R2. Are moral norms and obligations truly externalized? Do moral norms form a unified category?

Perhaps most fundamentally, a substantial number of the commentators argue that the sort of externalized or objectified phenomenology that I repeatedly characterize as “distinctively moral” does not actually exist and therefore needs no explanation of the sort I have proposed. Stich frames his argument for this claim around several criticisms of the work of Turiel and others, rightly pointing out that the particular characteristics they use to try to classify norms and violations as moral or conventional have been scrutinized and challenged by other scholars. Patel & Machery are similarly “skeptical” that “this distinction is central to human psychology.” And Davis & Kelly argue that neither “hardness” nor “objectivity” (two properties they suggest I have conflated in characterizing externalization; see sect. R5 below), nor both together, pick out a distinctively moral domain and indeed that “[n]o subcategory of norms makes up a psychologically distinctive or cooperatively indispensible set of moral ones.” Likewise, Theriault & Young are right to be concerned about the “stimuli-as-fixed-effects” fallacy and to suggest that we are far from having convincing evidence that “morals comprise a distinct category – at least on the basis of their objectivity, universality, and authority independence, as has traditionally been argued.”

But I am afraid that these authors misunderstand the use I intend to make of the work of Turiel and others on the moral/conventional distinction. Rather than identifying necessary and sufficient features shared by all and only moral judgments or even (much more plausibly) a “nomological cluster” of such features, my goal is instead to use this line of research to help us pick out a phenomenological difference between our experience of many prototypically moral norms (like those prohibiting lying or murder) and many prototypically conventional norms (like that specifying which spoon to use for soup). That phenomenological difference itself is what concerns us most directly, though of course I have suggested that a wide range of further empirical findings are rather neatly elucidated and unified by the evolutionary explanation I offer for it. And we already know that the features to which Turiel and others appeal offer at best a rough and ready guide to that phenomenological difference, for there are domains (like moralized disgust) in which the phenomenology of norms seems to exhibit some but not all of those features. In fact, I think the clearest indicator of objectification or externalization (moral or otherwise) is one that Goodwin and Darley (Reference Goodwin and Darley2008) use as an experimental probe for this phenomenological difference: the judgment that when agents disagree at least one party to that disagreement must be mistaken. But even here Goodwin and Darley's own results suggest that there is a continuum of such externalization or objectification along which various sorts of judgments tend to cluster (pace Theriault & Young) with considerable individual and contextual variation between them, rather than a set of qualitatively discrete categories with hard edges.

What all of this suggests is not that there is no phenomenon to explain, but instead that the phenomenon demanding explanation is considerably more complex than Turiel and others initially supposed: What requires explanation is the complex pattern in which humans tend to externalize or objectify various sorts of judgments to varying degrees under various conditions. The explanation I have offered does much more to illuminate some features of that pattern (e.g., why many prototypically moral judgments are strongly externalized) than others (e.g., why moralized disgust seems to exhibit some but not all of the features that Turiel and others suggested were characteristic of moral norms in general; though see my discussion of Wiegman in sect. R6). But it is the complex pattern itself we are seeking to understand. Recognizing this complex pattern as our explanatory target (especially in conjunction with the clear context-sensitivity of externalization, elaborated below) also leaves me largely untroubled by Davis & Kelly's further suggestion that only a minority of even prototypically moral judgments are in fact objectified or externalized.

This same recognition also reveals, however, that these authors are all quite right to criticize my repeated references to a special conceptual category of externalized norms (to which particular norms can simply be added or removed) or to a distinctive phenomenology supposedly distinguishing an independently specifiable “moral domain” from all others. Although this seemed a harmless simplification of both the relevant explanatory demand and my own proposed solution to it, I now see the error of my ways: This shorthand characterization indeed suggests that the attractions of the account I offer themselves depend on the existence of such a distinctive moral category. But this suggestion is mistaken: My account of the role played by externalization in protecting prosocial and cooperative tendencies from exploitation in humans is threatened neither by the recognition that such externalization comes in degrees that vary between individuals, contexts, and types of judgment, nor by the fact that the phenomenology of externalization does not itself pick out all and only norms that are judged to be moral on some independent ground (e.g., their content). What matters to the explanation is that we selectively externalize some norms (to varying degrees) and that shifting sets of prosocial or cooperative norms can be effectively and efficiently protected from exploitation by externalizing them in this way, even if the distinctive phenomenology of externalization does not characterize all and only norms independently identified (in some other way) as moral or pick out a categorically distinct moral domain.

Stich also rightly emphasizes that Sarkissian et al. (Reference Sarkissian, Park, Tien, Wright and Knobe2011) is only the first in a “recent cascade of papers, all of which cast doubt on the conclusion that people are consistently objectivist about moral judgments.” What much of this cascade reveals is that the extent to which we externalize or objectify judgments (moral or otherwise) is considerably influenced by a wide variety of different cues that include not only the content of the judgment, but also the degree of perceived consensus regarding that judgment, its valence (i.e., prescribing good conduct or proscribing bad conduct; though cf. Goodwin), and much else besides. But I resist any suggestion that I have merely accommodated Sarkissian et al.'s results or tried to explain them away. These authors suggest that their findings show the folk to become increasingly relativist as they are forced to confront or consider moral perspectives increasingly remote from their own. I have proposed a competing hypothesis: that externalization increases as contextual cues make the need for actual social interaction with a morally deviant agent a more salient, concrete, or realistic possibility. I have also suggested that there is a natural experiment to perform to evaluate these competing hypotheses: replicating Sarkissian et al.'s (2011) study with a further condition emphasizing the realistic possibility of ongoing social interaction with those who disagree with the subject's own moral judgment (e.g., in which the morally deviant Amazonian tribesmen or extraterrestrials described in the original experiment will soon arrive in our town and we will need to decide how to interact with them). This experiment is now underway.

Moreover, other findings in this “recent cascade” actually increase my confidence in the alternative hypothesis I have proposed, in particular the experimental work of Beebe (Reference Beebe, Sarkissian and Wright2014) to which Stich directs our attention. Most importantly, Beebe shows that when the deviant agent with whom the subject disagrees in a moral judgment is described more concretely (by providing a name and brief description of their academic courses and major, or even just a name and photograph), we find a corresponding increase in the extent to which we externalize or objectify the judgment or norm with which that deviant agent disagrees. This difference is perfectly intelligible if subjects are more concerned to objectify or externalize norms when faced with a more concrete, realistic, or salient possibility of actual social interaction with a deviant agent, but I do not see how they can be easily reconciled with Sarkissian et al.'s (2011) competing hypothesis that it is instead exposure to alternative moral frameworks that moderates our objectivist or externalizing tendencies.

