8.2.1 The Cognitive Space of Morality

A close and careful look at the phenomenology of morality gives rise to some important observations.Footnote ⁸ One is that something like a moral experience exists at all. Humans operate naturally within a mental space that has a normative dimension. There is a specific mental domain of morality, a qualitatively distinguished element of conscious thought, an introspectively accessible, distinctive, intuited, subjectively experienced aspect of our mental life, a qualia, as it is often said, or – to use standard understandings of this term – a certain phenomenal character of certain forms of experience. The availability of such a cognitive domain is not self-evident; rather, this domain represents an empirical property of the human mind that not all organisms share. Human beings do not perceive ultrasound, but bats do. Human beings see the world in the distinct colors of morality, but bats do not. This cognitive domain does not concern a side issue but defines nothing less than an element central to the identity of the human species: the moral dimension of human lives. Its existence consequently merits close attention.

A further interesting observation concerns the fact that there is a highly and intricately qualified limited set of possible objects of moral evaluation. This set already restricts the kinds of morality that are possible. The dropping of an apple from a tree into the hands of a hungry person is not a virtuous action on the tree’s part. Or, as Hume observed: “A young tree, which over-tops and destroys its parent, stands in all the same relations with Nero, when he murdered Agrippina,” but does not commit matricide.Footnote ⁹ This is because agency is a precondition for the moral evaluation of certain events in the world. If these events cannot be attributed to agents, questions of moral evaluation do not arise.Footnote ¹⁰

Moreover, placing your pen gracefully on your desk cannot be the possible object of moral evaluation either (except under very particular circumstances), even though an agent performs this act. However, this act is a possible object of aesthetic evaluation – another distinctive element of human experience.Footnote ¹¹ Kicking a ball so as to feel like Dzsenifer Marozsán for a few precious seconds is morally very different from kicking a defender or a dog that is in your way. A precondition of moral evaluation is thus – to put it very roughly – something like volitionally controlled or controllable bodily actions or omissions of agents with consequences for the well-being of sentient beings and other qualified objects of moral concern; intentions concerning such actions or omissions; states of affairs resulting from such intentions, actions and omissions; or qualified emotions directed at the well-being of others.Footnote ¹²

8.2.2 Principles of Morality

As we already have seen, in analyses of morality it is particularly important to distinguish behavior and motivational inclinations to behavior from the moral evaluation of this behavior and the intentions underlying it. Concrete behavior, including what is called prosocial behavior, and the inclination or preference for such behavior are one thing, the reflexive appraisal of this behavior, inclination or preference with deontic dimensions quite another. Only the latter falls within the proper realm of morality and ethics.Footnote ¹³

If we turn to the content of morality and carefully analyze some qualified (only seemingly simply structured) considered judgments of the kind described above, we see that these judgments seem to be guided by principles of egalitarian justice and altruism across a wide range of cases. That it is just to distribute party favors equally to the young guests of a child’s birthday party seems as uncontroversial as that it is a morally good deed to help prevent starvation in Yemen.Footnote ¹⁴ Respect for others is a further important principle. This state of affairs is not particularly surprising, as all of these principles run through the history of ideas, too, as the core of morality, across cultures and millennia, albeit accompanied throughout by influential skeptical voices from Thrasymachus to Nietzsche and beyond, who have argued (with greater or lesser philosophical sophistication) that morality has no content that lends itself to rational reconstruction and that apparently core normative concepts do not mean anything at all.Footnote ¹⁵

Substantial empirical work carried out in recent years points in the same direction, supporting the observation that certain identifiable elements – more precisely, egalitarian and altruistic principles – seem to play an important role in the evaluation of the justness and goodness of actions, despite some theoretical limits of parts of this research that already have been reviewed above.Footnote ¹⁶ Other frontiers of research include the prohibition of the instrumentalization of human beings that plausibly lies at the heart of a proper analysis of the extensively empirically researched trolley problems and distinctions between different subjective cognitive and volitional attitudes towards potential actions, including direct and oblique intentions and their relevance for human moral evaluation.Footnote ¹⁷

If we analyze the empirical work on justice and basic uncontroversial judgments about the justice and injustice of intentions, actions and states of affairs and do not forget what the struggles of social history seem to suggest about ideas of justice, then differentiated principles of equality as already discussed aboveFootnote ¹⁸ seem to have considerable explanatory power to account for many patterns of moral evaluation. We have said that a just distribution demands proportional equality between the value of the specific criterion of distribution reasonably related to a particular sphere of distribution on the one hand and the amount of the good distributed on the other. Other normative demands concern the equal treatment of persons, restitution and respect for the equal worth of each human being – as indicated, the ultimate yardstick of just treatment and a just state of affairs. Such egalitarian principles match the empirically identified patterns of moral evaluation recalled in this study, including standard examples such as the ultimatum or dictator game, the many debates about the proper interpretation of these findings notwithstanding.Footnote ¹⁹ The same also holds for other clues to the moral world in which human beings live beyond experimental data, not least social history and its many egalitarian struggles, some of which we have recalled. It should be noted that a main bone of contention in these political struggles was not the principle that equals ought to be treated equally, but the criteria of distribution and who and what actually fulfill these criteria. Regarding rights, for instance, an important question in history was whether fundamental rights are distributed in a society on the basis of the bearer’s humanity or some other criterion (say, aristocratic birth) and who fulfills this criterion (for instance, whether women are fully human or some kind of deficient being and thus not entitled to a full set of fundamental rights). If we pay due attention to such factors, the core of the human quest for justice becomes considerably more transparent.

To achieve sufficient explanatory depth, however, the analysis of these empirical observations needs to remain aware of the distinction between a moral competence and its actual use, the plurality of motivational factors influencing human action, including nonmoral interests and the complex structure of moral judgments, with their cognitive, volitional and emotional components (to be discussed in more detail below), which are not reducible, for example, to preferences or aversions.

Whether there is such a thing as genuinely other-regarding altruistic behavior or whether any action beneficial to others is ultimately motivated by some (albeit perhaps refined and hidden) self-interest of the agent is one of the traditional questions of practical philosophy.Footnote ²⁰ As in the case of justice, this is a huge debate that today is enriched by interesting empirical work.Footnote ²¹ In this context, it is important, too, to rely on a sufficiently complex theory of morality, in particular to distinguish between actual behavior and considered evaluation and thus distinguish the question of whether people are in fact acting because of a genuinely altruistic motivation from the question of whether genuine altruism is the precondition for evaluating something as morally good. There is not much reason to believe that people generally excel in altruistic behavior. However, this observation tells us nothing about the principles that guide moral judgment – for example, when evaluating the selfish behavior prevalent around us (including our own).Footnote ²² Concerning these principles of evaluation, there are reasons to think that such a genuinely altruistic motivation is in fact a core element of moral evaluation. More precisely, it seems plausible to assume that an action is morally good if it is performed with the direct intention (or purpose), not only the oblique intention (or knowledge), to foster the well-being of the patient. If this is so, it is irrelevant for the moral evaluation whether or not the fostering of the interests of the agent is – at the same time – a directly intended or foreseen (obliquely intended) consequence of the action and forms a second reason for action in a bundle of motives. The direct intention to foster the well-being of the patient of the action appears to be a necessary condition of morally good action.Footnote ²³

To illustrate the meaning of this principle, it is useful to look at one of the most refined versions of ethical egoism. This form of egoism holds that altruistic behavior is ultimately motivated by the desire to experience the satisfaction of having acted in a morally appropriate manner. This argument makes an important point, namely that moral action does indeed provide some particular form of satisfaction for the agent and that agents are certainly often aware of this. In addition, acting immorally can have unpleasant effects, too, such as feelings of shame. These observations do not settle the issue, however. Consider the following case: Pawel helps Mio, thinking: “I do not care for this person and her well-being at all (what a silly person she is, in fact!), it just happens (unfortunately) that I have to do something for her in order to reap the sweet fruit I really desire, namely to feel the satisfaction of being a truly nice person!” Is this really a morally laudable deed? If doubts arise about the moral praiseworthiness of an action with such an intention, it seems to confirm the analysis above. This is because the Pawel has only an oblique intention to help the other person and not the direct intention to be beneficial to her: His direct intention is to satisfy one of his own personal desires, and helping the other person is only a (perhaps even unwelcome) means to achieve that end.Footnote ²⁴

Another point is perhaps worth noting: Justice seems to be something like a limiting condition of morally good action. There are reasons to believe that there is no morally good intention that violates principles of justice. If Pawel, for instance, helps three out of four people in need, not because he cannot help them all, but just because he feels like excluding one person on a whim, this is not a morally good action, despite his direct intention to help the other three, because it violates principles of equal treatment.

