Hume on moral relativism

It is a matter of biographical record that Smith read Hume’s Treatise of Human Nature while an unhappy visiting student at Balliol College, Oxford in the early 1740s, at which point he may well have begun formulating his philosophical disagreements with the older thinker, who would later become a friend.Footnote ⁵ Yet Smith must also have been familiar with Hume’s Enquiry Concerning the Principles of Morals, published in 1751, eight years before the first edition of The Theory of Moral Sentiments. In both, Hume contends that virtue ultimately consists in that “which is useful or agreeable to the person himself or to others,” with vice being the reverse.Footnote ⁶ On this matter at least, whatever the technical differences between the Treatise and Enquiry—the latter, for example, appearing to give even greater importance to the role of utility—from Smith’s perspective, they are both crucially mistaken. Utility does not play a foundational role in the moral sentiments in the way Hume claims, and hence arguments deployed in The Theory of Moral Sentiments target both works.Footnote ⁷ By contrast, at least one thing that does distinguish the Treatise and the Enquiry is the latter’s “A Dialogue,” which has no direct equivalent in the earlier work.

The reasons Smith likely took special notice of “A Dialogue” will be presented below. It is first necessary to have a clear sense of what Hume achieved (and crucially, did not achieve) in that work. Although more complex a philosophical contribution than it first appears, “A Dialogue” may nonetheless be summarized as follows. Presented as an exchange between an unnamed narrator and his friend Palamedes, the latter launches into a description of a foreign land, Fourli, he has allegedly just returned from visiting. He tells of a people not only engaged in the most outrageous practices—infanticide, treason, parricide, and suicide, to name but a few—but who consider them as virtuous and in various ways to be applauded. When the narrator objects that his friend should stop trying to make a fool of him by getting him to believe in a land that is plainly unbelievable, Palamedes reveals he has been using pseudonyms to describe myriad practices associated with the ancient Greeks and Romans. After more ground clearing, Palamedes stakes his central claim: “[T]o represent the uncertainty of all these judgments concerning characters; and to convince you, that fashion, vogue, custom, and law, were the chief foundation of all moral determinations.” But if this is indeed granted, then severe consequences would seem to follow: “What wide difference, therefore, in the sentiments of morals, must be found between civilized nations and Barbarians, or between nations whose characters have little in common? How shall we pretend to fix a standard for judgments of this nature?”Footnote ⁸

But with Palamedes’s position stated, the narrator immediately hits back:

This is the crux of “A Dialogue.” The narrator purports to show that all moral judgments are ultimately derived from the principles of being either agreeable or useful to self or others.Footnote ¹⁰ Making this move is potent as a way of replying to Palamedes’s challenge. In the first instance, if all moral judgments have a single common source, just as both the Rhine and the Rhone spring from the same point in the Alps, then this has the effect of reducing apparent disparities between different moral practices in different times and places. Yes, on the surface they may look radically different, but the philosopher is able to trace things to “first principles” and see what they more fundamentally have in common, while in the process refuting the claim that there is nothing more to morality than fashion, vogue, custom, and law.

Furthermore, the narrator takes himself to be able to explain why moral practices differ, despite their common source. Different societies will inevitably develop different ways of combining usefulness and agreeability, given that there are many ways in which these can be balanced, traded, and offset. Hence, the French tend to licentiousness with regard to adultery and gallantry, and the Athenians to prudish severity. It is also quite possible that different societies get the balance wrong, actually hindering agreeableness and/or utility. The French go too far in one direction, the Athenians in the other. Similarly, the different circumstances that different societies find themselves in will likewise alter judgments and practices regarding utility and agreeableness: “It is not surprising, that, during a period of war and disorder, the military virtues should be more celebrated than the pacific, and attract more the admiration and attention of mankind.”Footnote ¹¹ Likewise, whether a society is governed according to the principles of a republic or a monarchy, is rich or poor, and so on, will also affect how it configures the virtues. In more favorable circumstances, societies will be apt to configure the virtues better, that is, in terms of promotion of the useful and agreeable over the harmful and the disagreeable.Footnote ¹²

“A Dialogue” presents the case of the narrator as thoroughly compelling. On the one hand, it is not “incumbent” for the philosopher to show that all peoples in all times and places “reason aright with regard to this subject, more than with regard to any other.” When it comes to morality, “[i]t is sufficient, that the original principles of censure or blame are uniform, and that erroneous conclusions can be corrected by sounder reasoning and larger experience,” as the narrator claims to have shown.Footnote ¹³ On the other, not only is it false that there is nothing more to morality than custom, fashion, vogue, and law. It is also possible to assess and make judgments about where morality is doing what it should—that is, promote usefulness and agreeableness—and where it is failing in those precise regards, as in some cases, the narrator happily admits, it can, and hence is in need of reform. But what this does not lead to is the conclusion that just because moral difference can be observed both between different societies and across different eras, this somehow reveals morality as a fraud, arbitrary, or otherwise impugned. Both the Rhine and the Rhone flow across different terrain, but the mere existence of the Rhine does not somehow impugn the status of the Rhone. They are both rivers, and the mere fact they may manifest differently, in different times and places, does not alter the fact that they spring from the same source and are governed by the same principle of gravity. Likewise holds for the status of morality.