I find it difficult, however, to reconcile Stich's enthusiasm for this recent experimental work highlighting the variability and context-sensitivity of our externalizing tendencies with his (apparent) further claim that there simply is no such externalized phenomenology in the first place. He says,

Here Stich seems to report that he himself doesn't experience even the most prototypical moral judgments as having the character I describe in these phenomenological terms. It might seem tempting to suppose that Stich instead means only to deny that such externalization is a consistent part of his moral phenomenology (i.e., is not reliably exhibited by all and only judgments he regards [on some independent ground] as moral in character). But this cannot be right: For one thing, Stich emphatically denies that there is any independently unified category of moral judgments whose degrees of externalization we might then go on to consider. It might therefore seem natural to respond simply by noting that the target article itself already recognizes the existence of reliable variation among human beings in whether they externalize or objectify even prototypically moralized judgments, and by suggesting that perhaps we should not be shocked to discover further intersubjective variation in this respect (I should emphasize that I do not mean to be suggesting that Professor Stich is a psychopath or sociopath). Perhaps Stich does lack the sort of externalized moral phenomenology I have described, but my remaining commentators seem to recognize such selective externalization of some but not all norms as part of their own moral phenomenology, even if they also think I have gone on to mischaracterize or misunderstand that phenomenology in a truly daunting number of further ways.

R3. Is externalization necessary for correlated interaction and/or human prosociality?

Patel & Machery suggest that even when externalization does occur it is probably not what protects prosociality from exploitation, because correlated interaction can be achieved by other means, such as merely conventional norms (see also Birch; Brusse & Sterelny; Handfield, Thrasher, & García [Handfield et al.]; Jebari & Huebner; O'Neill). Patel & Machery go on to note that what they suggest is “the best explanation of cooperative behavior” we have (Balliet et al. Reference Balliet, Mulder and Van Lange2011) makes no appeal to externalization, using instead punishment and reputation-tracking as mechanisms for achieving correlated interaction. It is certainly true that externalization is not necessary to produce correlated interaction, but this is also not all that externalization does. For one thing, externalization selects out only a subset of norms and judgments (rather than all or none of those we accept) as those by which the comparative desirability of social partners will be judged. In addition, externalization ensures that one and the same set of privileged norms is used simultaneously to motivate ourselves and to evaluate candidate social partners even as the membership of this set of privileged norms changes over time. It might even be possible to achieve correlated interaction sufficiently precise and robust to protect humans’ spontaneous prosociality from exploitation using only subjective preferences and conventional norms, but a wide range of evidence, including perhaps most importantly the systematic sensitivity of preferred social distance to specifically moral (but not other kinds of) disagreement (Skitka et al. Reference Skitka, Bauman and Sargis2005), strongly suggests that externalization is in fact the means by which such sufficiently precisely correlated interaction was achieved among early humans.

Birch and O'Neill similarly emphasize that objectification of the sort I have described is not necessary to achieve correlated interaction between cooperators. Birch, for example, suggests that once we distinguish the apparent source of a norm from its scope, it becomes clear that the latter rather than the former is what really matters for protecting cooperative dispositions from exploitation. He rightly points out that social norms need not be externalized to be of wide scope (e.g., a subjective commitment to removing litter that “applies to the behaviour of the whole community”) and need not be of wide scope to be externalized (e.g., externalized or objectified norms that apply to and concern only the conduct of the Pope). Although this latter example is an extreme case of the sort of “role-dependent or asymmetric norms” that I have suggested arose later in our phylogenetic history, it nevertheless illustrates that even norms of narrow scope can be externalized. The first illustration seems more problematic: Externalization leads us to apply a norm with unrestricted scope, to absolutely any potential social partner, so applying “to the behaviour of the whole community” does not exhibit the sort of unrestricted scope that externalization or objectification confers. It seems instead a textbook example of a conventional norm, applied only to the members of the relevant cultural group or community.

But let us concede the possibility of both externalized norms of narrow scope and non-externalized norms applied with at least very wide scope (e.g., to all members of a particularly expansive community). Does this show that it is wide scope rather than externalization that does the work of ensuring correlated interaction? Surely not, for my proposal is that the selective externalization of particular norms is precisely how such norms came to be applied by early humans with unlimited scope in the first place. Wide scope and externalization are not competing explanations for the stability of exploitable norms; instead it is by means of externalization that humans actually came to apply particular sets of norms with unrestricted scope. Birch is right to think that any alternative mechanism leading us to apply a subset of our norms with arbitrarily wide scope (and to exclude violators from interaction with us) could generate correlated interaction equally effectively, but the suggestion here is that externalizing norms was in fact the way that hominins came to apply them with unrestricted scope rather than only to the members of a particular (even very large) social group.

Similarly, O'Neill argues that generalization and desire/preference-independence are distinct features that I have conflated in characterizing the externalization or objectification of moral norms. She is quite right to emphasize that these features are different and play distinct roles in protecting prosociality from exploitation. She goes on to suggest, however, that generalization alone generates correlated interaction while desire/preference-independence serves merely to “stabilize” the resulting prosocial interaction by protecting it from particular kinds of exploitation and/or failure. Here I am less convinced, but I am also unsure of how much is at stake in any residual disagreement, as we seem to agree that both of these characteristic features of moral normativity were important for the evolution of human prosociality.

Moreover, I suspect that O'Neill's analysis here is flawed in an illuminating way. When she argues that generalization alone suffices to produce correlated interaction, she suggests that: “Presumably, the tendency to generalize obligations produces or comes with a preference for cooperation partners who believe themselves to be subject to the same obligations that one believes oneself to be subject to.” (Birch does not make this claim explicitly but may presuppose it; see also Wiegman.) But this is again simply to hide a crucial part of the problem for which externalization itself constitutes such an elegant solution – namely, the challenge of ensuring that at any given time I use one and the same set of some, but not all, of the norms I accept both to motivate my own behavior and to evaluate the desirability of candidate social partners, even as the members of that set remain open to modification, extension, and replacement on a cultural or historical (rather than biological) timescale. Nothing about evaluating candidate social partners by means of their adherence to a particular set of norms requires these to be the very same norms by which I myself am motivated – it is externalizing the source of moral normativity that establishes and maintains this crucial connection. My suggestion is that externalization is how humans actually managed to combine normative plasticity with the tendency to experience (at any given time) one and the same set of privileged norms both as motivating our own behavior in a distinctive, preference-independent way and as the standards by which we evaluate the desirability of candidate social partners.

