Concluding Thoughts
9
Constructivism
A wide variety of things are described as social constructions: from gender and race to property and class boundaries. Behind all such talk is a constructivist account in metaethics that provides an alternative to expressivism, error theory, and realism and has its share of notable adherents—from Gilbert Harman to John Rawls, from Christine Korsgaard to Sharon Street, from Ronald Milo to Roderick Firth, from Michael Smith to Stephen Darwall, from Tim Scanlon to David Copp, and plenty besides.
Asteroids and aardvarks are what we might call mind-independent entities, but what about pet rocks or paperweights? The latter are more plausibly thought of as mind-dependent in terms of their functions served. To be construed as such things is a conventional matter; they need to be thought of in those terms by someone or other. So, although some facts are independent of human minds, not all are, and a plausible question is whether moral facts are best likened to mind-independent entities like squares or salmon, or to mind-dependent objects like hassocks or etiquette.
Moral utterance often seems to presuppose that moral language is at least sometimes assertoric (that is, it asserts things to be a certain way), that on occasion such claims are true, and that morality is objective in one way or another. Nihilists and expressivists have a difficult time accommodating such notions. Speaking roughly, the former (moral error theorists) deny that any moral claims are true because there are no moral properties. The latter (emotivists) deny that moral sentences are assertoric, suggesting instead that they are noncognitive expressions of emotions, preferences, and the like.
Constructivists aim to rectify these deficiencies. Constructivism is a metaethical stance that tries to steer a middle course between realist and expressivist views. It is less an outright rejection of moral realism than a substantive revision. If moral properties are mind-dependent, according to constructivists, “then we could say that a central function of moral discourse is to represent these properties, that the content of some such discourse is true, and that morality is objective in a fairly robust sense.”1 Moreover, they think they can secure these advantages without susceptibility to the vulnerabilities of moral realism, especially its aforementioned ontologically odd or metaphysically extravagant features.
Let us now consider a typical range of proposals and briefly offer some explanation and critical comments on each of them. Given space constraints, of course, we will need to be deplorably brief. In fact, in light of how implausible individualist subjectivism is, we will skip it for present purposes and move straight to ethical relativism.2
Ethical Relativism
The problems with individualist subjectivism might lead some to think the way to avoid such problems is by relativizing morality not to individuals but rather to culture. On this view,
X is moral if and only if (henceforth “iff”), and because, A endorses X, where A is a group of agents.
Allan Bloom began his 1987 book The Closing of the American Mind by writing that there is one thing a professor can be absolutely certain of: Almost every student entering the university believes, or says he believes, that truth is relative.3 On the moral variant of this view, which we can call ethical relativism, X is right (or good) amounts to this: Society approves of x. This maneuver neatly avoids making morality a purely individualist affair; moreover, the view is a cognitivist one. Still, problems abound.
For example, most of us are inclined to think that just as an individual can get it wrong, so can a whole society. When the majority or preponderance of Germans, swayed by Goebbels’s propaganda, were supportive of the Nazi cause and the wholesale slaughter of Jews and Poles, surely that did not make such horrible genocide morally legitimate, contra the relativists. Moral relativism retains the category of morality, including language of moral truth; but the truth they affirm, in this case, would be the unspeakable claim that genocide was morally right. So much the worse for ethical relativism!
We made reference to the majority or preponderance of a culture endorsing a practice, and this raises a rather interesting challenge for ethical relativism. Who speaks for a culture or a society? The majority? What if, on a particular vexed question, there are three salient moral camps, none of which hold a majority? In that case, does the biggest voting bloc get final say, in which case a minority voice speaks for the whole culture? What about finer-grained gradations among the winning group? The winning bloc with the decisive say likely does not speak with one voice, so the majority or preponderance among it may have the authority to dictate what the group has to say as a whole. As a decision procedure, this policy is likely to result in increasingly fragmented perspectives, where ever smaller voting blocs could dictate the official moral stance of an entire culture.
Indeed, why, some might query, should the culture be forced to speak as a whole? A society is made up of disparate subcultures, after all, each of which might have its own moral commitments. Why relativize morality to culture per se rather than, say, subculture or sub-subculture? Moreover, since individuals are not part of just one subculture, but many, what would determine what is right for them? Which subculture gets assigned primacy when there is a conflict? And again, since hardly any two individuals even within narrowly defined subcultures share all and only the same moral views, what is to stop the decision procedure from eventually devolving into relativizing morality to the individual all over again?
If we stick to the level of the culture, let us think about a specific example in the history of the United States. There was a time when most Americans were in favor of one of the worst forms of slavery the world has ever seen. If this empirical fact was enough to render slavery morally permissible, if not obligatory, what can be said about those who actively opposed the barbaric institution? The altogether unpalatable implication of the view is that such prophetic voices were actually morally wrong! At least until their influence eventually won the day and the majority of people decided slavery was wrong—at which point slavery actually became morally wrong for the nation. But this makes the deliverances of morality a simple matter of taking a vote or a poll. Maybe the civil laws of the land are roughly a function of such a policy, but the notion that morality itself operates in such a manner strains credulity.
It would be natural to say that when the nation did eventually “change its mind” on slavery, it underwent moral progress. It went from sanctioning and condoning the treatment of fellow human beings as chattel to denouncing and condemning it. If anything qualifies as a moral move in the right direction, it would be something like that. But ethical relativism has a notoriously hard time making room for the very category of moral progress. After all, according to this approach, the nation was right when, in the main, it condoned the institution; and then it was just as right when it changed its tune and decided to condemn it. Individuals may have been wrong on both sides of the divide (for not echoing the prevailing convictions), but the culture as a whole was never wrong. Surely a reasonable starting data point in our analysis of metaethics, however, is that a culture on occasion can be morally mistaken.