At this point, it is germane to ask: But is the argument successful? The (surprisingly) small number of commentators who have focused on “A Dialogue” overwhelmingly conclude that it is.Footnote ¹⁴ I would agree with them—but only to the extent that we keep the debate within the remit that Hume himself lays down. The problem, however, is that this remit is unduly narrow and fails to take account of the most pressing objections. This can be seen as follows.

As reconstructed above, “A Dialogue” clearly engages with the challenge of what we now call moral relativism. Yet the version of the challenge that Hume constructs, as voiced by Palamedes, and that the narrator comprehensively engages with is a relatively superficial and comparatively easy one to address. A far more serious objection would go as follows:

At the very least, Hume’s narrator would need to say something in reply to an objection of this kind. But the philosophical seriousness of the challenge goes deeper:

This is the serious challenge from moral relativism that a sentimentalist account like Hume’s must ultimately say something in reply to. And yet, “A Dialogue” has nothing to offer on this score. Its success, therefore, is at best limited.

Disagreement in ethics versus aesthetics

Before turning our focus to Smith, it is worth considering an important difference between “A Dialogue” and Hume’s most famous statement on what we would now call aesthetics, his 1757 Essay “Of the Standard of Taste,” which Smith must surely also have known (a point to which I will return).Footnote ¹⁶ Although itself a highly complex philosophical contribution, one way to read the essay on taste is as applying roughly the same strategy as deployed in “A Dialogue.”Footnote ¹⁷

On the one hand, when it comes to questions of what Hume calls “criticism,” all seems ultimately to lie in the eye of the beholder. Insofar as, for example, beauty and quality by definition cannot reside in external objects but only in the minds, and therefore opinions, and therefore tastes, of perceivers, then it would appear that disagreement here simply lies outside the domain of rational disputationFootnote ¹⁸—or as the saying goes, “de gustibus non disputandum est” (“there is no disputing taste”). On the other hand, were someone to suggest (and here I update Hume’s somewhat unhelpful examples) that William Shakespeare’s Hamlet is no better than the sitcom Friends, they would be found to “defend no less an extravagance, than if he had maintained a mole-hill to be as high as Teneriffe.”Footnote ¹⁹ Thus while it is often maintained that there is nothing to say when tastes clash, that disagreement must simply be accepted, this seems incompatible with our widespread conviction that taste can vary in quality, that criticism is not all of the same order, and that the statuses of various things subjected to criticism legitimately differ, with some things (paintings, dramas, poetry, and so on) rightly considered better than others of their kind. But how to explain this conviction, if we are to grant—as Hume, the sentimentalist, does and must—that such qualities lie not in objects themselves, but in the minds of perceivers, and hence ultimately in matters of (nonrational) taste?

Notice here the parallel between claiming that there is no disputing with taste when it comes to judgments of criticism—that is, nothing further can be said beyond reporting that people’s tastes differ, and simply accepting this—and claiming that there is nothing to morality other than custom, vogue, fashion, and law. In both cases, the philosopher can try to show that this is not true, that much more can be said to account for observed differences beyond simply observing that such differences exist. And indeed, this is (at least one) strategy Hume adopts in “Of the Standard of Taste.” He proceeds to demonstrate the importance of factors such as constancy of shared judgments over time, delicacy in relevant faculties of taste, repeat practice and experience in particular domains, deferring to the acknowledged wisdom of others, being free from prejudice and other distorting influences when making judgments, and so forth, when it comes to matters of criticism. There is a great deal more to say about taste, Hume shows, beyond simply observing that tastes often differ. As a result, we can make sense both of the idea that some people’s tastes are more refined than others, but also that some people’s criticisms are more insightful than others, hence why we routinely rank objects of criticism as of better and lower quality—and that to do so is unobjectionable and makes perfect sense. In other words, by tracing matters higher, to “first principles,” the philosopher demonstrates the intelligibility of there being standards of taste, notwithstanding the fact that beauty, quality, and other aspects of criticism are ultimately founded in the subjective tastes of perceivers.

Yet here is where a crucial matter of divergence between “A Dialogue” and “Of the Standard of Taste” arises. While Hume contends that it is manifestly the case that criticism can vary in quality, he also allows that there are circumstances in which the appropriate response when it comes to what we would now call aesthetics is, as the saying goes, to “agree to disagree.” If we control for factors such as delicacy, refinement, freedom from prejudice, experience, and so on, and find that two well-calibrated judges nonetheless still have sincere differing opinions as to the quality of some object of criticism, then it may turn out that there is nothing further to say other than that two well-qualified people disagree. There would then be no final supreme standard one can appeal to in order to decisively claim superiority of judgment.Footnote ²⁰ This will not always be the case, and indeed we will not necessarily know in advance which cases turn out to be like this, although Hume indicates that the different idiosyncratic humors of men, age, cultural differences, and the epoch in which one lives play significant roles. Nonetheless, Hume allows that it manifestly is the case that sometimes disagreement gives out into rival matters of taste, and there is nothing more to be said. And here is where we can mark an important divergence with “A Dialogue”: namely, that in matters of aesthetics, there seems no equivalent to the sort of challenge raised by the figure of Persistent Palamedes. Assuming that we are both competent judges in the ways Hume explicates, above a baseline of plausibility if you and I disagree about the merits of, say, Shakespeare’s comedies versus his tragedies, or the music of Wolfgang Mozart and Ludwig van Beethoven, we may well be perfectly happy—in a way that we won’t when it comes to Hamlet versus Friends—to say that here the appropriate response is for us to agree to disagree. For indeed, at that level there really is no disputing with taste. Yet things seem importantly disanalogous in ethics. Should you claim infanticide to be permissible, perhaps even sometimes required, then even if I were informed that you reason sincerely from principles of usefulness and agreeableness to self and others, at source the very same ones that I use, it seems entirely unacceptable to simply shrug and say, “Oh well, we must agree to disagree—it’s a matter of taste, and there’s no disputing with that!”