I hope it is now clear why I decline to defend the “indispensability thesis” that Handfield et al. attribute to me, which they describe as the view “that externalized moral norms of this sort are necessary to achieve pro-social cooperation, at least at the high rate seen in humans.” Once again, moral externalization is certainly not necessary for human hypercooperation, because any source of (sufficiently) correlated interaction will do. I fully agree with Handfield et al. that the mechanism of correlated interaction I describe cannot be usefully assimilated to costly signaling, a green beard hypothesis, or social selection – indeed, I think it is not usefully assimilated to any mechanism for generating correlated interaction that we have found elsewhere in nature. Moreover, although correlated interaction can indeed be generated by Patel & Machery–style punishment and reputation-tracking (though costly punishment is itself a second-order form of altruism whose stability would require some further explanation), Birch-style application of norms with wide scope, O'Neill-style generalization, and in many other ways besides, my claim is that we have convincing evidence that externalization is the mechanism actually responsible for the distinctive forms of generalization, reputation-tracking, preference-independence, and so forth, which constitute salient features of our own moral psychology.

R4. Is externalization a cultural construct? How does externalization generate social networks of precisely correlated prosocial interaction?

Patel & Machery offer a further reason for worrying about the centrality or importance of the moral/conventional distinction for human psychology: They claim that the distinction itself is culturally parochial, found only in some human cultural traditions and not others, and they cite unpublished data suggesting that Indian and Muslim subjects with a variety of national origins simply do not recognize any distinction between moral and non-moral norms. I cannot evaluate the unpublished data that Patel & Machery mention, but I remain skeptical. The case seems at least superficially similar to many others in which we have concluded that concepts or distinctions found in a wide range of human cultures are simply absent from some particular culture or population, only to find that in fact we failed to probe for those concepts or distinctions in a way that was sufficiently sensitive to the subtleties of the particular culture in question. This sort of danger is highlighted by Poulin's suggestion (following Gray) that those who condemn apparently harmless moral violations often do so because they see those violations as in fact having victims who are harmed by them.

Nonetheless, suppose that Patel & Machery are right about the cultural parochiality of the moral/conventional distinction or even externalization itself. What follows? Simply that the selective advantages of externalization will have to be recast using only the machinery of cultural evolution, rather than the complex combination of cultural and biological co-evolution that I have proposed. In that case, we would still appeal to the role of externalization in generating correlated interaction and (thereby) protecting prosociality from exploitation and exclusion. But now this would be part of a larger explanation of how a particular cultural innovation (viz., the selective externalization of norms) came to play an important role in scaffolding human prosociality in whatever cultures do externalize some, but not all, norms.

This would also be the appropriate response if we were convinced by the very different reasons offered by Brusse & Sterelny for thinking that the externalization of some but not all of the norms we embrace “is a late-breaking cultural innovation.” But I do not think we should be convinced. The problems raised by Brusse & Sterelny for the idea that externalization evolved in humans to establish and maintain correlated interaction between prosocial or cooperative agents all depend on a subtle but important misunderstanding of how externalization itself works. Brusse & Sterelny are quite right to think (see also Voorhees, Read, & Gabora [Voorhees et al.]) that excluding or shunning any potential partner with whom we have identified any point of substantive moral disagreement would be far too costly to be adaptive in realistic ancestral environments. But we did not and do not simply shun those with whom we have moral disagreements, whether concerning specific moralized norms or (far more frequently) simply our moralized evaluations of particular cases. Instead, identified points of moral disagreement simply count against the desirability of a given potential social partner relative to others and increase the social distance we prefer to maintain from that potential partner (see Skitka et al. Reference Skitka, Bauman and Sargis2005). The extent or degree to which any particular disagreement influences our evaluation of a potential social partner is mediated by factors like the extent to which the norm in question is externalized (Goodwin & Darley Reference Goodwin and Darley2012), how strongly the norm itself is held (Skitka et al. Reference Skitka, Bauman and Sargis2005), and presumably much else besides. But what such disagreement generates is not a list of tribemates to shun because I have identified some point of moral disagreement with them, but instead simply a preference ordering (or perhaps a set of preference orderings relativized to different activities) over such potential partners that is a complex function of our identified moral agreements and disagreements with them (as well as non-moral considerations like their competence or physical prowess).

Mistakes and accidents are also, therefore, not as consequential as Brusse & Sterelny imagine, though it is nonetheless remarkable that human cultures so reliably include elaborate procedures of apology and repair by which members seek to advertise their commitment to precisely those moral judgments and convictions that recent mistaken or accidental conduct might lead others to suspect they do not share (see also Allidina & Cunningham on the moralization of everyday behavior and Wiegman on over-compliance with norms). Externalizing a norm leads me to devalue, rather than shun, interaction with those who violate or fail to externalize it, ensuring in turn that identifying points of moral disagreement with my tribemates neither precludes me from cooperative or other forms of prosocial interaction with them, nor leaves me (as Brusse & Sterelny suppose) with very few or no candidate partners with whom I am both willing and able to cooperate. Of course, there may well be some threshold amount or variety of moral disagreement beyond which I will indeed shun a fellow group member, but this is not a consequence of identified moral disagreement in general. This is also why the demands of adaptive plasticity and conformism do not conflict: Two agents can agree on how to extend or adapt an externalized norm in new ways or into new circumstances (or even to externalize an entirely new norm), thus protecting their further prosocial interaction from exploitation, without being shunned for doing so by other group members.

