussion usually starts from a point where a source of normativity—as in something that makes a normative force political, ethical, moral, etc.—is accepted as existing, ontologically speaking, but the origins of this very normative force are unclear or opaque. Korsgaard asks, ‘In virtue of what is a certain value, be it moral, ethical, aesthetic, or political, genuinely reason-giving?’. That’s not what the realists are after.

As Baderin says, realists face the desideratum of ‘nondistortion’ and ‘nontriviality’ (2021, 1736). In other words, realism must ‘succeed in distinguishing political theory from ethics, without relying on a distorted picture of the latter [and] assert/embody something about the practice of political theory that many contemporary political theorists, at least implicitly, deny’. No such nondistortive and nontrivial account of realism has hitherto been defended. This suggests that there is at least a pro tanto reason for the realist to abandon the current metanormative agenda. If the priority of contextualist political attitudes over universalist moral principles is the most central aspect of political realism, then realists can afford to make their case without raising an issue about sources of normativity. That is to say, HCS’s findings need not map onto the debate about sources of normativity (through bridging the is–ought gap) in order for me to justifiably recommend an amendment of the realist programme. HCS’s findings can be the starting point of a realist programme that avoids the metanormative mystery.

7 Conclusion

HCS’s paper supports the claim that political attitudes determine moral attitudes. I argue that the realist might wish to consider HCS’s study as inspiration to, first, discard any metanormative bickering—it simply doesn’t do the realist any service—and second, to revise her realist agenda accordingly. If political realism wants to go beyond a disagreement about existential quantifiers—which it should, because there is some potentially transformative work to be done in the realist vein—it needs to depart from the metanormative case, be less concerned with the realism, and more concerned with the politics in political realism. I suggest taking the idea of political foundations seriously.^[42] These political foundations can be the non-metanormative underpinnings of realist theory.^[43] Hatemi, Crabtree, and Smith’s study supports the realist hypothesis that politics is in some relevant sense ‘prior’ to morality. On the basis of political foundations, realists can make their case with confidence, at last.