A third principle is respect for human beings. Respect in a moral sense must be distinguished from admiration for achievements of other kinds – for instance, for a skillful free kick right into the corner of the goal. Unlike admiration of this kind, respect in a moral sense has prescriptive consequences: One ought to behave towards other human beings in certain ways. One important element is to treat them as ends-in-themselves and not only as means to achieve other purposes, to once again use the Kantian version of an ancient idea. Evidently, reducing somebody to a means of some ulterior purpose degrades this person. As we saw when discussing the trolley problem, substantial empirical data suggest that people do in fact apply this principle when evaluating certain morally salient situations: The recorded moral evaluations of the footbridge scenario and its most promising analysis show that these evaluations are best explained by an operative principle demanding that people not be instrumentalized based on a structural means/ends distinction.Footnote ²⁵

Furthermore, there are other forms of disrespect for human beings that do not instrumentalize them: Regularly flooding a prison cell with feces, not to torture the inmate but simply due to negligence concerning the sanitary conditions of inmates, is an illuminating example of this from constitutional case law.Footnote ²⁶

The importance of recognition as a being of equal worth has rightly been flagged as a key element of social struggles:Footnote ²⁷ These struggles are not only about bread-and-butter issues, but also about exploited and degraded groups of people’s demand to be respected as human beings. This is a central ethical and political dimension in the fight against slavery, old and new, and in the fight against the subjugation and exploitation of women. The same holds for struggles for the emancipation of the working classes, which likewise were not only about material concerns like decent wages, but also about respect for the humanity of workers, who demanded (and indeed still demand) that they not be reduced to beasts of burden.

These findings help us to tackle some of the problems that the supervenience of moral judgments over facts poses.Footnote ²⁸ They help us to identify the principles that determine, first, the moral evaluation and normative consequences of moral judgments that are triggered by certain facts (e.g. by the fact of a direct intention to harm somebody) and thus to determine how moral judgments supervene over facts (not just that they do) and, second, which facts are morally relevant in the first place (e.g. that the agents’ intentions to act matter morally but not their haircuts). These tentatively outlined principles, which are obviously in need of much refinement, are abstract but not without meaningful content, as can be exemplified by the justification of human rights, as we have seen.Footnote ²⁹ To recall: The principles underlying the attribution of rights to persons have to be equal for all potential rights-bearers. It would, for example, be unjust to let some people enjoy fundamental rights because of their personhood and deny these rights to others because for them the color of their skin (and not their personhood) is taken to be relevant. In addition, the reasonable – more precisely, the only reasonable – criterion for the attribution of rights is a person’s humanity. As a just system of rights has to preserve a relation of equality between the value of this criterion and the distribution of rights, and as all humans are equal in their humanity, only a system of equal rights is consequently a just system of rights. The theories of human rights that imply a diminishing humanity of older people and, correspondingly, a diminishing set of protected rights, which may seem to be at odds with this principle, actually confirm it: These theories simply entertain the (implausible) idea that the humanity of elderly people diminishes, even though the elderly, of course, enjoy the same humanity and consequently the same rights as every other human being – no minor point, as the Covid-19 pandemic has reminded us.

Moreover, fostering the enjoyment of rights is morally good. Accordingly, as has been said before, given the importance of the goods that rights protect and the significance of the rights themselves, the promotion of rights is a prima facie obligation of human solidarity. The theory of the justification of human rights has shown, too, that one of its building blocks is respect for human beings, today widely understood as grounded on their dignity. In sum, justice, the altruism of solidarity and respect for the worth of others take us a long way when trying to identify the normative principles that are key to the justification of human rights.

8.2.3 Basic Harms, Human Rights and the “Seeds of a Collective Conscience”

Human rights play an important role in John Mikhail’s pioneering work on the mentalist framework of ethics and law.Footnote ³⁰ His reconstruction of a set of principles of a “universal jurisprudence” based on common moral precepts that are “the seeds of a collective conscience”Footnote ³¹ includes reflections on human rights that constitute some of the most advanced thought on the topic of our inquiry. His main thesis is that the prohibition of basic wrongs that is an element of the inborn structures of human moral cognition is mirrored in fundamental human rights norms.Footnote ³² The universal moral grammar “implies that human beings possess tacit or implicit knowledge of specific, human rights-related norms.”Footnote ³³ The small set of principles that guide human moral judgments include the prohibition of intentional battery, the prohibition of intentional homicide, the rescue principle and the principle of double effect.Footnote ³⁴ Of particular importance for human rights are the prohibitions of intentional battery and homicide, which are “lesser included offenses of a wide range of human rights abuses, including murder, extermination, deportation, torture, rape, genocide, and other crimes against humanity.”Footnote ³⁵ Human rights protect against a set of “core human wrongs” like battery and other torts, “although they by no means exhaust these violations.”Footnote ³⁶ A “clear conceptual and empirical bridge between moral grammar and human rights can be built” because “[t]hese basic wrongs and, in particular, the tort of harmful battery likewise supply the basic perceptual and cognitive tasks of many influential research programs in the cognitive science of morality.”Footnote ³⁷ Mikhail’s work on the trolley problems is itself an outstanding example of this.Footnote ³⁸

These remarks lead to three questions that can be addressed fruitfully in the light of the results of our analysis.

The first is the question of foundational principles. Our analysis suggests that the principles of justice, altruism and respect that are the normative core of our normative theory of human rights are consistent with Mikhail’s analysis but seem to have additional explanatory power. As indicated, the obligation to foster the well-being of others is the flipside of the coin of the prohibition to inflict harm – the minimum you can do for others is not to harm them, we said. It is the normative core of demands for mutual support and solidarity as well. Moreover, human rights are not just about the prohibition of specific wrongs, but also about the allocation of goods, such as normatively protected spheres of liberty. Such an order of freedom presupposes yardsticks for the justified allocation of such goods – in our analysis, for instance, principles of justice. This is important not least for ascertaining the justified limitations of human rights – for instance, when rights of different persons collide. Moreover, human rights only make sense, as we have seen, if there are reasons to believe that the beings who enjoy them have some kind of worth that justifies protecting these beings’ goods. Principles of human worth, today spelled out in human rights ethics and law as human dignity, are therefore likewise key to the understanding of human rights. These principles and their normative consequences, such as the principle of noninstrumentalization, are arguably also key to the solution of the trolley problems.

The second question asks why goods are guaranteed by rights and not some other means in the first place. If normative means to protect human goods are employed (and not just sheer force), why are duties and prohibitions not enough? After all, influential theories argue for the importance of systems of duty and their superiority at least in some respects to systems of rights.Footnote ³⁹ Our analysis has suggested (apart from the analytical connection of duties and rights) that normative principles like justice and altruism provide the answer, as these principles give rise to the normative phenomenon of rights that have the important function of normatively empowering people, turning them from patients of others’ obligations to subjects of justified claims.