The challenge posed by Johnson suffers from a related misunderstanding of the role played by moralized agreement and disagreement in mediating our prosocial dispositions. He argues that my proposal implies that I would never act prosocially or cooperatively towards another agent unless and until I had abundant and detailed positive evidence of many specific points or respects of moral agreement (or, even more challenging, extensive evidence that the agent and I “share the same heritable attribute triggering cooperation”). He therefore suggests that anonymous prosociality and prosociality in environments where we have little or no information about others represent “a form of prosocial behavior than cannot be explained by moral externalization.” But nothing about externalization requires positive evidence of moral agreement (much less of heritable properties in common) to be a condition of prosociality. Humans are much more spontaneously prosocial than other primates, and what is needed to protect that spontaneous prosociality from exploitation is simply that an agent's enthusiasm for social interaction with a given partner should decline when evidence of moral disagreement with that partner arises. Nearly any particular degree or respect of spontaneous prosociality with in-group members (or even strangers) could serve as a (culturally and contextually variable) default starting point for humans, so long as acquiring evidence of moral disagreement with any particular agent reduces the degree (and/or forms) of exploitable prosociality we spontaneously extend to her. What matters is that we hold others responsible for living up to our own externalized norms and increase our preferred social distance from them if they fail to do so, and this does not require that we must have any evidence (much less abundant evidence) that they will in fact do so before we are willing to risk interacting cooperatively or prosocially with them. Indeed, it was precisely to protect and facilitate dramatic increases in the range and degree of our spontaneous prosociality that our moral psychology evolved mechanisms ensuring increasingly precise, specific, and powerful correlated interaction in the first place.

For related reasons, I must decline Ross's intriguing invitation to see the account offered in the target article as advancing a comprehensive explanation for the dynamics of tribe formation, according to which:

I have several reservations about seeing this as a general model of the dynamics of tribe formation. Perhaps most importantly, it obscures the fact that externalization is so powerful precisely because it can generate correlated interaction between agents prepared to moralize a given norm (or a particular extension or application of a norm) even while they remain accepted members of a larger social group in which that particular extension, application, or norm is not generally moralized (cf. Brusse & Sterelny, and see Böhm, Thielmann, & Hilbig [Böhm et al.] on moral homogeneity within groups). That is, moral externalization ensures correlated interaction between agents prepared to engage in (and demand) a particular form of prosocial interaction without the need to form a new (physically or spatially distinct) tribal group in which this behavior is homogeneously moralized. It does seem perfectly plausible to suggest that such moralized differences might emerge, accumulate, and grow into points of persistent substantive moral conflict or disagreement with others, ultimately leading to the formation of a new tribe in something like the process Ross suggests. But even here I suspect this would most often be a matter of members of two groups within a tribe finding themselves with a sufficient number of sufficiently important moral disagreements that members of each group are motivated to exclude members of the other from social interaction with them, with neither group consisting of “excluded villains [who] interact with one another for lack of an alternative” (see Baumard et al. [Reference Baumard, André and Sperber2013] on self-segregation in contemporary hunter-gatherer societies, and Böhm et al. on out-group hatred). This last difference matters, because it suggests that members of the two groups begin to moralize different normative demands long before either group's members are simply shunned or excluded from social interaction generally by members of the other, rather than one tribe being composed of “villains” or “cast-offs” who simply moralize fewer behaviors (and therefore might ultimately threaten to collapse into what Ross, following Edgerton [Reference Edgerton1992], calls “sick societies”). This proposal would, of course, still help explain why “human tribes manifestly bifurcate within shared physical environments” (Ross, para. 4), although it also seems likely that moralized disagreement and/or moralized group identity is just one of many different factors at work in the dynamics of tribe formation.

R5. Do we also externalize conventional norms, aesthetic judgments, and/or other kinds of value judgments? Might externalization have been an ancestral condition?

I hope it is already clear why the very thin sort of objectification that Patel & Machery rightly point out is attributable to conventional as well as moral norms was never our explanatory concern. Likewise, Van Prooijen is right to suggest that prototypically moral norms are “externalized” in the sense that they are imposed on us by others, but this sort of externalization is equally attributable to prototypically conventional norms and is therefore again simply beside the point. And the same is true for the sense of objectification (“hardness”) that Davis & Kelly argue characterizes any and all normative judgments whatsoever. From the beginning, our explanatory target has been the fact that many prototypically moral norms exhibit a further and much stronger form of phenomenological externalization or objectification, one revealed most clearly in the unwillingness of subjects to tolerate disagreement without error concerning those norms and/or their implications (the “objectivism” Davis & Kelly suggest I have conflated with “hardness”), and it is this further, more robust form of externalization or objectification for which I have sought to provide a convincing evolutionary explanation.

Similarly, perhaps Isern-Mas & Gomila are right to think that it has been undisputed at least since Kant that we ascribe some distinctive form of objectivity (“aim[ing] at universal validity”) not only to moral judgments, but also aesthetic judgments and indeed any value judgments whatsoever. However, this cannot be the same form of objectivity for which Goodwin and Darley (Reference Goodwin and Darley2008) probe by asking subjects about the possibility of disagreement without error. Subjects ascribe that form of objectivity in the highest degree to judgments of empirical fact, more modestly to judgments of moral norm violation, followed by judgments of conventional norm violation, and least of all by judgments of taste and preference, including putatively aesthetic judgments like “Frank Sinatra is a better singer than Michael Bolton” (Goodwin and Darley's example). It is unsurprising, then, that the process of objectification Isern-Mas & Gomila (following Darwall Reference Darwall2006) go on to describe is one that applies to conventional norms in just the same way that it does to moral norms and therefore cannot explain the emergence of the sort of objectivity that subjects ascribe to prototypically moral (but not prototypically conventional) norms. Fortunately, I do not think the problem they seek to solve by invoking this process (supplying a supposedly missing connection between our externalized norms/values and our motivations) actually exists: Moral judgments can be intrinsically motivating, I suggest, just like desires and preferences, despite the fact that externalizing the former but not the latter certainly generates other salient phenomenological differences between our experiences of intrinsic motivation in the two cases. Indeed, this and other characteristics shared between intrinsically motivating moral judgments and intrinsically motivating preferences or desires leaves me unsurprised by Theriault & Young's evidence that the neural and behavioral signatures of moral judgments are more like those of preferences than those of objective facts (see also Schulz).