Gilbert Harman is one of the best-known advocates for a theory of ethical relativism. He envisioned morality to be the result of a human agreement, which raises an important question. Ethical relativism may or may not involve an actual agreement in the “social contract” of morality. The problem with a hypothetical agreement is seeing why it would be actually binding, but Harman thinks human social agreements need to be actual. Since such agreements can markedly differ, moral claims are analogous to claims about uniform motion in the context of Einstein’s special theory of relativity. Just as claims about uniform motion are relative to spatiotemporal frameworks, so claims about morality are relative to specific contexts shaped by particular social agreements. They are not “absolute.”4 There is no one right moral framework that holds for all.
C. Stephen Evans raises questions about whether Harman’s relativism really differs significantly from nihilism or expressivism. Harman admitted his view cannot accommodate universal moral obligations. But the problem is worse: Consider a Mafioso told by his superiors to commit a murder. Harman asserted that it is a misuse of language to say of the assassin that he ought not to kill or that it would be wrong of him to do so. Evans demurs, replying that it is hardly a misuse of ordinary moral language to say that this employee of Murder, Inc. ought not to kill. To say that morality has authority is in part to say that such claims about what people ought to do can be true regardless of what people believe.5
In addition, if the motivation for the agreement is self-interest, as Harman claimed, it is not clear that it is unreasonable for people to fail to keep their moral agreements. Why should not a person agree to live by a set of moral rules and yet selectively disregard those rules when it is in his interest to do so? This is particularly a problem for Harman’s view that what creates moral obligations is the individual’s decision to accept a particular moral agreement as binding. Decisions are notoriously reversible. One can walk away or can choose to exploit the conformity of others while practicing covert nonconformity oneself. Harman anticipated this objection about “free riders” but without offering a good solution. Harman saw moral obligations as grounded in relative agreements we can always opt out of; such a view does not differ significantly from a denial that any moral obligations exist.
As before, why obey social customs? Either there are good reasons for them or there are not. If there are not, they are merely arbitrary, and though there may be prudential reasons to comply with them, there is no reason to think such arbitrary impositions carry any moral weight. If there are good reasons, it is those reasons rather than the social agreement per se that deserve our allegiance.
Finally, imagine two nations, one peace-loving and pacifistic, the other brutal and aggressive, and that the clear majority in each culture embraces their culture’s dominant ethic. Now imagine the violent and predatory country deciding to invade and subjugate the docile and peaceful country. Ethical relativism would declare that anyone in the peaceful country who resisted such domination would be in the wrong, and likewise would be anyone in the violent country opposed to the invasion. In fact, those in the predominantly peaceful country would be obligated to submit to the invasion, and those in the predatory nation would be obligated to comply with the violence their nation incurred. These are among the unpalatable and deeply counterintuitive implications of relativism.
In sum, there seem to be few reasons to take moral relativism seriously, despite its popularity in some circles. Like its individualist cousin, it conflates belief/practice with truth. It also assumes as sacrosanct at least the fundamental values of any and every culture, vitiates the category of moral progress, and reduces moral disagreement to disagreement over empirical questions about prevailing views. It invests social endorsement with an authority it does not plausibly possess. The alleged social agreements on which it relies are either imaginary or actual, but in either contingency, at least in a case like Harman’s, they are nonbinding. In light of the various subcultures of which we are all members, it is unclear which level of society has final say, and because of fine-grained distinctions even between members of the same subcultures, it likely generates a trajectory to reduce morality to an individualist affair all over again. Its problematic implications are numerous, but, again, more sophisticated variants of constructivism remain worth our while to consider.6
Rational Subjectivism
A rational subjectivist view like that of Michael Smith suggests that
X is moral iff, and because, A would endorse X under ideal conditions, where A is an individual agent.
Here the context of the endorsement matters; the key is not simply what one happens to desire but rather what one would desire in certain idealized conditions of reflection. This move from the actual to the ideal is a recurring strategy among constructivists aiming to capture a stronger sense of objectivity, which comes in various gradations.
The fact that constructivism is a cognitivist theory provides all the objectivity some may say they want, but recall that even error theory can accommodate such minimal objectivity. Error theorists agree that moral judgments have a truth value; it just so happens it is always false. Most constructivists insist on more objectivity than that, such as the possibility of being at least sometimes right in our moral judgments. On our view, such a nonnihilist condition is necessary for objectivity but still not sufficient. What more is needed? Well, on the other end of the spectrum from minimal objectivity is strong or robust objectivity that requires something like stance-independence (or mind-independence). Recall that this is the sort of objectivity associated with moral realism that we hope to defend. Clearly enough, rational subjectivism, like the other constructivist theories, is unable to achieve that kind of objectivity.
Despite that Smith calls his view a version of moral realism, this does not seem right. For reasons that will become clear soon enough, rational subjectivism does not qualify as a realist theory. But why think it is the wrong account? It is an appreciable improvement, after all, on those theories that do not make appeal to idealized conditions. Smith aspires to bring together objective moral facts and the intrinsic ability of morality to confer reasons and/or motivations to act in compliance with it, by moving the discussion from actual reasons to reasons generated in an idealized context—one in which a moral agent is well informed, calm, cool, and collected.