The third issue is the question of construction. Our analysis has suggested that a process of complex construction is necessary to transform concrete, principled moral judgments about moral claims into explicit, critically reflected human rights in ethics and law, a process that spanned thousands of years. This process includes only seemingly straightforward tasks like the generalization, universalization and objectification of the abstract core content of particular moral judgments. Moreover, human rights systems are highly selective. To account for this selectivity and its justification, among other things, a theory of human goods is essential. The importance of a political theory of human rights has been underlined in our inquiry as well. It is essential to include these complex issues in a theoretical account of the link between human moral cognition and the idea of explicit human rights in ethics and law.Footnote ⁴⁰

8.2.4 Volitional Consequences of Moral Judgment

A further important element of an analysis of morality is that human moral judgment has volitional consequences: A moral evaluation does not yield information about a fact of the world like a descriptive proposition, but has prescriptive content.Footnote ⁴¹ Saying “X is just” is different from saying “X is a table,” and an important element of this difference is that a moral judgment tells us how we ought to act. This can have direct volitional consequences: An ought creates a motive for action for agents who are evaluating their options for acting.Footnote ⁴² Their will is bound to form certain intentions to act, to use a metaphor for a common introspectively accessible internal state of an agent. Everybody knows how it feels when you come to the conclusion that you ought to hand in the smartphone you have found lying on the ground.

If an observer evaluates not his own options for action but the intention or action of another agent, moral judgments still provide information on how one ought to act. For example, they tell the observer how the observed agent ought to act if the observed agent is in a position to act as a person is obligated to act in the particular situation in question – for instance, to hand in the smartphone found, even though the observer himself is not in this position.

A moral judgment can also remain abstract in the sense that there is nobody currently in a position to act accordingly. “One should hand in the valuables of others one has found” is a meaningful statement even if all valuables on the planet are safely in their owners’ pockets. Moral judgment then tells us what one ought to do if the conditions for this obligation obtain.

The fact that a normative statement is about an obligation, a permission or a prescription – in short, about a moral ought that provides a motivation to act – does not mean that the obligation necessarily forms the predominant, decisive, let alone only motivation to act. To be sure, human motivation encompasses many other inclinations that have great power and have nothing to do with moral considerations. Human history is to a large extent the history of greed and the pursuit of power, not the history of moral niceties. The claim is thus that only an element – and perhaps a precious element – of human moral motivation derives from moral insight.Footnote ⁴³

The prescriptive content of a moral judgment can – and this is important for our particular topic – constitute a right.Footnote ⁴⁴ If an act is just, the agent has an obligation to act justly and the patient has a right to that action: In the limited time available to comment and set things in intellectual order after an abysmal lecture on the foundation of human rights by a legal theoretician from Zürich, every discussant has the right to the same amount of time because this is a just distribution of this scarce good. The chair of the discussion has the obligation to ensure this fair distribution of time – for example, by restraining the loquacious Swiss theoretician’s vacuous responses. The connection between obligation and right holds for an obligation stemming from a duty to benefit somebody, too, unless it is a supererogatory action: Not only is there an obligation to pick up your phone to call an ambulance if somebody in front of you collapses, but the person who has collapsed has a right that you do (at least) this. The obligation to respect others is the correlative of others’ right to be respected: Respect is not an act of grace, but means acting in a way to which others have a claim.

8.2.5 Questions of Metaethics and the World of Moral Emotions

The principles of justice and altruism that guide reflexive evaluation have cognitive content. Whether or not there is (for example) a relation of equality between patients of actions or between a criterion of distribution and the good distributed in the sense explained above is not felt physically in the same way that cold or heat are, for example, but stems from a complex structural analysisFootnote ⁴⁵ of the evaluated act that eventually predicates a relation of equality or its absence and thus constitutes a judgment with cognitive content. The same is true for other structural elements of the action that play a role for moral evaluation, such as agency, the properties of patients of the action (e.g. whether they are sentient or not) and intentions and their kind (direct or oblique) and object. The presence of these elements is not physically felt either, but is ascertained by a judgment with cognitive content – the obtaining of a direct intention to benefit somebody, for instance.

As we already have seen, this does not mean that the relevance of moral sentiments is diminished. They are evidently crucial to the impact of morality on a human life. There may even be emotions that are “geological upheavals” of moral thought.Footnote ⁴⁶ Nor is the importance denied that certain emotions have for the design of law, not least for the skeptical project of constitutionalism.Footnote ⁴⁷ However, insofar as they are moral emotions, these emotions are not simply emotions relevant for ethics (like the fear of tyrants) but are the consequence of moral judgment with the cognitive content just identified, and thus do not constitute moral judgment. The moral evaluation based on the obtaining of the necessary elements of a good or just intention or action is the precondition for the experience of moral emotions – for instance, the unequal treatment of a person who is equal in the relevant respects in question is the precondition for the feeling of indignation about injustice. If there is no unequal treatment of equals, we will not feel moral indignation because of an injustice.

In light of this observation, traditional arguments about the constitutive role played by human moral sentiment in moral evaluation fail to convince because they are not precise enough. Hume wrote:

Our argument so far indicates that two elements of the structure of moral evaluation are missing from Hume’s account. Hume rightly emphasizes the importance of ascertaining the facts of a case to be evaluated. What he does not address, however, is the structural analysis of the intention or action that determines the outcome of the evaluation – for example, the analysis of whether Norbert killed another person with the direct intention to do so for personal gain or whether he did so to defend himself against an aggressor. We have already clarified that this analysis is not a sentiment, but a judgment with cognitive content. The second missing element is the evaluation that ensues on the basis of this analysis – for example, that killing a person for personal gain is (deeply) immoral. This evaluation assigns a deontic status (good/bad, just/unjust) to an intention or action. This is the propositional content of statements such as “Murder is a heinous crime.” The moral judgment causes and is accompanied by a sentiment – for instance, the feeling of abhorrence towards murder for gain. The feeling of abhorrence does not constitute the entirety of the evaluation of murder for gain, however.

Interestingly, Hume rightly draws our attention to agency as a structural precondition of evaluating something as morally right or wrong. In Hume’s example of the young tree outstripping and eventually overwhelming and killing its parent, the tree’s growth does not elicit any moral feeling because a precondition of moral evaluation, agency, is not fulfilled.Footnote ⁴⁹ Once again, whether or not a tree is an agent is not felt in the same way as our skin feels a cold breeze. Rather, any such conclusion is a statement about a complex state of affairs constituting agency (responsible, spontaneous beginning of new chains of causation, authorship of intentions to act and of actions, etc.). This cannot be reconciled with simply identifying moral judgment with sentiment. There certainly are moral sentiments, but they depend on such a prior structural analysis of the eliciting situations, including of criteria such as agency, and a consequent moral evaluation.