What is going on here is that in matters of aesthetic criticism, once we control for quality of judgment, any disagreements that remain simply do not appear sufficiently important for us to be perturbed by them. The stakes just aren’t that high. By contrast, in ethics, there are cases where even if we control for quality of sincere underlying judgment—for example, on the basis of usefulness and agreeableness to self and others—remaining disagreement does seem sufficiently important to be disturbed by it. In fact, more than disturbed. Given what we believe to be right and wrong, it seems required not only that we be perturbed by other people doing something we hold to be immoral, but that we ought to condemn it, and quite possibly condemn them, too. (Note, however, that it is a yet further question what we might practically do in the light of such condemnation: intervene, begrudgingly tolerate, consider whether we should in fact be like them, and so forth; those are precisely further matters.) In other words, in ethics the stakes appear—sometimes at least, as for example in the case of infanticide—dramatically higher than in aesthetics.Footnote ²¹ And because they are sometimes dramatically higher, there are cases in ethics such that even when the philosopher has traced matters to higher first principles, and even if granting that we agree with those principles, we are not content simply to agree to disagree. What oftentimes seems acceptable in the “criticism” of aesthetics is importantly different in the “criticism” of morals, Hume’s running the two together, as he does at the outset of “Of the Standard of Taste”Footnote ²² notwithstanding. In turn, his strategy for constructing a “standard” in matters of aesthetics is not straightforwardly transferable to matters of ethics.

The role of custom in Smith’s moral sentimentalism

With these appraisals of Hume in place, we are now better equipped to turn to Part V of The Theory of Moral Sentiments, noting at least two reasons why Smith is likely to have felt it incumbent upon himself to dedicate an entire part to the influence of custom and fashion upon the moral sentiments. The first is that no matter how sympathetic Smith might have been to Hume’s strategy of tracing matters to higher first principles, he could not endorse Hume’s account. Having attempted to show in Part IV of The Theory of Moral Sentiments that Hume’s utility-orientated sentimentalist account was inadequate and needed to be replaced by his own version, at the very least he would need to supply an alternative explanatory matrix to Hume’s “utility and agreeableness” framework, should he take up the challenge of explaining the role of custom and fashion vis-à-vis the moral sentiments. But why take up that challenge? Because Smith’s own account was vulnerable to precisely the kinds of worries noted above. This can be seen as follows.

At the outset of Part III, Smith makes clear that all moral judgments are ultimately socially composed. While we judge of the propriety of the action of another by imagining ourselves into their situation and seeing whether the sentiments we would feel in such a situation correspond with theirs, this would be impossible had we not already been socialized into prevailing standards of what it is and is not appropriate to feel. The same is true when it comes to assessing our own conduct. When we “divide” ourselves “into two persons” and attempt to assess our own sentiments from the perspective of an impartial spectator, this would be impossible had we not grown up in a society in which we first learned to take on board the imputed estimations of others (TMS III.i.6). In Smith’s evocative phrasing:

As he goes on in the concluding chapter:

Yet if this is the case, the challenge Palamedes raises in “A Dialogue” can be raised against Smith’s position.Footnote ²³ What counts as, say, propriety in ancient Athens appears to differ radically from what counts as such in modern France. How to explain this? On Smith’s account, it must be that the underlying sentiments in each society are different. But how to account for that? One potential answer, given apparently widespread evidence of considerable sentimental disagreement across various times and places, is that at base our moral sentiments are themselves the product of, as Palamedes would put it, “fashion, vogue, custom, and law,” which seems especially likely if (as Smith claims) we only have these sentiments as a result of being socialized amid some particular group of humans, among whom we happen to have been born. But if that is granted, then the original Palamedes challenge bites: “How shall we pretend to fix a standard for judgments of this nature?” (The implication is that there isn’t one.) As we would put it in modern terminology, is Smith not committed, at root, to conceding the relativity of morality, given its allegedly socially composed basis in sentiment?Footnote ²⁴

In sum, from Smith’s perspective, the position taken by the narrator in “A Dialogue” would have been unacceptable insofar as it relied upon Hume’s wider utility-based account of the moral sentiments. Yet this would have mattered to Smith precisely because the challenge raised by Palamedes potentially applied to his own position.Footnote ²⁵ An alternative to Hume’s “A Dialogue” was therefore required. The proper roles of custom and fashion needed to be properly accounted for in a satisfactory theory of moral sentiments. Hence, we have Part V of The Theory of Moral Sentiments.

But furthermore, I suggested above that “A Dialogue” is lacking over and above any technical rejection of Hume’s utility matrix, insofar as an even more serious challenge to ethical sentimentalism can be mounted than the one Hume considers—that which I put into the mouth of Persistent Palamedes, and which applies even if we accept Hume’s appeal to utility or agreeableness to self or others. Hume’s essay shows no awareness of or engagement with this more serious challenge. Yet it extends equally to Smith’s version of ethical sentimentalism. If all moral judgments are ultimately socially composed, doesn’t that mean that they are therefore all relative to the society that composed them? But if so, that seems to impugn the status of moral judgments per se for the reasons articulated above. In what follows, it is therefore worth asking to what extent (if any) Smith was aware of this further challenge, and to what extent (if any) he engaged with it.