The main reason such a view is not a realist view of ethics is that there is nothing that moral agents discover about the content of morality. There is nothing antecedent to one’s desire to grasp. One’s idealized desires constitute what is right and wrong, as Moreland and Horner put it.7 Confirming the nonrealist nature of the account is Smith’s admission that since reasons are person-relative, rational subjectivism almost invariably leads to relativism, with all its attendant problems. Smith holds out the hope that this contingency will eventually cease manifesting once people “converge” in their moral judgments. It has not happened yet, but Smith holds out the hope it will. But on what basis? Converge on what? There is no objective morality to discover and thus converge on, according to this view, so hopes it will happen seem little more than Pollyannaish, unmoored wishful thinking.
Smith also seems to think, as Plato did, that eliminating ignorance will ensure good moral choices, but history is replete with examples of people with no shortage of education or ability to reason but whose moral choices are arguably abominable. To choose a stark example, consider Theodore Kaczynski, the Unabomber, notorious for his acts of domestic terrorism. He was also a math prodigy and brilliant mathematics professor. Reflecting on his life, Kaczynski’s sister-in-law, a philosopher named Linda Patrik, wrote this about what she learned from the whole sordid affair: “Based on this whole experience, I have lost respect for tremendous intellect. I have discovered that genius needs to be coupled with heart and loving relationships with people to have a positive impact on society. I now know that intellectual brilliance alone has great dangers.”8
Some might balk at this particular example because, they assume, he had severe mental issues. This is likely true, but this is consistent with our point. G. K. Chesterton once wrote that a madman is not someone who has lost his reason but someone who has lost everything but his reason. Moral propriety is not ensured by intelligence or education. Our desires can be warped, twisted, perverse; as Augustine said, they can be either ordinate or inordinate. Smith seems to be operating under the misguided notion that one’s desires are in proper working order, which is a large assumption.
Smith is also operating with an instrumental, self-interested understanding of rationality. That strategic choice immediately delimits the sort of conclusions one is able to draw. Morality is left bereft of the otherwise rich resources at its disposal when a more expansive conception of rationality is allowed to operate. For the most part, this attenuation of rationality represents a significant departure from ancient and medieval construals of reason that allowed for the possibility of its apprehension and appreciation of normative features of reality. It is, at the least, not at all obvious that this emaciated modernist conception of rationality should be assumed as authoritative. Interestingly, Patrik went on to add another lesson she had learned from her brother-in-law’s tragedy: “Sometimes ethical decisions need to be made that have nothing to do with self-interest or self-benefit.”9
Ideal Observer Theory
Another variant of constructivism that emerged in the twentieth century pushes yet harder for more objectivity: the ideal observer theory (IOT). According to IOT, morality is dictated by the hypothetical deliverances of someone perfectly well qualified and situated to make such judgments. Such an ideal observer would see clearly, reason impeccably, and not be at all afflicted with blind spots, distorting biases, or anything else that would cause epistemic deficiency, disadvantage, or dysfunction. Rather than an actual contract between the members of a society, IOT features a hypothetical and ideal one. By appeal to the verdicts of an ideal observer, someone eminently if not perfectly qualified to make authoritative ethical pronouncements about a situation, advocates argue that this feature of their approach lends itself to obligations with great prescriptive force that are perhaps even universal in scope. On IOT,
X is moral iff, and because, A would endorse X, where A is an ideal hypothetical agent.
The locus classicus of such a view is Roderick Firth’s piece on the subject titled “Ethical Absolutism and the Ideal Observer.”10 Recall that rational subjectivism would have us imagine, say, an idealized you or me, including potentially unhealthy desires, but IOT further idealizes the agent in question. What would a hypothetical, ideally rational agent endorse? This further step toward abstraction can be thought to serve as a corrective to the problems with rational subjectivism. Ethical judgments are interpreted by IOT as statements about the reactions of a fully informed and neutral agent.
What are the traits that this ideal observer possesses? According to Firth, the ideal observer is a normal person save for these characteristics: omniscient regarding relevant nonmoral facts, omnipercipient, dispassionate, disinterested, and consistent. Important to note is what is absent from the list: any moral attributes. This is necessary because, on Firth’s view, there are no moral facts save for those that come into existence as a result of endorsement by the ideal observer, and moral traits would require instantiation of prior moral facts, contra IOT. Complete moral knowledge is not an intrinsic feature of the ideal observer but rather an emergent property from the combination of the ideal observer’s other traits and its knowledge of pertinent nonmoral facts.
A variety of concerns can be raised about the adequacy of IOT as a metaethical theory. Among them Moreland and Horner raise four issues: First, again, an operative assumption is that the sole source of moral problems is an agent’s lack of nonmoral information or his being sidetracked by contrary emotions from proper self-interested reasoning. Second, it is difficult to imagine what a hypothetical nonmoral agent would endorse when his reasoning is utterly abstracted away from all normative conceptions. Anything like benevolence is a moral quality, but moral qualities are ruled out for the ideal observer. Third, there is no reason to think that on this picture the ideal observer would arrive at anything like fundamental concerns for justice and caring for others. And fourth, since we are dealing with a purely hypothetical agent, why think that the agent’s deliverances, such as they are, would have any authority over one’s actual values and behaviors?11
The objection we will point to comes from a different analyst—namely, Alan Donagan. In his Theory of Morality, he explores the prospect of IOT providing what he called the stock of common moral convictions. Donagan was particularly influenced by the fiction of George Orwell and the need to safeguard human value. Ambiguities and limitations in IOT, he thought, pose a problem for achieving such an end. He constructed an analogy: Imagine if our explanation for a certain sum of two numbers was that this is the result generated by the ideal calculator without any explanation provided of how the calculator managed the challenge of addition.