Schulz has a distinct worry about the motivational adequacy of externalized moral norms: Although externalizing such a norm can certainly motivate us to follow that norm itself, he suggests, he does not see how or why it would motivate us to disfavor social interaction with those who do not comply with (or do not externalize) the norm. Here I think he is misled by the idea that norms cannot motivate us to do anything that is not explicitly specified as part of the content of that norm itself (hence the need for a further, conjunctive component of that content). But this does not seem to be how our motivational psychology actually works. Notice, for example, that the content of a conventional norm does not include any explicit description of the consequences of noncompliance – it simply articulates the norm itself, and it is simply a matter of empirical fact about us that we respond to any particular violation of such norms with criticism, exclusion, forgiveness, shock, glee, disappointment, or in any other particular way. On the account I offer, the same is true of moral norms and obligations: The norm itself does not specify the consequences of noncompliance; instead it is simply an empirical fact about human beings that they devalue social interaction with those who do not comply with (or externalize) the norms that they themselves externalize. Indeed, it is because the consequences of noncompliance are not specified as a conjunctive part of the content of the norm itself (as, surprisingly, Schulz himself seems to recognize in his second paragraph) that we must discover that increasing preferred social distance is in fact the consequence of failing to comply with (and/or failing to externalize) a norm that we ourselves externalize (Skitka et al. Reference Skitka, Bauman and Sargis2005). Moreover, Schulz's description of how his conjunctivist/subjectivist proposal explains the apparent externalization of moral demands is unconvincing, because the fact that moral demands systematically motivate not only compliance, but also the consequences of others’ noncompliance (whatever they may be) also applies straightforwardly to the case of merely conventional norms – again, whatever kind of apparent externalization Schulz can explain in this way is simply not what we set out to understand.

We can now also usefully approach Kaznatcheev & Shultz's suggestion (following Joyce Reference Joyce2006) that the objectification or externalization of norms was a likely ancestral condition from which no shift to an externalized moral phenomenology would have been required. Sometimes it seems that these authors slip into answering the wrong question, as when they seek to establish that “an understanding of experience as subjective both in oneself and others develops from an objectivized phenomenological precursor.” The issue is not, of course, whether the general capacity to represent experiences as reflecting objective states of the world itself did or did not precede the capacity to represent experiences as reflections of my own subjective states, but rather whether the capacity to experience norms as reflecting objective states of the world itself preceded the capacity to represent them as merely reflecting our own and others’ subjective preferences and/or social conventions. The more general capacity to discriminate representations of states of the world from my own subjective states is far more fundamental, and indeed it seems we could hardly hope to formulate or even understand norms regarding our own or others’ conduct without having this more fundamental capacity already in place. Moreover, the considerations to which Kaznatcheev and Shultz appeal that do bear on the externalization of norms in particular seem intended simply to blunt the suggestion that an objectified or externalized normative phenomenology must have emerged from an ancestral subjectivist normative phenomenology rather than vice versa. I agree with this modal judgment, but given the exploitability of human cooperative and other prosocial dispositions, the fact that insensitivity to intersubjective variation in prosocial norms would indeed invite and generate such exploitation, and the fact that we strongly externalize only some of the norms we embrace (including disproportionately many that do in fact serve to protect prosociality from exploitation), I remain skeptical that an externalized or objectified normative phenomenology represents the ancestral condition of human beings.

R6. Are there lacunae in our understanding of how and why externalization occurs and/or any convincing alternative explanations for it?

Poulin suggests that the wide array of apparently heterogeneous externalized moral judgments we find in different human cultures may all simply be consequences of externalizing a more generalized “moral imperative to help or at least not harm others,” often in culture-specific ways that may appear harmless to those outside the culture in question (cf. Patel & Machery). This is certainly a possibility, though I am more impressed than Poulin with evidence that humans regularly externalize and/or moralize norms that they themselves see as unrelated to harm. Poulin's suggestion is, of course, an extreme version of one of the potential explanations proposed in the target article itself for the patterns of similarity in externalized norms we find across human cultures: strongly biased learning in the externalization of particular norms or types of norms (whether categorized by their content or in some other way). But I continue to think that the other (non-exclusive) explanations suggested in the target article (convergent cultural evolution, common descent with modification, and enhancing in-group identification) and perhaps others besides are also quite likely to be part of the story of how different human cultures came to externalize the remarkably diverse and heterogeneous, but nonetheless systematically related, collections of different moral norms that they do. But if (or to the extent that) Poulin is right, the adaptive advantage of the remarkable plasticity with which norms are selectively externalized by members of different cultures at different times is simply that it allows us to quickly update, extend, and modify our (culturally specific) judgments about whether and/or how others can be helped or harmed.

For similar reasons, I doubt that the account proposed by Voorhees et al. represents anything like a complete answer to their question, “how could a group come to externalize all of the same norms?” Again, I expect the other mechanisms noted previously to play an important part in this story, though I am happy to regard the intricate machinery these authors propose of “cultural idea systems,” “functional vehicles for the social expression of emotional responses,” and “complexes of beliefs and/or organizational rules that operate in a top-down manner so that individuals gain functionality only by adherence to these rules and/or constraints” as an attempt to provide a more detailed and systematic description of many of the processes of vertical and horizontal cultural transmission (and reinforcement) through which the members of any particular cultural group come to share many of the same particular externalized norms and judgments. In that case, however, these authors provide a sympathetic extension or refinement of the account I have offered rather than (as they seem to think) a competitor to it.

Sometimes, however, Voorhees et al. seem to suggest that it is the existence (or emergence in ontogeny) of externalization itself that they seek to explain by invoking this machinery of cultural idea systems and the like, and here I am unconvinced, in large part for reasons like the ubiquity (if not universality, cf. Patel & Machery) of the moral/conventional distinction across human cultures and the fact that human children in very different cultures seem to reliably and spontaneously start externalizing norms at the same point in ontogeny. Voorhees et al. are also wrong to think that my own proposal faces daunting challenges concerning the adaptive value of externalization for a single initiating agent and/or the second-order free-rider problem for punishment. With respect to the first, if I am motivated by a given exploitable prosocial norm, it is beneficial for me to externalize that norm (treating adherence to it as a relevant consideration in my own selection of candidate social partners) to avoid exploitation whether or not others embrace and/or externalize either that same norm or any norms at all. And the target article's account of “punishment” faces no serious second-order free-rider problem because the only cost incurred by a punishing agent is the loss of further opportunities for social interaction with a potential partner she already sees as prone to exploitation.