The role of Part V in Smith’s The Theory of Moral Sentiments

Smith opens Part V of The Theory of Moral Sentiments as follows:

He thus clearly avows the importance of custom and fashion to the moral sentiments and immediately draws attention to their role in doing damage by causing “irregular and discordant opinions,” the reader being primed for the argument that is to follow. Yet despite this opening paragraph being focused on the moral sentiments, what follows in Chapter 1 of Part V constitutes a discussion of the role of custom and fashion entirely with regard to beauty in (what we now call) aesthetics. By contrast, the discussion with regard to (what we still call) morality is delayed until Chapter 2. Smith therefore makes a move that, as we saw above, Hume does not: separating out for different and distinct treatment questions of aesthetics and ethics. To be sure, Smith retains the language of beauty in both cases, a feature of the fact that he maintains both to be a function of sentiment.Footnote ²⁶ But he clearly indicates, via his separation of topics in Chapters 1 and 2, that the analysis of each needs to be conducted separately. It is often claimed that a clear distinction between aesthetics and morality does not arise in the Western tradition until Immanuel Kant’s Critique of Judgment,Footnote ²⁷ and indeed this may well be true, facilitated in part by Kant’s wholesale removal of ethics from the realm of sentiment. But it should be appreciated that Smith had already begun to take very deliberate steps toward such a separation, and from within a wholly sentimentalist perspective. Did he do this as a response to perceiving the disjuncture between “A Dialogue” and “Of the Standard of Taste” noted above? Here we can only speculate, but it is certainly possible.

Chapter 1 proceeds as follows. Smith begins by considering beauty in questions of human artifice: dress, furniture, poetry, architecture, music. These appear entirely governed by prevailing norms of custom, of which fashion is really just a subspecies.Footnote ²⁸ But what about “the beauty of natural objects,” for example, flowers, horses, and human bodies (TMS V.1.8)? Here, Smith considers at length the suggestion by Jesuit philosopher Claude Buffier that this, too, is the product of only custom and fashion, and hence why different human societies seem to have radically different estimations regarding what counts as beauty. In China, women’s feet are bound into the tiniest dimensions possible; in North America, some tribes use wooden boards to turn skulls into square shapes, aesthetic appearances that in the West are deemed ugly, but in these places are considered beautiful. The reverse, of course, is also true of the practice of squeezing European women into ill-fitting, health-ruining corsets, Smith undercutting any complacent sense of superiority potentially held by his readers. Smith, however, resists Buffier’s claim that aesthetic beauty in natural objects is a product of custom all the way down. Like Hume,Footnote ²⁹ Smith maintains that some aspects of beauty are recommended to us by nature prior to and independent of, custom. These include the utility of particular forms, the immediate agreeableness of certain colors, a preference for the smooth over the rough, for variety rather than tedious uniformity, and particularly for sequences of variety that appear connected to each other in significant ways as opposed to “a disjointed and disorderly assemblage of unconnected objects” (TMS V.1.9).

Nonetheless, Smith concedes that, despite these reservations, when it comes to aesthetic beauty, the roles of custom and fashion appear overwhelmingly important. Though he “cannot admit that custom is the sole principle of beauty,” nonetheless it is so powerful that

When it comes to aesthetics, therefore, custom and fashion appear overwhelmingly to determine what is considered beautiful or ugly, and hence what deserves praise or disapproval. But what about when it comes to morality? Here, things are different.

Chapter 2 of Part V is, even by Smith’s standards, dense. This is easily underappreciated, however, so it is worth carefully examining how he proceeds. Smith opens by first conceding that because morality is a function of sentiment, and therefore rightly assessed in terms of beauty, it must to some extent be influenced by custom and fashion, so extensive are the power of these (as just established in Chapter 1). And yet, “[t]heir influence here, however, seems to be much less than it is every where else” (TMS V.2.1). When it comes to aesthetics, there is perhaps nothing so ridiculous that custom cannot reconcile us to it, fashion perhaps even render agreeable. But when it comes to morals—to “character and conduct,” for example, that of Nero and Claudius—some things lie forever beyond the pale, regardless of custom and fashion. Why? Because there is a category difference between aesthetic sentiments and moral sentiments:

We will need to return to Smith’s use of the language of morality as being something that can be “warpt” but not “perverted.” (On what basis can he even intelligibly talk of morality in such terms?) But for now, let us consider the argument Smith proceeds to make in the rest of Chapter 2, composed of the following strategies.

First, Smith concedes that in those cases where custom and fashion do have an impact on the moral sentiments, although this is never as pronounced as it is in cases of aesthetics, the mechanics of how things work are essentially the same. “When custom and fashion coincide with the natural principles of right and wrong, they heighten the delicacy of our sentiments, and increase our abhorrence for every thing which approaches to evil,” with the reverse also being true—again leaving aside, for now, Smith’s striking use of the language of “natural principles of right and wrong” (TMS V.2.2). Those who are well brought up amid justice, humanity, and modesty, are especially shocked when these virtues are absent or violated, whereas those brought up in harsh conditions are rendered by customary habits far less shocked by acts that are properly considered of enormity and rightly deserving punishment. Similarly, under the rein of Charles II licentiousness became fashionable, insofar as it served to mark gentlemen out as not being puritans, a function of the context of the era.