We would be dissatisfied by such an answer, as we should be when someone points to the deliverances of the ideal observer with precious few specific details for how the ideal observer generates moral answers. Donagan anticipated that one might wish to say that the operation of the moral sense does not lend itself to providing those details. Since Donagan traced an important strand of the history of IOT to Adam Smith, it would make sense that an appeal to something like a moral sense would be made since this was a salient feature of several Scottish Enlightenment figures (especially Hutcheson). As for Donagan, though, as with most today, he was quite skeptical of any such sense.
Donagan thought that if such skepticism is warranted, this meant the deliverances of the ideal observer should lend themselves to a rational explanation instead. If so, he thought, some description of those rational processes should be forthcoming. This is especially pressing when adjudicating conflicting moral claims from equally consistent moral theories. How would an ideal observer, for example, decide between sacrificing one for the sake of many or upholding the rights of the one not to be thus sacrificed? Donagan was profoundly skeptical that IOT had a good solution.
The objection he considered fatal for IOT was that either there are no principles that one must accept by following the procedure prescribed or there is no way of assuring oneself that one has succeeded in following it.12 This is no merely academic matter. Tom Carson observed many years ago about ideal observer theories, “For any view or attitude, however absurd or objectionable, one can easily conceive of a possible creature who knows all of the relevant facts, etc., and who holds the view or attitude in question.”13
Contractarian Constructivism
Suppose we leave behind the individual variants of hypothetical constructivism and opt for a group version instead. This leads us toward what has been called contractarian constructivism (CC), according to which
X is moral iff, and because, it is permitted by rules that free, equal, and rational people would agree to live by, as the condition that others obey these rules as well.
For a paradigmatic contractarian constructivist, consider Ronald Milo, who also attempts to steer a middle course between expressivists and robust moral realists. He agrees with expressivists that there are no stance-independent moral truths, and he thinks that robust realists mistakenly treat moral truths too much like those truths found in science. He thinks that the point of ethics should be kept front and center in order to regulate and direct human behavior. As such he attempts to provide an account of moral truths and facts that is neither expressivist nor robustly realist. As he puts it, “Contractarian constructivism, as I shall call it, holds that moral truths are most plausibly construed as truths about an ideal social order, rather than the natural (or some curious nonnatural) order of things. It is true (or is a fact) that a certain kind of act is wrong, for example, just in case a social order prohibiting such acts would be chosen by rational contractors under suitably idealized conditions.”14
On this view, the ideal social order described by moral truths is constructed; it is not an independent moral order. Moreover, Milo insists that the moral principles chosen by the hypothetical contractors are taken by those contractors as action guides rather than truth claims. Moral principles are true just in case the action guide in question is the object of a certain kind of rational social choice. This is to see moral truths as practical truths rather than theoretical truths. We are justified in accepting the latter, Milo writes, because of the explanatory necessity of positing them. Practical truths, in contrast, are truths about what there is reason, for some individual or group of individuals, to prefer, choose, or do from some point of view. Among such points of view is the moral point of view.
Practical truths are discovered through practical reasoning. Whereas theoretical reasoning aims at determining what to believe, practical reasoning aims at determining what to choose. Rather than inferences to the best explanation, like those we often find in theoretical reasoning, the modus operandi of practical reasoning tends to be something more like an “inference to the best means”—that is, an attempt to figure out how best to achieve reflective equilibrium among our beliefs and our desires.
Rather than endorsing a normative theory of constructivism, which would leave underdetermined the nature of moral facts, Milo defends a metaethical variant of constructivism. This leaves unfinished the task of delineating the subject matter of moral judgments. He is attempting to say something about the ontological status of moral facts like moral wrongness. Such facts are, he thinks, evidence-independent; in this way, his view resembles moral realism. This is to say, in both cases, there may well be moral truths for which there is currently a lack of evidence. But unlike robust moral realism, his constructivism concedes to the antirealists that moral facts are stance-dependent, reliant on human volition in ideal conditions.
CC is an attempt to achieve a greater level of objectivity than the aforementioned forms of constructivism. The objectivity achieved is not of the sort that puts us in touch with a transcendent realm of reality; rather, it derives from the fact that moral facts describe an objectively preferable social order. In sum, Milo thinks that he has shown that his view provides a plausible account of what moral thought and talk is really about that captures the essentially action-guiding function of moral principles without depriving them of their aspirations to substantive truth or the dignity of a cognitive status.15
This view remains afflicted with several of the aforementioned deficiencies. Even though its orientation is more social than individual, the view still seems predicated on according primacy to self-interested considerations that are far from morally neutral and make it challenging to account for the other-directed concerns of morality.
In this connection, although CC might help explain certain constraints on morality, morality is much richer than that. For example, even if we suppose that CC carves out room for rights and duties in some form, there is ever so much more to morality beyond those deontic concerns. Consider, for example, the value of human persons. Arguably the recognition of such a thing, though it clearly has implications for our behavior, goes beyond that one consideration alone. Human value is most usually thought to be not conferred but recognized. Additionally, unless there is deep conviction about such a matter, its prescriptive grip on our behavior and imagination, as history has tragically shown time and again, is notoriously tenuous. The problem becomes especially acute in cases where some simply choose not to adopt the “moral stance.” As an example, rhetoricians speak of reification among demagogues, the treating of persons as things in order to demonize and silence them, for which there is no shortage of historical examples.16
Relatedly, as Moreland and Horner argue, CC does not even explain moral constraints very well because of the existence of free riders. If one were to think one could get away with being a free rider—exploiting the fact that most others will not do so—what compelling reason, moral or rational, is there against doing so? And at any rate, why consider as binding the deliverances of hypothetical agents—especially in light of CC’s watered-down account of what counts as moral truth?