Most importantly, however, Voorhees et al. simply do not offer a plausible account of how the distinctive phenomenology of externalization itself arises or emerges in the course of ontogeny. They claim, for example, that “[i]f our ancestors’ moral norms are part of the cultural idea system acting in a top-down manner within social systems organized through kinship relations, then kinship itself provides the objectivity and coherence of norm exteriorization.” Here they seem to suggest that the distinctively externalized phenomenology of many prototypically moral judgments derives from the fact that they are embedded in kinship structures, but this is very implausible. For one thing, the fact that a norm is acquired from a member of one's kin gives us no reason to expect that the norm itself will be experienced as objective, even if the set of kinship relations structuring this acquisition is nonetheless itself viewed or experienced as an objective fact about the world. But in any case, moral normativity and/or externalization do not seem sensitive to whether or not norms are acquired from (or concerned with) kin. If they were, why would we not also externalize the merely conventional norms that we acquire from the same kin at the same time? A similar problem applies, of course, to Voorhees et al.'s more general suggestion that “cultural ideas, acquired through enculturation, are internalized by culture bearers and seen by them as objective reality.” If this were indeed the source of the distinctive phenomenology of externalization, then again it seems we should expect to find conventional norms externalized in just the same way.

Jebari & Huebner propose yet another supposedly alternative explanation of moral externalization that I think does not conflict or compete with that offered in the target article. These authors suggest that “a plausible understanding of the evolution of objective morality must look beyond human psychology, to the objective features of the world that govern cooperative human ways of life (henceforth ‘lifeways’)” – that is, to the sorts of objective facts about the world and ourselves that render humans obligate cooperators, such as our affiliative tendencies, conformism, greater social tolerance and docility, and the loss of adaptations like sharp teeth for hunting and defense and the ability to extract nutrients from uncooked food. Jebari & Huebner go on to say:

Indeed, they suggest that “the felt objectivity of … obligations emerge as a consequence of our relationship to the social order, and our moral motivations are determined by our (often tacit) recognition of this relationship.”

There is very little in this proposal with which I am inclined to disagree, besides the explicit claim that it obviates the need for any appeal to externalization itself. A full explanation of the evolution of human prosociality will indeed appeal to objective facts about the world including the sorts of evolutionary changes that Jebari & Huebner rightly suggest have helped render humans obligate cooperators. But these sorts of facts about humans are simply not in competition with “feelings of objectivity” as a proximate mechanism for motivating and protecting human prosociality. Far from “obscur[ing] the significance of … the emergence of complex and adaptive social networks … structured by rich patterns of social interaction” (as Jebari & Huebner allege), I appeal to externalization itself to explain how those very social networks become established, maintained, and modified over time, and to explain the emergence and persistence of many of the very adaptations facilitating human prosociality that Jebari & Huebner describe. The only alternative mechanism to which Jebari & Huebner appeal is “the role of affiliative tendencies in producing automatically coupled values and preferences” – they suggest that “such forces are sufficient to drive complex, open-ended, and cooperative forms of behavior” and therefore that objectification is unnecessary (cf. sect. R3) to explain our preference for partners who resist “contra-normative” behavior. This appeal, of course, faces precisely the same problem as Van Prooijen's much simpler claim that moral norms are experienced as objective because they are imposed on us by the expectations and demands of others: It offers no explanation whatsoever for the fact that we externalize prototypically moral norms but not prototypically conventional norms. Conventional norms and norm-violations (not to mention subjective preferences) also mediate our affiliative tendencies and are also part of the social reality of obligate cooperation that Jebari & Huebner describe, so their proposal offers no explanation of the fact that we externalize some, but not all, of the norms and obligations we embrace at any given time.

Wiegman seeks to extend the account offered in the target article with the intriguing suggestion that “disgust provides a simple and economical way of implementing externalization of norms in some moral domains and hence may have been an early and influential driver of externalization.” As he explains: “if one comes to feel that certain acts are disgusting (e.g., acts that violate certain kinds of norms), one will avoid committing such acts oneself and one will also avoid those who commit such acts (because they are contaminated thereby)” and, most importantly of all, “one may also think that others have reason to avoid such acts (since they would be contaminated thereby).” Consider, first, the suggestion that those who commit disgusting acts thereby become contaminated, and that we avoid them in order to avoid being contaminated ourselves. Although this does provide a means by which we might acquire simultaneous motivations to avoid both particular acts and potential social partners who commit them, notice that the proposed motivation itself consists in nothing more than a strong subjective preference for avoiding contamination wherever we find it. Externalizing such motivations and/or the norms they motivate is instead a matter of coming to regard myself and others as responsible for or obligated to refrain from the disgusting acts in question, rather than simply seeking to avoid potentially contaminating social partners as well as contaminating actions. Thus, if disgust is to provide “a simple and economical way of implementing externalization,” this will have to be a consequence of the fact that when an agent finds an act disgusting she “may also think that others have reason to avoid such acts (since they would be contaminated thereby).” But again, thinking that others have good practical reasons for refraining from disgusting acts provides no ground at all for thinking that they are in any way responsible for or obligated to refrain from such acts. Externalization requires that we devalue social interaction with others because they have violated such responsibilities or obligations, not simply because they have failed to act on the good practical reasons they themselves have for avoiding particular actions, may therefore be or become contaminated themselves, and therefore threaten to contaminate us if we interact with them. Wiegman's claim that “disgust naturally lends itself to the thought that others should not contaminate themselves via contact with what I deem disgusting” (para. 2, my emphasis) trades on an ambiguity between such prudential and moralized senses of “should.” It remains possible that paired aversions to both particular acts and those who commit such acts (simultaneously generated by my more general desire to avoid contamination) played some important role in the emergence of genuinely externalized norms and obligations, but disgust itself does not seem to provide the “immediate route to externalization” that Wiegman suggests.

This complexity also helps to highlight further reasons we might be well-advised to regard moralized disgust as an unusual or perhaps even unique form of externalization in any case. Following Kelly (Reference Kelly2011), Wiegman suggests that moralized disgust is ultimately generated by the co-optation of a pre-existing system evolved to avoid poisons, pathogens, and parasites by a further and distinct system responsible for “motivat[ing] norm compliance and enforcement.” If so, it seems more likely that moralized disgust emerged against the backdrop of an existing normative psychology, rather than providing the bridge by which such a normative psychology might have emerged in the first place. Note, however, that this complicated and unusual (perhaps even unique) phylogenetic history might well make it less surprising that moralized disgust often exhibits only a subset of the features suggested by Turiel and others to be characteristic of moral norms in general (sect. R2).

Zinser also seeks to propose an alternative explanation for moral externalization, suggesting that it is “merely a psychological by-product of underlying affective responses; it is a story we tell ourselves to make sense of a particular class of pre-existing subjective states” in which “our executive, conscious system has to make sense of what our automatic, non-conscious system has decided.” As he notes, such “conscious explaining of our innate intuitions is called confabulation.” This proposal, he suggests, offers not only a more parsimonious explanation of moral externalization itself, but also one that better coheres with both dual-process models of the mind and the sort of social intuitionism ably defended in recent years by thinkers like Jonathan Haidt.