A central concern for Hume and the many thinkers following in his footsteps is to explain moral motivation. How could a judgment of reason cause a motivation to act? A proposition stating a matter of fact has no motivational impact. “There is a table” is, even if true, motivationally neutral, as we clarified in the preceding section. Only sentiment, it thus seems, can cause people to act. There is, however, a third path beyond the traditional but not exhaustive dichotomy of moral judgment as an act of reason and as sentiment. This third path offers the key to the problem: The motivational effects of moral judgments we highlighted above are crucial in this respect. Moral evaluation has a direct motivational effect, we said, because of its prescriptive content. Moral evaluation is not limited to stating the deontic status of an intention or action. It does not merely inform us about this status like a proposition about a fact of the world. Agents do not react to moral judgments in the same way they react to factual propositions. They do not say after a moral judgment, “Oh, this is unjust, how interesting!” in the same way they may observe, following some scrutiny, “Oh, this stone is in fact blue, how interesting!” Moral evaluation gives rises to a moral ought, and this ought alone can motivate people to act: It binds the human will, to use this basic metaphor. This is very much an everyday experience. Take the lost object example: You see a wallet that somebody has lost lying in the street. You ask yourself whether you should pocket the money in it. You look around – nobody is watching you. Do you do it? Perhaps you do not, because you have come to the conclusion that taking the money would harm the owner and thus would constitute an immoral act. This moral evaluation does not leave you untouched, because it creates an obligation not to do what is immoral that affects your will and may ultimately make you hand in the wallet. Perhaps you overcome this moral impulse by thinking of something nice you want to buy with the money, but even in this case the moral impulse is not nothing. It was simply not strong enough to outweigh the attraction of the thing you want to buy. Unsurprisingly, the direct motivational force of the moral ought has not escaped the attention of important contributions to moral philosophy.

In discussions about the nature of morality, we sometimes find the idea that a cognitivist approach to moral judgment is wedded to conscious, explicit reasoning based on deontological moral rules. As there is ample evidence for spontaneous moral judgments, such intuitions are understood as showing that moral emotivism is correct: These intuitions are emotions, not acts of reason-based cognition. This assumption, too, fails to fathom the nature of morality. Evidently, a structural analysis of the kind discussed – for example, as regards agency – is a quick, largely unconscious process that is not necessarily transparent to the agents themselves. However, this does not mean that the structural analysis is based on a sentiment – it clearly is not, as we just have seen. One thus has to broaden the analysis of moral judgment and include these structural analysis mechanisms, which, while they are largely intuitive, are not simply emotive. Such mechanisms can be made explicit, as Hume did in the case of agency, a process very important for the understanding and practice of morality.

Thus, it turns out that the simple dichotomy between moral judgments as cognitive acts of reason or as expressions of sentiment does not adequately capture the intricacies and rich content of moral judgments, their cognitive, volitional and emotional dimensions that need to be differentiated and accounted for in a theory of human moral cognition.

To repeat: None of this doubts the importance of moral feeling. It is no less than one of the most important elements of human identity. It makes morality the powerful force that it sometimes is in human life. The rich colors of moral emotions are sources of both the beauty and the profound sorrows of human life. These remarks simply intend to clarify what the role of these emotions in moral thought actually is.

A further central function of emotion in moral judgment is to fathom what an action means for the patient of the action. It is a piece of moral heuristics: Without empathy for the experience of victims of racial discrimination, for instance, without the emotional understanding of how it feels to be degraded and humiliated, nobody will be able properly to evaluate the significance of this injustice. If one thinks that being relegated to the back of a bus is a minor issue, the issue that the Freedom Riders in the USA were fighting for will remain inexplicable.

8.3.1 Ruled by Moral Taste Buds?

Contemporary contributions to moral philosophy and psychology with a – broadly understood – emotivist background do not call these findings into question. We have already investigated the explanatory power of the mental gizmo thesis. A further influential perspective, the social intuitionist model or moral foundations theory by Jonathan Haidt, does not change this perception either. This theory argues along emotivist lines that emotional intuitions form the basis for moral judgments. Arguments that appear to be rational are used to justify such emotion-based moral attitudes post hoc. Moral reasoning is not engaged in to critically improve one’s own moral point of view.Footnote ⁵⁰ Rather, moral argument is directed at manipulating others – in fact, reason is a “public relation firm” of emotional intuitions for “strategic purposes such as managing reputation, building alliances, and recruiting bystanders to support your side.”Footnote ⁵¹ At the same time, however, rational argument is supposed to have an influence on these emotional reactions, although the theory does not clarify exactly how this influence is to be understood, in particular whether reasoning causes and changes moral judgments or overrides moral judgment.Footnote ⁵²

Jonathan Haidt claims that there are six “moral taste buds of the righteous mind.”Footnote ⁵³ Haidt intends to expand the ethics of autonomy of WEIRD (Western, educated, industrial, rich and democratic) people (who form the sample of many psychological studies) to include moral systems that are “hierarchical, punitive, and religious” because they include ideas of loyalty, authority and sanctity.Footnote ⁵⁴ The six foundations of morality are the pairs care/harm, fairness/cheating, loyalty/betrayal, authority/subversion, sanctity/degradation and liberty/oppression.Footnote ⁵⁵ Fairness is about the allocation of goods proportionally to merit; demands for equality are matters of nondomination and therefore are based on the liberty taste bud.Footnote ⁵⁶ These pairs are moral modules that form the basis and common ground upon which all of the different human moralities develop. They define the possible contents of any given moral code. It is a misconception, Haidt argues, to focus on just some of these foundations as Western individualism does – for example, overlooking other moralities where loyalty and sanctity are of great importance.

The thrust of the argument is illustrated well by the following example:

This, then, explains why “good people” can disagree about right and wrong (for instance, the right and wrong of spanking) – they are simply operating in different moral matrices.Footnote ⁵⁸

Haidt uses these moral matrices to explain US politics. If care, fairness and liberty are dominant, one becomes a liberal, if loyalty, authority and sanctity are guiding, a conservative.Footnote ⁵⁹ The dominant influence of these taste buds is genetically fixed.Footnote ⁶⁰ These moral taste buds are explained by their evolutionary functions. The moral foundations have evolved, it is argued, in response to the adaptive challenges of “caring for vulnerable children” (care/harm); of “reaping the rewards of cooperation without getting exploited” (fairness/cheating); of “forming and maintaining coalitions” (loyalty/betrayal); of “forging relationships that will benefit us within social hierarchies” (authority/subversion); and of “the omnivore’s dilemma, and then to the broader challenge of living in a world of pathogens and parasites” (sanctity/degradation).Footnote ⁶¹ There is no clear account of why the liberty/oppression foundation evolved.

Group selection and gene–culture coevolution may have played a role in this evolutionary development, it is argued.Footnote ⁶² The point of morality is to enable cooperation, albeit limited to the groups of which the agent is a member. Parochial altruism is the ultimate frontier of ethics. Oxytocin is important for explaining this within the framework of the theory of group selection: “Oxytocin should bond us to our partners and our groups, so that we can more effectively compete with other groups. It should not bond us to humanity in general. Several studies have validated this prediction.”Footnote ⁶³ The benefits of cooperation also are why groups evolved that entertain religious beliefs: These beliefs make people better cooperators.Footnote ⁶⁴

Human rights are explained in this framework: They are post-hoc rationalizations of the intuitions of people with particularly receptive liberty taste buds.Footnote ⁶⁵ The different moral matrices complement each other – liberals are, for instance, right about the necessity of some regulation, while conservatives are right that markets are “miraculous.”Footnote ⁶⁶

This psychological theory is descriptive, it is underlined. Haidt notes that if taken as a normative definition of morality, the simple acceptance of loyalty and authority could lead to a defense of political orders such as fascism, which would be given “high marks” as long as they produced high levels of cooperation.Footnote ⁶⁷ For questions of normative theory, therefore, other arguments are necessary – more precisely, what is needed is rule utilitarianism, which should be decisive and guide public policy decisions that take into account the importance of loyalty, authority and sanctity.Footnote ⁶⁸ For the individual, virtue ethics is tentatively endorsed.Footnote ⁶⁹

The overall vision of morality presented is a narrow one: “Parochial love – love within groups – amplified by similarity, a sense of shared fate, and the suppression of free riders, may be the most we can accomplish.”Footnote ⁷⁰

8.3.2 A Testing Case: Corporal Punishment – A Question of Taste?

Spanking is an interesting example for assessing moral foundations theory in the context of human rights, because the corporal punishment of children is a standard human rights issue and in a leading decision of the ECtHR, for instance, was declared to violate Art. 3 ECHR on inhuman or degrading treatment or punishment.Footnote ⁷¹

Was this decision just an emotional gut reaction on the part of the majority of judges of the ECtHR, who were led by their particular moral taste buds in their interpretation of Art. 3 ECHR, while the dissenting judge arguing for the permissibility of this kind of sanction simply used other taste buds?