Humean and Kantian Constructivism
Sharon Street’s important constructivist proposal, in the spirit of Hume, aims to provide a ground for all normative facts, not merely moral ones. According to her view,
X is a reason for a person to do Y iff, and because, the judgment that X is a reason for the person to Y withstands scrutiny from the standpoint of the person’s other judgments about reasons.17
According to Street, whenever we make a normative judgment, there are standards involved by which our reasoning can be judged (whether we consciously recognize it or not). She writes, “Simply by judging yourself to have reason to Y, you’re thereby—as a constitutive matter—also judging yourself to have reason to take the means to Y, whatever those may be. So even if you don’t know that Z is a means to Y, and think you have no reason whatsoever to Z, you do have a reason to Z—according to you. Your very own normative judgment says so; it sets the standard according to which you are making a mistake if you think you have no reason to Z.”18
For a normative judgment to withstand scrutiny is “for that judgment not to be mistaken as determined by the standards of correctness that are constitutively set by those other normative judgments in combination with the non-normative facts.”19
As Christian Miller sums up Street’s view, “What normative judgments we happen to have now are not likely going to all withstand scrutiny in Street’s sense. Rather, it is only the hypothetical set of a person’s normative judgments which have been appropriately filtered and purged, that Street thinks is the basis for the normative facts which pertain to that person.”20
Unlike the Kantian constructivists, Street rejects the idea that the result of the filtering process will be one or more values that everyone must have. As she writes, “If you had entered the world taking entirely different things to be reasons, on my view, you would have had entirely different reasons. You might even have had reason to throw yourself under an oncoming boulder, or to die for the sake of eating a particular bowl of chili, or indeed to prefer the destruction of the world to the scratching of your finger.”21
At this point, Miller argues that what he calls the constructivist dilemma can be brought to bear against Street’s view. Patterned after the famous Euthyphro Dilemma, the constructivist dilemma poses a problem for both restricted and unrestricted versions of constructivism. The unrestricted versions fall prey to what corresponds with the objection to the voluntarist horn of the Euthyphro Dilemma. Miller calls it the “Anything Goes Concern”: In principle at least, any action, no matter how heinous, could become morally positive if the respondents were to respond positively to its being performed.
Miller writes that we end up getting a conditional like this: “If [respondents] were to have [some positive response] to someone at some time bringing about the torture to death of an innocent child, then it would be [some positive normative evaluation] for someone at that time to bring about the torture to death of an innocent child.”22 Obviously, this implication is clearly morally disturbing, and no plausible theory should allow it.
As an unrestricted approach, Street’s view is vulnerable to just this concern. Indeed, as Miller notes, Street herself provides us with the possibility on her view of there being a fully consistent Caligula, bent on the destruction of other people. As she writes, “If Caligula is aware of all the non-normative facts and has recognized every normative conclusion that follows from his own values in combination with those facts, then there is nothing he is failing to see.”23 This implication is clearly disturbing, Miller argues, and no respectable theory should countenance it.
Whereas Hume assigned primacy to intersubjective human sentiments, Kant wished to move the discussion to the dictates of rationality. As a result, Kantian constructivism, contra Humean constructivism, insists that there are at least some moral values that everyone must have. Such values are thought to be the result of rational deliverances. Christine Korsgaard is a notable contemporary example of a Kantian constructivist, and on her view,
X is moral iff, and because, X is rationally willed by ideally rational agents constrained by necessary structures of practical reason.24
On this construal of Kant, moral reasoning has a built-in structure with formal rational constraints. Moral facts thus do not find their locus outside the psychological and volitional activity of agents. Nevertheless, they are not, as on the Humean picture, desire-based reasons for action grounded in self-interested rationality but rather grounded formally in the structure of practical reason itself. They are, as a result, not contingent but necessary “to all reasoning about what to do. They apply categorically and universally to all rational agents, regardless of the agent’s contingent desires and aims.”25
Moreland and Horner, however, offer this critique of Kantian constructivism: “Though it overcomes many of the weaknesses of other constructivist views, it fails to capture their strengths in terms of connecting morality to persons and personal motivation. On the standard interpretation, Kant’s understanding of morality is austere and impersonal, abstracted away from contingent human concerns and desires. Morality amounts to formal rational consistency—a far cry from seeing it as embedded within the context of persons and personal relationships.”26
An arguably even stronger objection is that constructivism gets the explanatory order inverted. The deliberations of ideally rational agents will disclose to us what is moral just in case objective moral values exist explanatorily prior to their deliberations. Their deliberations cannot be explanatory prior to the existence of objective moral values because, in that case, ideally rational agents might all agree with Mackie and be moral antirealists!
Perhaps even more importantly, for present purposes, is that Kant was one of the great luminaries in the history of the moral argument for God. A merely constructivist construal of Kant’s views along the lines of Korsgaard would fail to account for how this could be so. Owing to the likes of Karl Ameriks, John Hare, Stephen Palmquist, and others, a recent sea change has taken place in Kantian studies that has assigned greater pride of place to theism in his construal of morality. One piece of the resultant picture is the way Kant argued that moral duties should be understood as the commands of God.
Because of the centrality of Kant to the historical development of the moral argument, and because theistic ethics on occasion is interpreted along constructivist lines, we need to spend a bit of time on this question of God and ethics before we close this chapter.