There is little question that humans engage in a great deal of this sort of post-hoc confabulation, especially to defend or rationalize strongly held moral intuitions, and I have considerable sympathy for both dual-process models of the sort Zinser references and Haidt's social intuitionism. But Zinser's proposal nonetheless offers an unconvincing explanation of human moral externalization. For one thing, it ignores the fact that we feel (rather than merely believe) moral norms and obligations to be somehow imposed upon us externally. While there is little question that holding particular beliefs can influence both the phenomenological character and the content our experiences, I am aware of no other case in which our conscious minds are proposed to have engineered a phenomenologically novel form of experience (viz., that of feeling morally obligated) and introduced it back into the stream of our conscious experiences in order to rationalize or scaffold the confabulation we have adopted in response to that very experience. A further class of reasons includes facts like the ubiquity (if not universality, cf. Patel & Machery) of the moral/conventional distinction across human cultures and the invariance of the age at which children in different cultures begin to spontaneously externalize some, but not all, of the norms they accept. If, as Zinser goes on to suggest (see next paragraph), moral externalization were simply a conscious, post-hoc rationalization of the fact that some of the norms we embrace are other-regarding, we should expect to find considerable cultural variation in both the particular confabulations we arrive at and the point in ontogeny when they first appear.

But most importantly, there is simply no genuine phenomenon demanding explanation that Zinser's proposal actually serves to explain. He suggests that the moral intuitions that invite or provoke us to externalize do so because they are other-regarding:

But the fact that a preference is other-regarding in this way does not mean it requires a special form of motivational rationalization: such other-regarding preferences are still just my subjective preferences about those others. It is not as if the fact that others are external to me somehow implies or even suggests that my own motivations to behave prosocially or altruistically towards them must also themselves have an external source. I have lots of other-regarding preferences that I do not externalize: I might, for example, want very much for my children (or yours) to have fulfilling and happy lives without thinking that either they or anyone is morally obligated to live such lives – these are simply the subjective preferences I hold regarding others. So there is no special puzzle about the source of motivation for our other-regarding preferences, and no need for us to confabulate moral externalization to make sense of them. Zinser's suggestion cannot actually explain why some (but not all) other-regarding attitudes are externalized, and of course, it also makes no sense of the fact that we do sometimes seem to externalize attitudes or demands that are purely self-regarding, in cases like moral duties to the self (of the sort familiar from the work of Immanuel Kant) and moral obligations of the sort that Birch suggests might both apply to and concern only the Pope and his own conduct. No special motivational rationalization is required for our other-regarding norms, preferences, and obligations, and even if there were such a demand, Zinser is wrong to think that confabulating externalization as such a rationalization would help satisfy it.

R7. Extensions, elucidations, and friendly amendments

A number of commentators seek to supplement, extend, or refine the account given in the target article in a variety of further ways that also repay careful consideration. Bruner, for example, argues that the set of norms and beliefs we should regard as likely to be or become externalized is wider than I have suggested (see also Wiegman) and includes norms governing “conflictual coordination” problems in which agents must coordinate to achieve a desired end but disagree about which of several different possible coordinative arrangements is most desirable. With such norms, free-riding and exploitation are not the outcomes that must be avoided, but instead miscoordination as well as any obstacles to resolving disagreements quickly, easily, and peacefully, which provides agents an “incentive to selectively interact with those adhering to the same norm as themselves” (para. 5 in Bruner's commentary). Bruner acknowledges the target article's claim that it may well be possible to moralize nearly any norm or behavior, but he suggests that norms governing conflictual coordination problems are particularly likely to be moralized for just the same reasons I suggest that norms protecting prosociality from exploitation are – namely, externalizing these particular norms is what generates the positive fitness consequences of externalization itself. I think Bruner is entirely correct and in fact points the way towards an even broader moral: It is not just norms enhancing cooperation and/or preventing exploitation of prosociality we should expect to be widespread among human societies, but instead any and all norms whose externalization would have significant positive consequences for our fitness. This will include norms governing conflictual coordination, just as Bruner suggests, but it will likely include a wide array of other norms and/or types of norms, too. In the target article, I focused on the fact that we can explain the ubiquity of norms enhancing cooperation or protecting prosociality from exploitation in this way, but Bruner's larger point is that the same explanation will apply to any norm whose externalization reliably ensures substantial positive consequences for our fitness. I should emphasize, however, that this recognition leaves entirely open the question of what process or combination of processes (descent with modification, convergent cultural evolution, biased learning, etc.) are those by which any particular norm or type of norm became and/or remains widespread across human cultures, just as it did in the more specific case of norms enhancing cooperation and protecting prosociality from exploitation.

Böhm et al. seek to extend the account offered in the target article in a different way. These authors suggest that the target article emphasizes the benefits of moralizing for facilitating cooperation and prosociality (especially within in-groups) to the exclusion of recognizing that these same moralizing tendencies play a central role in “fuel[ing] aggression and conflict between groups.” My only reservation about their description of the target article's central morals is that the function of moral externalization is not simply (nor even primarily) to enhance and protect cooperation among all and only the members of a particular in-group; it does serve that end, but it does so by forming much more fine-grained networks of correlated interaction even within a well-defined in-group between agents inclined and/or willing to externalize particular norms and judgments. In fact, this is part of what makes externalization under normative plasticity so powerful – existing norms can be extended and applied to new situations and new norms moralized for the first time by just one or a few members of a community (generating correlated interaction between them) even when those agents constitute only a tiny fraction of the larger social group to which they belong (cf. Brusse & Sterelny; Jebari & Huebner; Ross).