This example reveals a major shortcoming of moral foundations theory quite clearly: It does not distinguish between factual, traditional morality and critically reflected ethics. One major point of moral philosophy is to critically investigate certain moral issues that may turn out not to stand the test of such critical scrutiny, such as the morality of loyalty to fellow slaveholders or respect for the sanctity of moral rules subjugating women – or the fairness of the physical punishment of children. The effects of this critical scrutiny can make history – as the abolition of slavery and the women’s liberation movement illustrate. Only critical reflection on morality, based on considered judgments, will tell us what human morality ultimately is about, while factual, traditional morality, by contrast, will tell us about precisely those influences that skew moral judgment – say, racist ideas, misogyny or an authoritarian tradition of child-rearing. Studying a traditional morality that sanctions violence against children as a good starting point for understanding moral cognition is comparable (with a dose of exaggeration) to studying mathematical cognition by looking at the belief that 79 + 86 = 164. Such studies may tell us a lot about the factors influencing counting, such as a lack of attention, but they are not key to the psychological foundations of mathematics.

Let us take a closer look at the example of corporal punishment to better understand what the further problems of the moral foundations theory may be. There is much to be said about this example, particularly in relation to children. Thinking about the effects of such punishment on children, its educational merit, brutalizing consequences and so on, is an important exercise if one wants to do justice to children. None of this means simply to rationalize emotional intuitive reactions. Rather, this exercise clarifies the issues – in particular, whether corporal punishment does any good or simply harms children, as very many people (rightly) assume today, and whether it should come with corresponding legal consequences. Beyond these important arguments, the ultimate moral evaluation relies on the elements of the action identified above – for instance, the kind of intentions of the agent of corporal punishment. If the intention is to gratify the cruel instincts of the punishing adult, the evaluation will differ from the evaluation of the ill-advised but well-meaning intention of a loving parent who uses corporal punishment as means of education, even though in the latter case, too, it is ultimately an unjustified harm done to children.

This search for moral understanding is the daily business of any responsible human being. It is thus analytically unconvincing to regard morality simply as a manipulation device to strategically influence others, not as a guide to one’s own actions. One distinguishing feature of moral judgment is that is aspires to correctness based on reasons – and there are certainly good reasons for certain forms of action. This is what motivates the painstaking efforts of many human beings who honestly want to get it right. Interestingly, the social intuitionist model underlines this itself because it includes reflection in its reconstruction of morality, albeit ambivalently and without clarifying its precise role. The fact that it refers to rule utilitarianism as the ultimate yardstick for the evaluation of social policy and as a bar to “high marks” being awarded to fascism bears witness to this. By trying to limit the harmful conclusions that may be drawn from its arguments by reference to the principles of an ethical maxim such as rule utilitarianism, moral foundations theory denies intuitions generated by moral taste buds the ultimate rule that it appears to defend. In this context, we can note once again that utilitarianism itself is based on notions of equality that are not the consequence, but the normative precondition of utilitarian evaluation. Utilitarianism takes us right back to moral principles of justice and offers no escape route from them, as we have seen. Moral foundations theory’s inconsistent impression is only deepened by making virtue ethics a yardstick for individual action. Virtue ethics is, like all ethical systems, contested. However, it does not endorse the rule of emotional intuitions as the core of ethics.Footnote ⁷²

These observations are important for assessing the merits of the six foundations of morality. They point to further deficiencies in the analysis of morality: Care and harm, fairness and betrayal point to notions of justice and altruism, albeit not very precisely, as betrayal, for instance, is simply one form of unfairness. The relationship of proportionality and equality is analytically misunderstood, too: If the criterion of distribution is equally fulfilled in two agents, equal treatment is the consequence of principles of proportional equality, as we have seen. Therefore, for example, the equality of rights is a consequence of proportional equality: The humanity of humans as a criterion for the enjoyment of rights is simply the same for each person.

Liberty is a central good for human beings, to be sure. Moral foundations theory, however, does not discuss the normative principles that are relevant in balancing the liberty of one with the liberty of all and that may even create rights. The reference to a “liberty taste bud” can be seen as a metaphor that can be reconciled with (but is no substitute for) the theory of liberty as a human good sketched above. However, this reference does not answer the problem faced by any normative theory of liberty that has been at the heart of many debates since antiquity, namely the question of who justifiably enjoys which liberty and to which degree in relation to others.

Another example: Loyalty, authority and sanctity are only secondary virtues. Loyalty and respect for authority are morally justifiable if they serve the good of people; loyalty to dictators and reference of their authority is not a virtue but a vice, as Haidt himself admits, realizing that its content may cause “high marks” to be awarded to fascism if it manages to achieve social cooperation. Sanctity is a difficult concept, too. Respect for the sanctity of a certain conception of marriage may be bad news for gay couples. Here, too, other moral principles, such as respect for the autonomy of other human beings and their dignity, need to be considered.

Moreover, the theory does not engage in any detail with what basic notions of justice or altruism actually entail – for example, as to the intentions of the agents, the effects of actions, means/ends distinctions and so forth, as roughly outlined above.Footnote ⁷³

Finally, as a last point: The examples of political differences explained by the different operation of the moral taste buds are loaded with prudential arguments, such as arguments about the question of when regulation actually works or whether and under which conditions markets deliver good results. Regulatory choices or the design of markets evidently have ethical implications. But discussing such prudential arguments as part of the core normative machinery of morality does not help to clarify what morality is about.

This notwithstanding, these analytical shortcomings teach us a constructive lesson: They show that the six categories of moral foundations theory do not adequately capture the normative principles that are central for ethical systems. In light of its critical discussion, the principles of justice, altruism and respect continue to seem valuable candidates for the foundational elements of ethics that a psychological theory needs to account for.

The evolutionary framework of moral foundations theory suffers from a failure to present any evidence whatsoever that the evolutionary story it recounts is actually true. It remains one of those “just-so” stories that are falsely taken as valid evolutionary theory-building.Footnote ⁷⁴ Moreover, as indicated above, the first step for any convincing evolutionary theory of morality is to construct a theory of morality that possesses sufficient analytical precision. As this is lacking, Haidt’s evolutionary theory is unable to get off the ground.

Moreover, as we have seen, the idea of human rights has highly complex roots in history and human normative reflection. To reduce it to the effects of a liberty taste bud and the post-hoc rationalization of its operations seems not to do full justice to these findings.

The social intuitionist model illustrates a danger already identified previously: When the idea that corporal punishment is admissible, for instance, is interpreted as the expression of certain moral taste buds that simply are different from the taste buds used to criticize this form of punishment, the criticism of corporal punishment loses its ground because there is no reason to prefer the one taste over the other. The theory tries to deal with this problem by introducing rule utilitarianism as a normative yardstick for public policy and virtue ethics for individual acting, contradicting its emotivist message in doing so. Nevertheless, ideas of loyalty to in-groups, respect for authority without questioning its legitimacy and reverence for sanctity without inquiring into its origin and content are understood as expressions of moral judgment on par with any other moral code, in particular one based on human autonomy. In this way, psychological theory may serve to shield unjustified moral precepts against criticism by presenting them as products of a genetically fixed and thus unchangeable natural morality.