Theistic Ethics
Constructivism is a fascinating effort in metaethics to salvage mind-dependent ethical truths that are nevertheless sufficiently objective. If the effort succeeds, it shows a middle course can be steered between expressivism and moral realism. The array of constructivist proposals is impressively wide, and we have been able to briefly consider only a handful of examples here.
This makes us hesitant to dismiss all constructivist proposals as problematic or inadequate, but we have come to see that a number of recurring problems do tend to beset and plague constructivism. If there is a theoretical framework of metaethics at least in the vicinity of constructivism that holds potential to withstand critical scrutiny, some might wish to argue that it is the effort to make morality, or at least aspects of it, dependent not on human minds but rather on a divine mind. This maneuver introduces the topic of theistic ethics.
The sort of realism that we have been most interested in locating, if we are able, is moral realism of a robustly objective sort. Most constructivist proposals, as we have seen, do not provide what we are looking for, because they remain problematically mind-dependent. The moral truths construed along such lines may be objective in a weak sense but are not objective in the stronger sense of being mind-independent.
Now, what about morality that finds its foundation in a supernatural being like God? For purposes of discussion, suppose by “God” here we are referring to God as traditionally conceived, a maximally great being. So God is, thus construed, all-powerful, all-good, all-knowing, and the like. What might it mean for God to provide the foundation of morality? Let us put it generally and suggest that it means that the various aspects of morality essentially depend on God—perhaps God’s nature, desires, will, or commands. At this point we are trying to keep the definition rather general because there is a variety of ways in which morality has been argued to depend on God: natural law theory, divine command theory, divine desire theory, divine will theory, exemplarism, and more besides. The differences between such theories are not unimportant, but what we want to consider for now is the general idea that morality, in one way or another, importantly depends on God, derives its authority from God, or something in that neighborhood.
To say that morality depends on God might be thought to be another variant of constructivist ethics. To illustrate why, consider the paradigmatic example of divine command theory (DCT), which is usually offered as an account of moral duties, as follows:
X is morally right (dutiful) iff, and because, X is commanded by God.
This analysis suggests that moral duties result from God’s commands, but of course this leaves other moral concepts and phenomena—like moral goodness, moral rights, and so on—unexplained. If someone were to offer an unlimited divine command theory in order to cover the full range of moral realities, that would be a fully or radically voluntarist theistic account of morality. But such a view tends to be rare in the scholarly literature on the subject. It is much more common to see DCT, when it is referred to at all, as delimited to moral duties—following the advice from William Alston several decades back.27 This is what we can call a weakly voluntarist account, and it is the theory we wish to consider for a moment.
Why some argue that this is a constructivist account should be fairly obvious. If something is morally obligatory because it is ultimately grounded in God’s commands, then there is a divine mind-dependence at work here. At least an aspect of morality would be a function of an aspect of the divine mind. As such, Michael Huemer argues that theistic ethics, thus construed, is an instance of constructivism and, as a result, subject to most if not all of the various criticisms that have been lodged against constructivism.28
In particular, an objection to divine command theory that is quite common derives from Euthyphro. Recall Miller’s use of a generalized Euthyphro-styled objection to certain variants of constructivism. The Euthyphro Dilemma, put in contemporary terms, goes something like this: Is something morally right (dutiful) because God commands it, or does God command it because it is right? Both horns of the Dilemma seem problematic, as we saw in part 1.
The former horn would affirm a limited account of morality: a divine command theory of obligations. But this seems problematic for various reasons, including an arbitrariness concern. Imagine if God were to command us to torture children for fun. Then we would be morally obligated to do that, which is obviously a problem for the theory. The latter horn of the Dilemma affirms that God issues commands, but what he tells us to do is what is already antecedently right. So this renders God at most as the source of some moral knowledge but not the ultimate foundation of morality. So moral rightness is either arbitrary or beyond the scope of God’s authority; so much the worse for DCT.
With that cursory foundation in place, consider now the constructivist proposal advanced by Stephen Darwall, which illustrates how the Euthyphro Dilemma can be used to criticize constructivism more generally and what might be said by way of reply. Darwall has argued for a “juridical account” of moral duties inspired by themes in Fichte and Kant but also developed in conversation with early modern “theological voluntarists” like Pufendorf and Locke.29 Darwall argues that moral obligations are grounded in the demands of persons. Moral obligations are linked to being accountable for such demands, but God is not needed. Rather, all that is required is a moral community of equal, autonomous human agents who hold each other accountable.
For Darwall, a moral obligation is not simply an act one has a moral reason to do. Rather, moral obligations derive from claims or demands that persons make on one another. Moral obligations are moral demands, and moral demands are whatever we are legitimately held to or answerable for—that is, where violations are blameworthy if the agent lacks a valid excuse. If someone steps on your toe and you request the offending party to stop doing so, refusal would be wrong in a new way because ignoring your justified request shows you a lack of respect.
Darwall, however, does not think that all duties derive from claims made by actual persons. Some behaviors are wrong simpliciter, and these are linked to a demand that one would be blameworthy for not recognizing even if the demand is not made by anyone. So he thinks that there are many moral obligations not grounded in actual demands made by anyone. C. Stephen Evans argues that this poses a problem for Darwall, whose sympathies seem to be with a “procedural” constructivist view. Evans writes, “While Darwall draws on both Kantian and contractualist forms of constructivism, he seems to be offering yet another form by proposing that moral obligations are grounded in the claims human persons make on each other.”30 This introduces a tension in Darwall’s view.