Having conceded that custom and fashion can influence moral sentiments, Smith next however goes on to replicate a strategy found in “A Dialogue”: that the philosopher can explain many of the ways in which custom and fashion work their influence beyond simply stating that morality is a function of them, with no further elaboration (the original Palamedes challenge). Where Smith differs from Hume is that he makes no appeal to a single higher “first principle” that allegedly explains all downstream moral difference, and certainly not any appeal to utility and agreeableness. Nonetheless, he contends that there are multiple factors we can account for regarding why morality differs in different times and places. First, different professions will inculcate different characters in people. We expect clergymen to be sober and serious as befitting their spiritual burdens, but soldiers (a little paradoxically) to be gay and lively, a function of their needing to cope with the psychological stress of military service. Similarly, we expect the old and the young to display different characters (and for our own characters to change over time as we age) because the different stations of life properly influence the way we act and feel. Likewise, the circumstances that different societies find themselves in, given varying historical contexts as well as differing economic and political situations, ensure that which sentiments are deemed appropriate must fluctuate to some degree as per local conditions. In Russia, the generally tougher circumstances mean that a moderate degree of politeness would be interpreted as intolerable effeminacy, whereas in France, where things are much more luxurious, the same degree of politeness would be received as barbarity. Generally speaking, Smith maintains that it is inevitable that in “barbarous” conditions the harshness of life ensures that virtues of self-denial are more highly praised than those of humanity, whereas in more “civilized” nations the reverse will be true, a function of how the moral sentiments have to adapt to the realities of different socioeconomic contexts.

A great deal more could be said about the (often intricate) details of Smith’s examples, all aiming to examine in greater depth the influence of custom and fashion upon the moral sentiments in different times and places. But for present purposes it is most fruitful to focus on the conclusion that Smith aims to draw from the many examples of moral divergence he spends the bulk of Part V discussing: namely, despite the fact that there are many differences that are a function of custom and fashion, the vast majority of these do not matter. While it may be true that barbarous societies are deficient regarding the humane virtues, they are not incapable of them. Likewise, polished civilized societies may struggle with the virtues of self-denial, but they can acquire them, for example, in times of military conflict. While we expect the old and the young to conduct themselves differently, this applies only to what are ultimately trivial matters. By contrast, we expect the rules of “truth and justice” to apply equally to both, as indeed we do when it comes to soldiers and clergymen (TMS V.2.13). In sum, Smith adopts a different central strategy from Hume. Rather than trying to trace observable differences to higher first principles, Smith mounts what is ultimately an empirical claim: insofar as custom and fashion produce moral differences, these are all minor, and need not trouble us:

We might term this strategy of Smith’s an attempted deflation of the challenge posed by moral relativism. Hume’s narrator in “A Dialogue” engages in this strategy, too, but makes it subservient to the agreeableness and utility framework. By contrast, Smith makes it the linchpin of his argument.

Smith, however, cannot stop matters there. For an evident reply can be made that he himself acknowledges: it is all very good appealing to the examples that he does, but there are surely others that plainly tell against him. The following, surely, is a difference in morals that is anything but trivial: the ancients routinely practiced infanticide, as do “all savage nations” in his own time, and both consider it entirely morally permitted (in the case of the Athenians, for no better reason than to avoid inconvenience). For modern Europeans, by contrast, there can be no “greater barbarity … than to hurt an infant” (TMS V.2.15). When moral sentiments can differ that much, isn’t Smith’s suggestion straightforwardly refuted?Footnote ³⁰

Yet Smith still has cards left to play. The most important of these is an alleged distinction between “the general style of conduct or behaviour” and that “[w]ith regard to particular usages.” Concerning the former, Smith reasserts his empirical claim. When it comes to that which is “general,” custom and fashion do not authorize important departures from what all epochs and societies agree is fundamentally morally acceptable. By contrast, when it comes to the “particular,” here is a realm in which custom and fashion’s “influence is often much more destructive of good morals, and it is capable of establishing, as lawful and blameless, particular actions, which shock the plainest principles of right and wrong” (TMS V.2.14). Infanticide, Smith declares, exists in the latter category of cases.

As it stands, however, this distinction does not get Smith very far. A critic might easily reply: “Well, so what? What you call ‘particular’ cases appear awfully important. Indeed, they appear to arouse precisely the kind of worries that you have been trying to deflate.” Hence, Smith is required to make further moves to defend his position. These go as follows. First, he offers another variant on the deflation strategy. Infanticide must have arisen, he claims, in the most barbarous periods of society. It could only have originated as a response to situations of extreme scarcity and danger, in which keeping an infant alive would merely condemn both the child and its parent to death: “One who, in flying from an enemy, whom it was impossible to resist, should throw down his infant, because it retarded his flight, would surely be excusable; since, by attempting to save it, he could only hope for the consolation of dying with it” (TMS V.2.15). Hence, it was only in conditions of extreme savagery that the practice of exposure could have arisen; only the most desperate of external circumstances could have overpowered the natural tendency to love and care for children universally instilled in human nature. But if so, there are two things to note. One implication of Smith’s position is that exposure of infants is not something we need to be worried about in modern conditions, because our civilized situation is such that circumstances will never overwhelm our natural principles of humanity and love for infants. The second implication is that Smith is warning his readers not to be too hasty in judging those in circumstances so perilous they feel the need to expose their young. If such circumstances really do obtain, then infanticide is “undoubtedly more pardonable than in any other.” Although Smith does not himself spell out the further implication, we might well go that extra step on his behalf. Notice that if you were living in such savage circumstances, then you would think that exposure of infants was permissible, too (TMS V.2.15). Killing children only (approaches) permissibility in the most extreme of circumstances, which we moderns will never find ourselves in, so there is less to be alarmed about than first appears. Furthermore, we can recognize even by our own standards that infanticide, at least in certain circumstances, is not as morally outrageous as it seems prior to careful reflection. There is less moral disparity, even in this case, than there may have seemed.