In line with various theorists of morality, moral foundations theory argues that only a morality of “love within groups,” a “parochial altruism” is the horizon of human moral possibilities. Concern for all humanity is simply too much for human beings. The human rights project, which is a real practice, not an ephemeral dream, indicates that these theses do not fathom what morality is really about.

8.3.3 Sentimental Rules

In another very interesting approach, Shaun Nichols argues for a sentimentalist picture of morality. It takes as its starting point the body of data suggesting that children from an early age distinguish between conventional and nonconventional moral rules.Footnote ⁷⁵ The former can be changed by authorities, the latter not. The former deal with prudential arrangements, the latter are centered on the avoidance of harm. The latter are more serious than the former. The latter can be generalized over situations and contexts, the former less so.Footnote ⁷⁶ Moreover, moral objectivism is the psychological default position, Nichols argues.Footnote ⁷⁷

Nichols claims that moral rules are based on moral judgment, which he argues has two components: information about normative violations and a noncognitive response. The former is provided by rules proscribing certain behavior, while the latter consists of specific emotions. Only those rules relating to behavior that triggers reactive concern and distress cues are nonconventional moral rules.Footnote ⁷⁸ It is the particular emotional engagement of the evaluating person that turns them into moral rules. These emotions are also the reason why these rules are able to survive historical change and development.Footnote ⁷⁹

As we saw above, the mental gizmo theory already claimed that emotions explain moral judgment such as that reached in the footbridge case, namely that it is impermissible to kill the bystander to save five lives. The sentimental rules approach goes further than the mental gizmo theory, however, in claiming that this is not only because of the emotionally salient features of the act. Rules on the prohibition to perform this act are also important. However, these rules need to be backed by emotions in order to yield the impermissibility judgments observed in the footbridge case.Footnote ⁸⁰ A third factor influencing moral judgment consists of cost–benefit assessments along utilitarian lines.Footnote ⁸¹ These factors interact in a complex way, without there being a unified normative theory that captures all moral intuitions.

Metaethically, morality therefore is not objective, but bound to feelings relative to particular groups of people. There is no argument why one should prefer one set of emotions to another. It is not possible to evaluate the emotions determining the content of morality with these very emotions.Footnote ⁸² The illusion of objectivity hides the respondent-dependent relativity of emotionally based sentimental rules. This notwithstanding, the objectivist intuition persists and guides human lives.Footnote ⁸³

This theory faces the same problems already identified as decisive for emotivist accounts of moral judgment: The emotional reaction to some harm is not all there is to moral judgment. A harm may cause considerable unease without it being immoral – to use Nichols’ own example, a medically justified operation, for instance. It is only immoral if the intention and the action have certain properties – for instance, the intention to torture the patient and not to heal them. These properties need to be determined based on an analysis of the intention and action, as outlined above in some detail. This analysis then elicits the moral evaluation. This, in turn, gives rise to moral feelings, which are different and need to be distinguished from feelings such as unease at witnessing an operation. Feeling nauseous at the sight of blood during such an operation is not already a moral feeling, while feeling abhorrence because of the wrongness of an attempt to torture a person is. The evaluation of an action as morally wrong is, however, the precondition and cause of these moral feelings. Like an Escher cube, the same action can change its moral nature, depending on what evaluation the analysis of the action’s intentions and effects yields.

Nichols developed a theory of moral learning that adds a further qualification, underlining another influence that reason has on moral judgment. It is best considered in the framework of the discussion in Section 8.5 on the acquisition of moral knowledge.

8.5.1 How Do We Learn to Be Moral?

One question that remains to be explored is the ontogenetic origin of moral cognition. The development of moral cognition forms the subject of landmark debates in moral psychology.Footnote ⁹³ Our analysis thus far suggests that central issues here include – among other factors – the acquisition of the cognitive domain of morality, the restrictive principles that determine the possible objects of evaluation, the material principles of morality discussed, the prescriptive, volitional effects of moral judgments, the moral ought, including necessary connections between duties and rights, and the emotional consequences of moral experience. Such phenomena could be constructed and acquired by secondary learning processes (instruction, repetition, role-taking, peer pressure, sanctioning, etc.), as many normative principles are (e.g. the intricacies of Swiss law on unjust enrichment), or alternatively they could be at least in part the product of the unfolding of innate cognitive structures triggered by experience. The latter is the way a mentalist approach to ethics would approach the issue.

It is important to emphasize that the importance of social and cultural influences on ethics and law is not denied in the latter case: As noted, the freedom of the press, for instance, as a legal norm and the underlying principles of political morality presuppose the cultural achievement (a late achievement with plenty of preconditions) of the press and, in addition, the experience of its suppression even by democratic governments, among many other things. “The freedom of the press is protected” is certainly not an inborn principle of human moral cognition. As we saw in our survey of the history and the theory of justification of human rights, fundamental moral judgments about the wrongness and justice of certain intentions and actions are the seeds of the idea of human rights. The real question is therefore: What kind of mind do you need to possess in order to develop such an idea? What is so special about the human mind that only humans and no other organism developed a concept like human rights? What are the cognitive preconditions for starting the long cultural process leading, after many thousands of years of human cultural and social development, to the idea that one can not only wish for or have an interest in, but in fact enjoy a right to a free press? None? Is simply being smart enough? Or does one need to have certain specific conceptual tools with which to build a system of rights?

The simple observation already recalled above that any child acquires a differentiated set of moral concepts and categories whereas no nonhuman animal acquires any of the same, even if it is raised in the same environment, seems to indicate strongly that very different cognitive abilities are in place. Some of the empirical work on this difference between humans and nonhuman animals has been discussed in Chapter 7. To repeat: The question therefore is not whether there are species-specific inborn cognitive structures enabling the formation of moral precepts, but what exactly these cognitive structures are.

8.5.2 The Moral World of Infants and Toddlers

Jean Piaget pioneered the idea that children apply complex moral ideas and concepts, not least when they play, as illustrated by his fascinating analysis of children playing with marbles.Footnote ⁹⁴ Recent years have produced many more highly creative studies on the moral psychology of children, increasingly focusing on young children. We have already encountered some important examples in our inquiry. As we have seen, these studies operate with different theoretical background assumptions and sometimes yield contradictory results. Some research is constrained by theoretical and conceptual problems, limits of the design of experiments and contestable interpretations of their results. One important problem is the lack of a sufficiently fine-grained analytical theory of moral judgment that does not commit, for instance, the category error of conflating moral judgment with simple preferences and aversions. The thrust of this research is, however, that children operate in a richly textured moral world from very early on. This fits well with research into other cognitive domains that also indicates that human beings are not born blank slates.Footnote ⁹⁵

A classical set of experiments concerns the moral–conventional distinction, for instance: As mentioned above, children distinguish between conventional norms and moral, nonconventional norms.Footnote ⁹⁶ Other experiments suggest that toddlers by the age of eighteen months and even earlier exhibit spontaneous helping behavior towards others.Footnote ⁹⁷ They engage in so-called paternalistic helping: Children want to accommodate the well-being of others, not just their wishes.Footnote ⁹⁸

Children act with an intrinsic moral motivation: They also help when nobody is watching or when others do not know that they are being helped.Footnote ⁹⁹ External rewards even seem to undermine their intrinsic motivation.Footnote ¹⁰⁰ Helping somebody themselves and seeing somebody being helped render them equally content.Footnote ¹⁰¹ There are studies on extended sympathy – for instance, for the victims of some harmful action after the deed.Footnote ¹⁰²