Since not every claim is a legitimate moral claim—the demands of a bully, for example—justified or warranted claims are normative in character, and these normative facts must hold antecedently to any claims any actual person makes. This is no criticism of Darwall, just a feature of his view; and it parallels what we find in the divine command theory like that of Robert Adams. Moral obligations can be thought to presuppose use of other normative concepts; it would seem that both Darwall and Adams do this. In fact, Darwall sees a parallel between his own view and DCT, but he replaces God with the human moral community.
DCT is confronted with a “prior obligations” objection: Why obey God’s commands? Simply to say it is because God commands it seems objectionably circular. Similarly, Darwall’s view holds that some obligations do not arise from the demands of actual persons; there is a prior obligation to respect the legitimate demands of other people even when no demands have been made. The problem Evans raises is this: If Darwall is understood as a constructivist, are the “prior obligations” that hold antecedently to any actual claims made by actual persons something that the constructivist can explain? Can “the moral community,” understood as a “regulative ideal,” do the job? If not, then there might seem to be moral obligations that are not second-personal in nature after all.31
A divine command theorist like Robert Adams could make appeal to normative truths rooted in God but not rooted in God’s commands, and this is exactly what he did, predicating his DCT on a theistic Platonic account of the good. The challenge for Darwall is arguably worse because the prior normative truths he appeals to are not rooted in any actual commands. Darwall can, and does, attempt to appeal to hypothetical demands, but this approach faces its own Euthyphro problem. Does the claim in question make the demand moral, or is the claim based on what is already morally right? The former leads to arbitrariness, whereas the latter departs from Darwall’s paradigm.
As we have seen, it is also unclear how moral obligations are constructed from nonactual persons or communities. Evans adds, “Just as it is sometimes claimed that Kant’s categorical imperative is too abstract and formal to give rise to determinate moral principles, it may be that the idea of the ‘moral community’ is too abstract and formal to give rise to determinate moral obligations.”32
Antecedent moral principles sound suspiciously more like moral realism than constructivism. Likewise, the moral autonomy to which Darwall appeals as partially constitutive of intrinsic human value also sounds like an appeal to moral realism. So, despite resonances between Darwall’s view and constructivism, it seems as if something much like moral realism is needed to salvage his view in the face of Euthyphro-like objections. The hypothetical demands Darwall appeals to sound much like the timeless moral truths of the sort that nonnatural realists accept.
Moreover, at just this juncture, DCT holds an advantage over Darwall’s view. God is taken, on DCT, to be an actual being, whose commands are not just hypothetical. Recall a problem with IOT: Why care about the judgments of a nonexistent being? “However, if God actually exists,” Evans writes, “God’s judgments might matter greatly. For example, it may be argued that humans owe a great debt of gratitude to God as the creator of every good we have, or it might be argued that our ultimate happiness is linked to a loving relationship with God.”33
It is not part of our agenda at this point to argue for DCT or theistic ethics more generally, although we are keen on leaving such options on the table for consideration. We tend to find many dismissals of them too hasty, especially those posed by the Euthyphro Dilemma against radical voluntarism. As Evans just argued, too, it is not at all clear why DCT might not be used to avoid pitfalls encountered by Darwall’s inventive account.
We have admitted that it is difficult to dismiss constructivism entirely and that some variant of it might pass muster. To our thinking, the best prospect for doing so is something like a robust theistic ethic, which makes aspects of morality stance-dependent but dependent on God’s mind rather than that of human beings. And it does so potentially in a way that retains cognitivism, nonnihilism, robust objectivity, and even moral necessity, while still retaining a personalist component to the ethical enterprise. Often these priorities seem at odds, essentially irreconcilable. But if such a project is doable, a sort of synthesis of key elements of moral realism and constructivism may well be possible and preferable.
This brings to mind some words from Thomas V. Morris, who presciently anticipated just such a development, arguing that key elements of Platonism and theism can be reconciled, yielding “a view which both retains the commitments of realism concerning the objective existence and status of abstract entities and modal truths, while at the same time capturing the conviction of anti-realists and conventionalists that such items must be in some sense mind-dependent. And this in itself is quite significant.”34
Concluding Thoughts
This chapter has been a circuitous journey. We consider constructivist theories to be nothing less than fascinating, and many variants of it are clever indeed and well worth careful consideration. We even suspect that certain key elements of constructivism and realism can achieve something of an integration, helping to account for both the strongly objectivist flavor of morality and its deeply personalist aspects as well. We traversed quite a bit of terrain: We discussed, in turn, ethical relativism, rational subjectivism, ideal observer theory, contractarian constructivism, Humean and Kantian constructivism, and theistic ethics.
Constructivism, in general, seems to fall prey to several recurring deficiencies. Often it fails to capture the requisite level of objectivity that moral language, logic, and phenomenology require. Where it refers to actual persons and their preferences, agreements, and the like, there often seems to be the option of backing out from such contracts. Movement toward more idealized situations and hypothetical contracts leaves us wondering where the authority of such nonactual entities comes from and why their deliberations should constitute rather than merely disclose to us the moral good. The problem of free riders is chronic.
With the possible exception of DCT, constructivism has a difficult time making good enough sense of binding moral duties. Moreover, what is ultimately required in a robust metaethical account is more than merely an account of moral duties. A whole panoply of other moral phenomena—from moral values to basic human rights, the demands of justice to moral knowledge, and more besides—cry out for more adequate explanation than what constructivism can seem to explain.
Since we are on the lookout for a defensible variant of moral realism serviceable in a moral argument for God’s existence, constructivist aspects of theistic ethics are not problematic. Moreover, if a theistic ethic can be successfully argued, as it has been by the likes of Robert Adams, C. Stephen Evans, John Hare, and others, to be reconcilable with due regard and adequate principled space for human teleology, binding normativity, necessary moral truth, and robust objectivity, then there is all the more reason to think that moral realism may well evidentially point to God’s reality. It is imperative, however, not just to dismiss theistic ethics based on its worst and weakest examples, like radical voluntarism, but rather to give serious considerations to its best formulations.