This is all well and good, except that it fails to account for the stand-out example: Athens, at the height of its civilization in the ancient world, where citizens not only practiced exposure, but apparently did so for reasons as trivial as avoiding inconvenience. That was not a savage society subsisting in extreme circumstances, so Smith must make yet another move. The problem, he claims, is that after the establishment of exposure among the primordial Greeks, “uninterrupted custom” saw this practice passed down as “authorized” to societies even as advanced as Athens. Indeed, the power of custom in this regard was so extreme that it shaped not just the “loose maxims” of ordinary people, but also the conclusions of philosophers as great as Plato and Aristotle. By their own moral standards, however, the ancient Athenians—or at the very least, the reflective, philosophically minded among them—ought to have seen that the practice of exposure was no longer authorized, but a relic of customs passed down from a previous primordial age. Instead, even philosophers appealed to “far-fetched considerations of public utility” (TMS V.2.15) as post hoc rationalizations for a “horrible practice” (TMS V.2.16) that they should have recognized as such for themselves.Footnote ³¹ But such was the power of custom that when it came to infanticide, it could authorize an immoral phenomenon even among a sophisticated people.

This seems to serve Smith’s deflationary purposes well. It turns out the ancient Athenians were not all that different from us; they were just led astray by inherited customs. Had they reasoned correctly and by their own lights, they would have condemned the practice of exposure on the same grounds that we do. Unfortunately, however, this in fact lands Smith straight back into the problem noted above. As he himself puts it: “When custom can give sanction to so dreadful a violation of humanity, we may well imagine that there is scarce any particular practice so gross which it cannot authorise” (TMS V.2.15). And so here, Smith plays his final card. While it is sadly true that with regard to specific practices, custom may authorize terrible departures from what all would otherwise consider to be morally acceptable—see ancient Athenian infanticide—these are the exceptions, not the rule. When it comes to “the general style and character of conduct and behaviour,” custom can never “pervert our sentiments” to the “same degree” regarding “particular usages.” Why? Because “[t]here never can be any such custom. No society could subsist a moment, in which the usual strain of men’s conduct and behaviour was of a piece with the horrible practice I have just now mentioned” (TMS V.2.16). So ends Part V.

Smith’s failure to meet the challenge of moral relativism

Part V is thus a dense and sophisticated component of The Theory of Moral Sentiments. As I will begin to explore in this section, however, it is also a candidate for the most puzzling, and indeed ultimately unsuccessful, aspect of Smith’s moral oeuvre. To see why, we can begin where Smith ends: his strikingly underdeveloped claim that the style and character of conduct and behavior can never severely depart from core, widely agreed-upon, human universals in the same way that “particular usages” can.

What Smith appears to have in mind is the same basic argument developed in Part II regarding justice. Given baseline facts about human nature and the conditions therefore required for humans to live successfully in groups, limits will inevitably be placed on how they can do soFootnote ³²: “If there is any society among robbers and murderers, they must at least, according to the trite observation, abstain from robbing and murdering one another” (TMS II.ii.3.3). In other words, it is a fact about human beings that any society will need to have some conception of justice organizing its internal relations, if it is to continue as a society. While the exact rules of justice may differ according to circumstances, they can only differ so much before society falls apart, the rules thus failing as rules of justice. At the close of Part V, despite being notably under-spoken about the details, Smith appears to be generalizing this strategy to something like morality tout court. Given baseline facts of human nature and our need to live successfully in groups over time, there are necessarily going to be limits placed on what can be established regarding permissibility of moral conduct. Hence, although it might be possible in certain circumstances for a subset of innocents within a society to be permissibly (or “permissibly”) killed according to local standards—for example, (some) babies in savage conditions (or in ancient Athens)—it will never be possible for a society to declare that all and any innocents may permissibly be killed, because to adopt such a general principle would lead to the destruction of any such society itself. Hence, hard limits are imposed upon the extent to which moral systems in different times and places can vary, given underlying facts about human nature and the stability requirements of any society able to persist over time.