Studies indicate that – unlike great apes – small children act according to principles of distributive egalitarian justice, particularly in collaborative contexts.Footnote ¹⁰³ While there are studies suggesting that children limit fairness to in-group members,Footnote ¹⁰⁴ there is also substantial evidence that three-year-olds apply egalitarian principles universally.Footnote ¹⁰⁵ Such intuitions extend to procedures of distribution.Footnote ¹⁰⁶ Three- to five-year-olds take need or merit as criteria of distributive justice. They protect the entitlements of others, create social norms, employ normative categories and engage in third-party punishment.Footnote ¹⁰⁷ Intentions are a central factor in this punishment.Footnote ¹⁰⁸ There are studies on the specific moral feelings of guilt and shame – for instance, arguing that three-year-old children feel guilt over the harm they inflict, not just sympathy for the victims of this harm.Footnote ¹⁰⁹

According to various studies, it is plausible to assume that infants already operate with at least some kind of proto-morality. Twelve-month-olds categorize actions on the basis of their social valency.Footnote ¹¹⁰ Studies on social evaluation indicate that six- to ten-month-old infants base their evaluation of another individual as appealing or aversive on this individual’s actions towards others: For instance, they like individuals better who help others and who act more cooperatively to facilitate the achievement of these others’ aims than those who hinder the achievement of others’ goals. They also like helpers better than neutral individuals, and the latter better than hinderers.Footnote ¹¹¹ These experiments concern actors who are unknown to the infants and actions that have no effect on them as observers. This rightly has been identified as an important element of moral judgment: It is not personal experience with the agent that causes the agent to be evaluated a certain way, but the latter’s action affecting unrelated others. Eight-month-olds showed a preference for individuals with good intentions, basing their evaluation on mental states of the agent, not outcome.Footnote ¹¹² Other research indicates that actions towards inanimate entities do not enter into infants’ evaluation of intentional agents.Footnote ¹¹³ Knowledge of an agent’s goals is relevant for ten-month-olds’ evaluations: Only an agent who knowingly helps another is preferred to other agents. It is not far-fetched to draw the conclusion from this research on infants that the capacity to evaluate others based on their behavior towards third parties is universal and unlearned, given their early age. It should be noted once again, however, that a social preference is not the same as a moral evaluation of an intention or action. The former is at best indirect evidence for the latter.

Many of these studies understand these mental capacities as biological adaptations for cooperation and interpret them as tools to identify free riders, cooperators and reciprocators. As we have seen already, this biological interpretation is too narrow and overlooks the many forms of cooperation that are both possible and entirely reconcilable with the constraints of a plausible evolutionary trajectory.

8.7.1 The Power of Statistical Learning

It is instructive to look at an account of moral learning developed by Shaun Nichols that takes seriously the argument made thus far, namely that there is no direct input from parents and peers that explains the development of moral cognition.Footnote ¹²¹ He underlines that there is a consensus that for some capacities an empiricist account and for other capacities a nativist account is more plausible.Footnote ¹²² His account, already referred to briefly above, complements his moral sentimentalism with a rationalist elementFootnote ¹²³ and presents an alternative to assuming that concrete moral rules are based on certain inborn structures, claiming that statistical Bayesian learning abilities suffice to infer the content of the existing rules of a community from the given evidence of identified violations of rules.Footnote ¹²⁴ This is how deontological rules arise: “Our hypothesis is that non-utilitarian judgement derives from learning narrow-scope rules, i.e. rules that prohibit intentionally producing an outcome, in a way that approximates Bayesian learning.”Footnote ¹²⁵ Nevertheless, there is a nativist element in his theory, or his theory is at least consistent with a nativist approach: Learners have an aptitude for concepts like agent, intention and cause, and the capacity for acquiring rules.Footnote ¹²⁶

More concretely, Nichols’ argument takes its start from the observation that a distinct evaluation of doing on the one hand and allowing on the other is a reality of human moral psychology. This distinction presupposes that humans apply act-based rules, not consequence-based rules. Moral rules are about prohibitions of or prescriptions for action and not about minimizing unwelcome outcomes or maximizing welcome consequences.

Nichols goes on to ask how children can learn that the moral rules that they apply are act-based, not consequence-based.Footnote ¹²⁷ His argument is that they are exposed to act-based input by adults, such as “Don’t do X!” or “Do Y!” Through mechanisms of statistical learning, the children acquire act-based rules, rather than consequence-based rules, as experimental evidence confirms in his view. The core statistical mechanism is the size principle: When a learner has to choose between two hypotheses, one of which is a nested subset in the other, it is rational on probabilistic grounds to choose the hypothesis with the smaller scope if all of the evidence available is consistent with this smaller hypothesis. This is because there is a higher likelihood that the smaller hypothesis is in fact correct. It would form a “suspicious coincidence” if all of the evidence falls in the smaller hypothesis while the larger hypothesis is true.Footnote ¹²⁸ Humans are able to perform this statistical operation because they possess substantial statistical learning abilities, Nichols holds.Footnote ¹²⁹

As the input that learners are exposed to is based on the application of act-based rules (for instance, the command “Don’t do X!”), this evidence is consistent with both the hypothesis that act-based rules are applied and the hypothesis that consequence-based rules are applied, the latter hypothesis also including act-based rules. As the first hypothesis is a nested subset of the latter and the evidence that the learners is exposed to is act-based, it is rational to conclude that act-based rules are applied. Act-based rules are thus acquired by the learners. Nichols concludes, albeit somewhat hesitatingly, that this learning process could explain why people apply the principle of double effect – for instance, in the trolley cases.Footnote ¹³⁰ It can also account for the acquisition of parochial moral codes.Footnote ¹³¹

Furthermore, there is evidence that humans expect new rules to be act-based, showing a “pronounced prior” in this respect.Footnote ¹³² The reasons for this are overhypotheses in the sense defined by Nelson Goodman – because the experience of rules consists of act-based rules, an overhypothesis is formed that rules tend to be act-based.Footnote ¹³³

Normative systems need a default principle of the normative status of those intentions and acts that are not explicitly prohibited or permitted. Everything that is not prohibited is allowed or everything that is not permitted is prohibited are such closure principles, for instance.

Nichols argues that learners acquire the one or the other depending on the input received: If they encounter permissions, they conclude that what is not permitted is prohibited. If, however, they encounter prohibitions, they conclude that what is not prohibited is permitted. This conclusion is based on pedagogical sampling: Learners assume that their teacher is using methods of rational and efficient instruction. Based on this principle, it is rational for learners to conclude that if they are only presented with prohibitions of certain acts by the instructor, then other acts are permitted, and vice versa.Footnote ¹³⁴

Nichols also explains perceptions about the universal or relative validity of norms along these lines. If learners encounter widespread consensus, they acquire the idea that the respective norms are universally valid, while if they encounter disagreement, they conclude that these rules are of relative validity. The statistical principle at play is the trade-off between fit and flexibility: The hypothesis has to fit the data but must remain flexible so as to accommodate other factors producing consensus or variety of opinion (for instance, unreliability of the persons evaluating the topic), without becoming too flexible and thus empty, accommodating any data.Footnote ¹³⁵ These mechanisms also explain the distinction between moral and conventional rules. If people (on statistical grounds) judge an action as universally right or wrong, they should regard it as a wrong in itself, independent of authority, too.Footnote ¹³⁶

8.7.2 The Limits of Statistical Learning

Importantly, this approach rightly underlines the significance of a descriptively adequate account of morality. It acknowledges the necessity for theory-building to determine properly the “acquirendum,” the mental structure that human beings actually acquire in the moral domain.Footnote ¹³⁷ It convincingly refutes various theories that identify morality as the expression of a small set of primitive emotions, seeking to deny that the operations of moral cognition are based on a rich systems of structures, rules, principles and representations.Footnote ¹³⁸ It applies the poverty of stimulus test but arrives at the result that there is in fact no such poverty: The input available to children and statistical principles suffice to acquire the basic elements of the moral world of human beings it investigates. The theory is, therefore, an important constructive contribution to the understanding of human moral ontogeny.