We do not make that full case at this point. Instead, we are simply setting the table for such a discussion, arguing that theistic ethics deserves a place in the discussion once it is seen there are good reasons to take moral realism seriously. To make the full case would require articulating a defense of theistic ethics, including Euthyphro objections, and criticizing nontheistic efforts to account for moral realism. For now, we end our discussion of constructivism, concluding, at least tentatively, that when divorced entirely from moral realism it appears to encounter insuperably steep challenges.
1 Russ Shafer-Landau and Terence Cuneo, introduction to “Constructivism,” in Foundations of Ethics: An Anthology, ed. Russ Shafer-Landau and Terence Cuneo (Blackwell, 2007), 79.
2 For a treatment of individualist subjectivism, see David Baggett and Jerry L. Walls, The Good, the Right, and the Real: Is Value a Fact? (Oxford University Press, 2025).
3 Allan Bloom, The Closing of the American Mind (Simon & Schuster, 1987).
4 “Moral right and wrong (good and bad, justice and injustice, virtue and vice, etc.) are always relative to a choice of moral framework. . . . And no moral framework is objectively privileged as the one true morality.” Gilbert Harman, “Is There a Single True Morality?,” in Explaining Value and Other Essays in Moral Philosophy (Oxford University Press, 2000), 78. See also Gilbert Harman, The Nature of Morality: An Introduction to Ethics (Oxford University Press, 1977).
5 C. Stephen Evans, God and Moral Obligation (Oxford University Press, 2013), 132.
6 For a popular and practical treatment of ethical relativism, see Francis J. Beckwith and Gregory Koukl, Relativism: Feet Firmly Planted in the Mid-Air (Baker, 1998).
7 J. P. Moreland and David Horner, Metaethics: A Short Companion (B&H Academic, 2024), 76. This delightful book is a terrific primer on metaethics.
8 “A 1998 Union College Magazine Interview with Professor Linda Patrik,” Union College, May 14, 2008, https://muse.union.edu/newsarchives/category/stories/page/202/.
9 “Interview with Professor Linda Patrik.”
10 Philosophy and Phenomenological Research 12 (1952): 317–45.
11 Moreland and Horner, Metaethics, 78–82.
12 Alan Donagan, The Theory of Morality (University of Chicago Press, 1977), 220–21. See, too, Richard Brandt, “The Definition of an ‘Ideal Observer’ Theory in Ethics,” Philosophy and Phenomenological Research 15, no. 3 (March 1955): 407–13. Brandt was a staunch defender of IOT.
13 Thomas L. Carson, The Status of Morality (D. Reidel, 1984), 76. A problem with IOT is that it concerns moral epistemology rather than ontology. The ideal observer can guide us to the content of the moral good, but no reason has been given to think that the verdict of the ideal observer constitutes the moral good. Only if the moral good exists objectively would the deliberations of the ideal observer be any help in discerning what is good.
14 Ronald Milo, “Contractarian Constructivism,” in Foundations of Ethics: An Anthology, ed. Russ Shafer-Landau and Terence Cuneo (Blackwell, 2007), 121.
15 Milo, “Contractarian Constructivism,” 131.
16 Jennifer Mercieca, Demagogue for President (Texas A&M, 2020), 24–25.
17 See Sharon Street, “Constructivism About Reasons,” in Oxford Studies in Metaethics, vol. 3, ed. Russ Shafer-Landau (Oxford University Press, 2008), 223. See also Sharon Street, “What Is Constructivism in Ethics and Metaethics?,” Philosophy Compass 5 (2010): 369.
18 Street, “Constructivism About Reasons,” 229.
19 Street, “Constructivism About Reasons,” 230.
20 Christian Miller, “Constructivism,” in Moral Arguments for God’s Existence: Navigating the Terrain, ed. David Baggett and John E. Hare (Oxford University Press, forthcoming). Miller’s adroit analysis at several points enhanced our work in this chapter.
21 Street, “Constructivism About Reasons,” 244.
22 Miller, “Constructivism.”
23 Street, “What Is Constructivism in Ethics and Metaethics?,” 371.
24 See Christine M. Korsgaard, The Constitution of Agency: Essays on Practical Reason and Moral Psychology (Oxford University Press, 2008), 321.
25 Moreland and Horner, Metaethics, 87.
26 Moreland and Horner, Metaethics, 88.
27 William Alston, “What Euthyphro Should Have Said,” in Philosophy of Religion: A Reader and Guide, ed. William Lane Craig (Rutgers University Press, 2002), 283–98.
28 Michael Huemer, Ethical Intuitionism (Palgrave, 2008), 54–60.
29 Stephen Darwall, The Second-Person Standpoint: Morality, Respect, and Accountability (Harvard University Press, 2006). More recently, see his Morality, Authority, and Law (Oxford University Press, 2013).
30 C. Stephen Evans, “Does Darwall’s Morality of Accountability Require Moral Realism? (And Would It Be Strengthened by Adding God to the Story?),” Religions 12, no. 3 (2021): 187.
31 Evans, “Darwall’s Morality of Accountability,” 187.
32 Evans, “Darwall’s Morality of Accountability,” 187.
33 Evans, “Darwall’s Morality of Accountability,” 187.
34 Thomas V. Morris, Anselmian Explorations (University of Notre Dame Press, 1987), 178.