A particular advantage of reading Smith this way is that it makes much less puzzling his otherwise striking use of language. As noted above, Smith discusses morality as something capable of being warped or perverted, likewise invoking “natural principles” of ethics, while condemning infanticide as a “horrible practice” in a tone that appears independent of the judgment being somehow relative to the context of any given society. Yet how is Smith entitled to speak in such terms? Given his claim in Part III that moral distinctions are inherently socially composed, is he not straightforwardly contradicting himself by adopting such apparently nonrelativistic terminology? By what right, if any, can he use such language? The trick here is to avoid the natural temptation (at least, natural to readers today) of interpreting Smith’s language as a species of objectivism, and specifically of what we would now call moral realism, that is, somehow claiming that moral principles exist prior to and/or independent of the mental states (the moral sentiments) of moral practitioners (that they are, to use modern terminology, stance-independent). If this were what Smith was trying to claim, then he would indeed be using language to which he is not entitled, given Part III, and which would generate self-contradiction. Instead, I take it that Smith’s use of language presupposes the correctness of his own wider theoretical commitments. We must take as given Smith’s contention that human nature, combined with the practical limits imposed upon all successful societies that persist over time, generates upper bounds regarding the extent to which moral practices can differ across different times and places. If that is true, then it is entirely comprehensible to talk, for example, of general moral principles as capable of being warped, but never entirely perverted, that is, that there is only so much alteration they can exhibit. Likewise, that there are “natural principles” of morality is simply a function of baseline facts about how human nature necessarily is, plus the practical realities of social living in groups over time, but neither presupposing nor being committed to any metaethical principles one way or the other. Similarly, to speak of infanticide as a horrible practice, and in unqualified terms, is unproblematic if the default assumption is that all normal human beings—that is, excluding those corrupted by local customs—will simply agree on this due to the very fact that they are human. Smith need neither invoke nor surreptitiously help himself to any metaethical principles incompatible with the position taken in Part III (or indeed, any at all). He need only claim that universal facts of human nature prescribe the limits of how much morality can in practice differ. As striking as Smith’s language in these regards may at first appear, this is a puzzle that can be resolved in his favor.

Smith’s account, however, faces much more serious difficulties. As we have seen, Smith bets the house on what is ultimately an empirical claim: that the vast majority of differences in observable moral practices fall into the category of not mattering, and any that do matter belong to a particular subset of cases that can be bracketed-out, and hence need not perturb us. Of course, Smith adds a further layer of complexity, purporting to be able to explain this alleged empirical uniformity as being due to underlying facts about human nature plus the constraints imposed upon any society able to persist over time. But be that as it may, all ultimately stands and falls with Smith’s insistence that the overwhelming majority of observable human moral differences are not sufficiently significant to matter. Yet here, Smith is on shaky ground.

To twenty-first-century readers this will seem particularly evident. Given now-widespread knowledge of profoundly incompatible practices such as, for example, female genital mutilation versus Western standards of gender equality, or of state-sanctioned persecution of homosexuality (leaving aside for now the fact that not so long ago Western societies engaged extensively in such behavior), Smith’s gambit appears hopeless. The empirical record tells firmly against the claim that the vast majority of moral differences do not matter. Abundant examples indicate that it very often does—and in significant ways. Furthermore, even if we grant Smith’s supposition that the facts of human nature, plus the constraints of social living for creatures like us, necessitate that certain things can never be instantiated as morally acceptable in any feasible human context, this falls well short of lending support to the claim that the vast majority of moral differences do not matter. Whatever the facts of human nature might dictate regarding the limits of moral difference, these are just not substantive enough to guarantee a sufficiently high degree of convergence across epochs and cultures to allay the concerns generated by observable cases, as the empirical record itself proves. Upon closer inspection, Smith’s empirical gambit, and his suggestion that human nature sets limits to how much moral difference is possible for the species, are not mutually supportive in the manner he supposes.

In response to this, it might appear only fair to excuse Smith on grounds of historical ignorance, to plead that he could not know then what we know now about the empirical extent of severe moral difference. Yet this lets Smith off the hook too easily. Given the wealth of traveler reports that had been flooding into Western Europe throughout the seventeenth and eighteenth centuries, many of which claimed astonishing levels of moral difference, and thus incompatibility, found in places such as North America—dubious tales of which Smith himself repeatedly appeals to in Part V—it seems highly unlikely that Smith could have expected his contemporaneous audience to react with easy acceptance. Even if granting that many traveler reports were unreliable and exaggerated, the picture painted by European societies’ encounters with other cultures, by the time that Smith was writing, was not one of convergence in moral affairs, but of profound divergence. Footnote ³³ Furthermore, Smith’s readership hardly needed to resort to traveler’s reports of varying quality to find his empirical claim suspect. Their own Bibles contained Old Testament evidence of extensive polygamy in ancient Hebraic societies, a significant moral difference with the present, and which was at the forefront of ethical debate in the eighteenth century.Footnote ³⁴ Similarly, infanticide was hardly the only practice conducted by the ancient Athenians that was deeply shocking. Although it might be too impolite to openly discuss it in print, eighteenth-century readers were well aware that the Athenians practiced institutionalized pederasty, and considered this unproblematic, at least in various circumstances (they, too, read their Plato). This, if anything, was even worse than infanticide. In other words, it is not at all clear that readers in the eighteenth century would have found Smith’s empirical gambit any more persuasive than it appears today.Footnote ³⁵

Similarly, Smith’s attempt to shore up his gambit by appealing to a distinction between general moral practices and “particular usages” appears implausible when subjected to scrutiny. Infanticide, after all, was not somehow insulated from wider Athenian beliefs about the family, the nature of death, the duties of parenting, civic responsibility, and so on. In other words, there simply is no bright-line distinction between general moral practices on the one hand and “particular usages” on the other. The latter are derived from, indelibly conditioned by, and bound up with the former. Smith’s distinction here looks suspiciously like it is invoked not because it tracks how things are, but because it would be very helpful for Smith’s deflationary strategy if it were so—which rather implies it is damned twice over. Furthermore, there is the point touched on above that even were Smith able to separate out general practices and particular usages in a principled and consistent manner, the empirical record would appear to show that even if we somehow grant that “general practices” are hemmed in by the constraints of human nature, this just is not enough to deflate relativistic worries generated by the remaining “particular usages.” The extent of severe moral differences observable between “particular usages”—just think of infanticide!—across the species and throughout history, mean that such usages are too significant a feature of social reality to be safely boxed off and put aside in the way Smith proposes. They are rather precisely the problem on many occasions, not merely a regrettable subissue we can safely put to one side. What does it matter to find out that the Athenians perhaps shared with us certain general structural features regarding their beliefs in justice, prudence, beneficence, and so forth, once we find out they thought it acceptable to kill babies?