However, several problems arise with regard to the reach of statistical learning. As just discussed, cognitive access to basic elements of the moral world is the precondition for understanding what moral judgment and explicit norms are about in the first place. You need to know what ought means, for instance, before you can grasp what you ought to do, as Hume already argued. Statistical learning is unable to bridge this gap. No quantity of references to ought by others will tell you what ought means from a first-person perspective, just as no quantity of references by others to the pleasant smell of coffee in the morning can teach you the nature of this odor from a first-person perspective if you cannot smell. The same holds for the other elements of human beings’ moral world, including moral emotions like shame.

Another problem stemming from an insufficient determination of the acquirendum consists in the following: The moral principles human beings acquire are not captured with sufficient precision if one looks only at the doing/allowing distinction. The distinction between (intentionally) doing something and allowing X to happen when you have a duty to act to prevent X from happening on the one hand and (intentionally) allowing X to happen when there is no such duty on the other comes closer to a central element of human moral cognition.Footnote ¹³⁹ The prohibition against a girl hitting her brother as punishment for taking her ball is morally equivalent to the prohibition against the child not preventing her younger brother from falling from a swing as punishment for taking her ball when she is playing with her sibling at a playground and has a duty to see that her younger brother does not hurt himself. In contrast, there is no equivalent moral duty to prevent all possible harm to all other children on the playground. Unsurprisingly, the relevance of this distinction is mirrored in basic provisions on criminal omissions in penal law around the world.

Moreover, children do not just learn that rules are act-based. They acquire the principle that internal states like direct intention (purpose) or oblique intention (knowing) matter when evaluating an action. Moreover, these internal states are relevant in complex ways for moral assessment: If one commits armed robbery of a bank with the purpose of getting money and knowingly creates a risk that a guard will be killed, and this indeed happens, in many legal systems this will count as intentional (knowing) homicide. The foreseen but not intended negative side effect of a purposive morally wrong act only makes it worse. In other constellations, for instance, when one intentionally acts to prevent harm but foresees harm to third parties as a side effect of the benevolent action, the act may be justified, as in the standard trolley bystander case. Thus, the acquirendum seems to be of a different and much more complex nature than is allowed for in the doing/allowing distinction that Nichols addresses. That any of these acquired complex structures can be learned through the statistical operations Nichols employs is far from clear, given our findings thus far. Take his example of intentional act-based rules. The evidence he considers (“Do not do X!” etc.) allows for the conclusion that somebody wants the agent to refrain from doing X, perhaps backing the prohibition to do X with the threat of sanctions. It does not allow for the conclusion that one is to refrain from acting with a specific internal state – for instance, an intention – or that one is to see intentions and actions causing harmful foreseen consequences as sometimes prohibited, sometimes justified, let alone that one is to acquire the capabilities to experience in oneself a moral ought and the other cognitive, volitional and emotional elements of the subjective world of morality already discussed above.Footnote ¹⁴⁰

The discussion of closure principles is also of interest. It tells us something relevant about the reactions to explicit instruction by prohibitions or permissions concerning the specific type of action (in Nichols’ study, for instance, mice entering a barn).Footnote ¹⁴¹ For moral theory-building, however, the problem that closure principles deal with is a different one. It concerns the question of how people assess the permissibility of action even without explicit permissions or prohibitions concerning a specific type of action: Do you need an explicit prohibition to take an action as prohibited or do you need an explicit permission to regard it as permitted? What is the default principle (if there is indeed one) if there are no such explicit prohibitions or permissions? Nichols does not investigate this question because he is concerned with reactions after explicit instruction, though he indicates that there may indeed be a default principle of liberty. Only if one addresses this question, however, can one reasonably discuss the problem of how such a principle – for instance, of residual liberty – becomes part of the moral and (in liberal orders) legal world of human beings.Footnote ¹⁴²

As to the discussion of universalism and relativism, it is important that the validity of a norm is not dependent on the quantity of assent it finds, but on material validity conditions. Thus, the norm that Jews are just as entitled to life as other human beings was valid even in German extermination camps. Statistical learning thus allows for the hypothesis that many people agree about the validity of a certain norm, not that it is (in fact) valid. So-called authority independence is not related to consensus either. The point is rather that such universally valid, authority-independent norms are the basis to challenge even majority opinions. The phenomenon of critical reflection creates the possibility of changing a consensus using reasons – a process in which the universality of some norms is a possible argument. In Nichols’ account, any norm could be understood as universal and authority-independent, as its universality depends solely on factually existing consent patterns. If the input that a child is exposed to consists of a consensus that it is justified to kill Jews, this norm enjoys universal validity, for instance. Nichols is not very clear on when and under which conditions the acquired rules change, but it seems that this change is dependent merely on new experiences of consensus or varieties of opinion. This account appears to miss the point that content matters for the question of universality, and that only certain content is a serious candidate for universally valid norms – for instance, norms prohibiting harm, but not norms prescribing gratuitous cruelty. This is also observed in the child psychology studies on the moral/conventional divide that Nichols accepts – not just any norms are at issue, but in particular prohibitions of harm.Footnote ¹⁴³ That any of these norms constantly change with the shifting of opinions is hard to reconcile both with experience and with evidence.

As to some concrete rules that we discussed, it is not clear what kind of statistical evidence there is, for instance, for the prohibition of the instrumentalization of people. Do children observe a lot of this? How come human beings are able to critique and transcend norms that were robustly enforced in their environment – for example, on the permissible instrumentalization of women in patriarchal societies?

Nichols rightly underlines the difference between possible learning processes and actual learning and emphasizes that his account is only concerned with the former, not the latter.Footnote ¹⁴⁴ The reviewed literature on the ontogeny of moral cognition seems to underline the relevance of the identified problems of his account. Moral development in fact seems to take not the course that statistical learning foresees, but a rather different one, including differentiated moral concepts in young children, cross-cultural equal age trends of developments of moral cognition, intention-based moral evaluation and so forth.

An interesting point that Nichols raises is the origin of the limited hypothesis space with which child learners operate. The theory offers no explanation (as it underlines itself) of how learners arrive at the hypothesis space. This hypothesis space is thus understood as possibly innate,Footnote ¹⁴⁵ though there is some counterevidence against even such constraints, it is argued.Footnote ¹⁴⁶

The problem of constraints on hypotheses of possible morality is very important: What if children were exposed to the punishment of unintended actions? Some kind of strict liability morality? Would they then acquire a morality in which intention does not count in the moral evaluation of actions? Is there any evidence of this happening? Or is rather the opposite indicated not only by experimental work, but also by the passionate protests of children around the world when they are punished for something that they did not intend to do? The answer must be based on the poverty of stimulus argument: If the learning child does not receive sufficient input to form the hypothesis space – including the constraint that intentions count in a morality, for instance – then the constraints must be innate.Footnote ¹⁴⁷

In sum, the reference to statistical learning thus does not seem sufficient to account for the acquisition of the basic elements of morality.Footnote ¹⁴⁸ Exploring the problem of the origin of the initial conception of morality that frames the acquisition of moral knowledge, the hypothesis space and its possible nativist explanation is an important point of the theory and highlights the need for open-minded research in this area. In any case, nothing in this research rules out that the Bayesian learning mechanism would lead to the acquisition of a morality incompatible with the content of human rights – on the contrary, this theory argues that human beings can learn any rule.Footnote ¹⁴⁹ They are in no way naturally limited to narrow parochial tribalism. Statistical learning is open to acquiring inclusive norms,Footnote ¹⁵⁰ such as – one may add – human rights.