Setting aside this quibble, I certainly concur with Böhm et al.'s suggestion that: “While morality may indeed foster cooperation and harmony within groups, it may also fuel aggression and conflict between groups.” Moreover, I suggest that the further experimental evidence which Böhm et al. go on to report (Weisel & Böhm Reference Weisel and Böhm2015) elegantly coheres with the account offered in the target article itself. These experiments measure out-group hatred as the “willingness to actively diminish out-group members’ resources at personal cost in an intergroup social dilemma game.” As Böhm et al. describe their results:

This fascinating finding seems to offer further evidence of the distinctive role that specifically moral (but not other kinds of) disagreement plays in regulating human prosociality. Notice that mere enmity was insufficient to motivate agents to pay to punish members of another group – actual moral disagreement (for which membership in groups with moralized identities serves as a proxy) was required. So it is not just that externalization and moral advertisement often interact with in-group/out-group dynamics to leave us without sufficient information about members of out-groups to be willing to risk spontaneous prosociality with them: Weisel and Böhm show in addition that when group identities are themselves moralized, our resulting awareness of our moral disagreements with members of another group can motivate even costly efforts to actively harm the members of that group. This result nicely complements Skitka et al.'s (2005) finding that goodwill and cooperativeness among the members of a group trying to solve a problem are not lowered by the existence of moral disagreement between them but instead by their awareness of that disagreement.

Goodwin offers two similarly sympathetic elaborations of the account provided in the target article. The first are a pair of proposed mechanisms by which moral objectification might give rise to correlated interaction. One such possibility is that we may direct increased social attention towards those who are discovered to objectify a given norm that we ourselves also objectify, which seems plausible and is simply overlooked in the target article. I am somewhat puzzled by Goodwin's description of the other mechanism he proposes, which seems to assume that objectification is itself a public and intersubjectively available process, such that “[t]he objectification of a moral norm creates a strong expectation that others abide by this norm, but it also conveys this social expectation” (para. 3 in Goodwin commentary, original emphasis). What I have called externalization or objectification is a matter of private experience, in which I become motivated by a particular norm or obligation in a way that I experience as externally imposed on me and (therefore) upon others as well. It is instead the advertisement of one's externalized commitment to a given norm or obligation that represents a public and intersubjectively available act conveying “the strong expectation that others abide by this norm.” It therefore seems to me that the additional mechanism Goodwin proposes just is the mechanism of moral advertisement. Accordingly, I think we must resist Goodwin's further suggestion that moral advertisement “could equally well be achieved by indicating a strong subjective preference to abide by” the norm in question. Not only does advertising our objectification of that norm (rather than simply a subjective preference for compliance with it) convey the expectation that others should also comply (as Goodwin clearly recognizes in formulating his own version of this mechanism), but it also assures others that we are motivated by it in the right way, such that we do not feel free to simply change our minds (as we do with merely subjective preferences) or faultlessly trade-off our subjective preference for compliance with the norm against other subjective preferences we might have. Those who externalize the demand to oppose Nazis prefer to interact with those who externalize the demand to oppose Nazis rather than those who regard their opposition to Nazis as a mere subjective preference or desire. This also illuminates the process by which moral demands become iterated: We devalue not only Nazis as social partners, but also those who tolerate Nazis as social partners, those who tolerate those who tolerate Nazis as social partners, and so on, further correlating interactions between those who do in fact externalize this norm (although, unsurprisingly, the relevant moral demand appears to become progressively weaker at each stage of such iteration).

But Goodwin also wants to know whether further features of norms (like their content) play a role in determining the extent to which disagreement concerning them motivates social exclusion. Re-analyzing the data reported in Goodwin and Darley (Reference Goodwin and Darley2012), he finds a significant correlation for each particular norm between its degree of objectification by an agent and social avoidance or discomfort with a disagreeing party, though also quite a wide and heterogeneous range in the magnitude of those correlations (suggesting that some forms of moralized disagreement are much more important than others in determining our preferred social distance from those with whom we disagree). Given earlier results, it is somewhat surprising that he finds no effect of valence, but he is quite right to suggest that this null result is perfectly consistent with the account offered in the target article, as the valence of a norm does not seem a particularly reliable indicator of the extent to which adhering to it will enhance cooperation and/or protect prosociality from exploitation. In any case, Goodwin is certainly right to suggest that important open questions remain concerning why objectification is “tied closely to social exclusion for some moral norms and not others.”

A further sympathetic extension of the account proposed in the target article is offered by Allidina & Cunningham, who suggest that because we have relatively few opportunities to witness one another's genuinely consequential moralized behavior, “societies develop norms, games, and conventional rituals that allow for moral behaviour to play out in a relatively more symbolic form” that “allow[s] people to form impressions and predict how others will act in more serious moral situations” (para. 2 in their commentary). Although I suspect this is only one of a number of different ways in which norms seemingly unconcerned with protecting prosociality and cooperation from exploitation can become moralized, it does seem to offer a natural explanation (or partial explanation) for the fact that we moralize so much of our ordinary or everyday behavior, such as standing quietly for the national anthem, dressing modestly, recycling, and our conduct in sports and games (Allidina & Cunningham's examples). This explanation would also seem to have the endorsement of legendary men's basketball coach John Wooden, widely regarded as the original source of the adage “sports do not build character; they reveal it.”

R8. Conclusion: In defense of the folk

Let me close by thanking my commentators once again for their insightful contributions and by returning to a broad theme that runs throughout many of the commentaries addressed specifically to the phenomenology of moral experience and to the beliefs of “the folk” concerning moral objectivity. Moral philosophers will certainly want to know whether the proposal I have offered constitutes a so-called “error theory” of morality (or a “debunking” view of morality), which is to say one that implies that moral claims are generally false because the folk are gravely and systematically in error concerning the character of the objectivity they attribute to moral norms, obligations, motivations, and the like. There are a number of reasons to resist this characterization of the account I have offered, not least of which is the fact that I am certainly prepared to accept the reinterpretation it offers for my own moralizing as a broadly accurate description of what I myself have been up to all along. But I find it hard to even make sense of this question as applied to the beliefs and/or experiences of the folk. The folk tend not to have considered views on subjects like the nature of moral objectivity. They don't mistakenly think that moral norms and obligations are external in the same way that rocks and trees are, or in any other particular way – what the folk know is that such norms and obligations are somehow external to us and motivate us in a way that is somehow importantly different from that of mere preferences and conventional norms (which is true). Perhaps the collection of claims to which many of the folk would assent regarding the character of moral objectivity and/or motivation even includes demonstrable falsehoods or inconsistencies, but we have known since Plato wrote the Euthyphro that nothing could have the entire collection of properties the folk are thought by philosophers to attribute to moral norms and obligations. And it is philosophers who attribute such sharp and determinate beliefs about the nature of moral objectivity and motivation to the folk, not the folk themselves. Or to put things another way, it seems to me a slander against the folk to attribute to them beliefs about the nature of moral objectivity sufficiently clear and determinate to be falsified by the account I have offered and defended here.