Part V is thus deeply puzzling. It is not only that Smith’s empirical strategy appears unsuccessful, given its implausibility. It is that its implausibility appears so evident (then as well as now), such that one struggles to understand how a thinker as sophisticated and accomplished as Smith could have stood by it. And yet he did, across six editions and thirty-one years, subjecting this part of the work to only the most inconsequential of later editorial changes. Smith appears to have remained happy with the position he took in 1759, feeling no need to revise it (unlike many other aspects of The Theory of Moral Sentiments) for the rest of his life. Accounting for why constitutes, I suggest, an outstanding puzzle of Part V.

Concluding thoughts

It was noted above that it is worth considering the extent to which Smith was more sensitive than Hume to the destabilizing worries raised by relativism, which I put into the mouth of Persistent Palamedes. In particular, that mere awareness of moral difference between different cultures and epochs can have the effect of destabilizing one’s own moral commitments; moral difference can lead to awareness of moral contingency, which appears to threaten moral arbitrariness. By way of drawing matters to a conclusion, we may ask: How does Smith fare on this score?

No better, or so it would seem, than Hume. Smith does not explicitly discuss this kind of existential challenge, nor how it might be alleviated, and yet he comes extremely close to being required to. Recall his discussion of how, in sufficiently savage conditions, the exposure of infants was “undoubtably more pardonable than in any other” (TMS V.2.15). As noted above, however, the logic of Smith’s position seems to generate a different and more unsettling conclusion: In sufficiently harsh circumstances infanticide is permissible, and we would agree with this should we ever find ourselves in the same unfortunate position as such “savage” peoples. But if this conclusion is granted, then it immediately engages the anxieties of Persistent Palamedes: that our moral principles are a function of social context, and therefore of accident of birth, and that infanticide is not somehow wrong simpliciter, but only in sufficiently advantageous circumstances. But if that is granted, then how can we say, with any further confidence, that it is wrong? Destabilizing skepticism threatens.

At least two possibilities might be at play here. The first is that Smith simply failed to see the full implications of his position, thus explaining why he did not in turn address the sorts of existential concerns that his argument regarding infanticide generates. Yet this explanation is unsatisfying insofar as it relies on Smith exhibiting a theoretical myopia that appears inconsistent with his otherwise clear and consistent level of philosophical genius. Hence, a second possibility is that Smith was aware of the relativistic logic of his own position, but shied away from stating it outright, leaving this matter for the astute reader to discern themselves. Why might he have done this? One possibility is that while Smith may have been aware of the kind of challenge posed by Persistent Palamedes, he did not know how to answer it, and so opted to leave that can of worms unopened. A second, related possibility is that Smith feared that openly admitting the (ultimately) relativistic underpinnings of his argument would seriously undermine the attempt at deflation that Part V otherwise constitutes. Rather than allaying a reader’s worries about the roles of custom and fashion, these might be seriously exacerbated, should Smith in the end openly concede that moral judgments are irreducibly context-dependent. Hence Smith perhaps deliberately stopped short of doing so outright, for broadly prudential reasons. “Smith, unlike Hume, did not have the temperament of an iconoclast,” after all, and potential caution on his behalf fits with what we know about his wider aversion to courting controversy.Footnote ³⁶ On the other hand, postulating that Smith hid what he truly thought from sight, deliberately presenting an incomplete version of his own argument in order to make it more palatable to the average reader, invites an interpretative approach regarding which extreme caution should be exercised.Footnote ³⁷ In any case, what exactly Smith is attempting at the close of Part V remains, once again, puzzling.

This essay has attempted to address the scholarly lacuna currently surrounding Part V of The Theory of Moral Sentiments by placing Smith’s discussion in the context of a complex intellectual inheritance bestowed by Hume. Despite occupying a relatively short portion of The Theory of Moral Sentiments (a mere eighteen pages in the Glasgow edition), Part V is a dense and sophisticated intervention. And yet, as I have also tried to show, it is a candidate for the least successful as well as most puzzling feature of that work.

Why does this matter? If the above readings are correct, then one consequence is that we must accept that neither Hume nor Smith offered satisfactory engagements with the problem posed by moral relativism. Despite remaining the two preeminent theorists in the tradition of ethical sentimentalism, both fall short on this matter. Smith, it seems, keenly felt the need to improve on Hume, but ultimately did not fare well on this score, his steps toward facilitating a conceptual distinction between ethics and aesthetics notwithstanding. The wider lesson is that sometimes even world-historical geniuses, who knew each other’s work and confronted the same problems, did not get it right. More time, and future perspectives, can be required for progress.Footnote ³⁸

Be that as it may, Part V has lain unduly neglected for too long. This essay seeks to have shown that this situation is unsatisfactory, while also attempting to change it. Indeed, even if every aspect of the reading offered above is ultimately rejected, insofar as this encourages closer engagement with Part V and with the underlying challenges to ethical sentimentalism that moral relativism poses, such rejection is only to be welcomed.