Introduction
After reading the first edition of Adam Smith's Theory of Moral Sentiments (Smith Reference Smith, Raphael and Macfie1976, hereinafter: TMS) David Hume wrote to Smith and pointed out what he took to be a major shortcoming of Smith's account of sympathy in TMS. The section of the letter in which Hume takes up this issue is remarkably dense. By my count, he makes four philosophical objections in six fairly brief sentences. Examining each of those objections in turn and considering what resources exist in Smith's system to meet them proves to be an excellent ‘stress test’ for this core piece of Smith's sentiment-based moral theory. That being said, this is not a discussion of Smith's ethical theory per se. Smith and Hume were both advocates of applying, ‘the experimental method of reasoning [to] moral subjects’ (cf. the subtitle of Hume's Treatise), and both conceived of their dispute over the nature of sympathy as an empirical matter (for commentary on this point, see Alexander Broadie [Reference Broadie and Haakonssen2006: 170–74]). In particular, Hume's claim is that Smith has got the psychology of sympathy wrong. In this essay, I hope to settle that debate in Smith's favor. Accordingly, this is a project in ethics only in the sense Elizabeth Anscombe had in mind when she wrote, ‘it is not profitable for us at present to do moral philosophy; that should be laid aside at any rate until we have an adequate philosophy of psychology, in which we are conspicuously lacking’ (Reference Anscombe1958: 1). It is my hope that giving a proper defense of Smith's account of sympathy can be a first step toward a philosophy of psychology able to underpin a version of sentiment-based ethical theory that can withstand the traditional criticism that such systems have inadequate resources to guard against bias.
1. Smith's Account of Sympathy
For Smith, sympathizing with someone crucially involves an imaginative exercise in which we attempt to discern what we would feel if we were in that person's predicament. Here is how Smith introduces the notion on the first page of The Theory of Moral Sentiments in the chapter entitled ‘Of Sympathy’:
By the imagination we place ourselves in his situation, we conceive ourselves enduring all the same torments, we enter as it were into his body, and become in some measure the same person with him, and thence form some idea of his sensations, and even feel something which, though weaker in degree, is not altogether unlike them. His agonies, when they are thus brought home to ourselves, when we have thus adopted and made them our own, begin at last to affect us, and we then tremble and shudder at the thought of what he feels. For as to be in pain or distress of any kind excites the most excessive sorrow, so to conceive or to imagine that we are in it, excites some degree of the same emotion, in proportion to the vivacity or dulness of the conception. (TMS I.i.1.2: 9)
So sympathizing is experiencing an emotion as a matter of ‘fellow-feeling’ as a result of ‘changing places in fancy’ with another person (TMS I.i.1.3: 10). The adverbial expression can be useful here: to sympathize, in Smith's sense, is to grieve or rejoice or be angry sympathetically. As it is a process through which we experience such emotions sympathetically, Smithean sympathy is not itself an emotion (for discussion of Smith's differentiation between his technical sense of ‘sympathy’ and the nontechnical sense in which it does denote an emotion, see Charles Griswold [Reference Griswold1999: 78–83] and Geoffrey Sayre-McCord [Reference Sayre-McCord and Schliesser2015: 211–12]).
As the block quotation above focuses on unpleasant sentiments, such as pain and sorrow, it is important to note that Smith thinks we can, and often do, have sympathetic experiences of happier emotions as well:
Whatever is the passion which arises from any object in the person principally concerned, an analogous emotion springs up, at the thought of his situation, in the breast of every attentive spectator. Our joy for the deliverance of those heroes of tragedy or romance who interest us, is as sincere as our grief for their distress, and our fellow-feeling with their misery is not more real than that with their happiness. We enter into their gratitude towards those faithful friends who did not desert them in their difficulties; and we heartily go along with their resentment against those perfidious traitors who injured, abandoned, or deceived them. In every passion of which the mind of man is susceptible, the emotions of the by-stander always correspond to what, by bringing the case home to himself, he imagines should be the sentiments of the sufferer. (TMS I.i.1.4: 10)
As we can see, Smith uses ‘sympathy’ as a technical term referring to a psychological process through which we have affective responses upon conceiving the situations of others. As the above examples of characters in theatrical productions suggest, the ‘person’ with whom we sympathize in this way need not be someone particularly well known to us. Perhaps somewhat surprisingly, it is also a feature of Smith's account that it is possible to sympathize with someone who does not, in fact, actually experience the emotion in question. Smith discusses three examples of this sort. It will be useful to look at each of them in turn and in some detail: both because the examples draw out important features of Smith's account of sympathy and because they will help us differentiate Smith and Hume on this issue when the time comes.
First, we have a situation in which a person is so socially unaware as to not notice that he or she is being rude. Smith takes it to be an important empirical datum that the other person’s lack of awareness does not lessen our sense of embarrassment on his or her behalf, but instead increases it:
We blush for the impudence and rudeness of another, though he himself appears to have no sense of the impropriety of his own behaviour; because we cannot help feeling with what confusion we ourselves should be covered, had we behaved in so absurd a manner. (TMS I.i.1.10: 12)
One of the reasons Smith thinks this sort of case is so important is that it counts as evidence that the emotions we feel in sympathy with others are largely the product of the imaginative exercise of thinking of how we would react if we were in that person's situation.
Smith then considers a second sort of case in which a spectator might sympathize with a sentiment that the person in question does not herself experience: that is, where mental illness is a factor. His example involves a person whose disease keeps him from noticing the direness of his condition and who even ‘laughs and sings perhaps’. Here again, it seems like an exercise of the imagination must be a crucial catalyst in our having a sympathetic emotional response:
The anguish which humanity feels, therefore, at the sight of such an object, cannot be the reflection of any sentiment of the sufferer. The compassion of the spectator must arise altogether from the consideration of what he himself would feel if he was reduced to the same unhappy situation, and, what perhaps is impossible, was at the same time able to regard it with his present reason and judgment. (TMS I.i.1.11: 12)
Smith takes this as further empirical evidence that imagination, rather than perception, is the key driver of sympathy.
The third category of this type is constituted by sympathetic emotions we feel when we consider the plight of the deceased. Smith begins by noting the peculiar mismatch between our sympathetic sense of loss of the experiences the person could have had if still alive, on the one hand, and the fact that the person can, of course, feel no sense of loss (nor anything else, for that matter), on the other hand (TMS I.i.1.13: 12). Smith then directs the reader's attention to an importantly related issue by noting that it makes us unhappy that our sympathy is impotent to comfort the deceased: ‘That our sympathy can afford them no consolation seems to be an addition to their calamity; and to think that all we can do is unavailing, and that, what alleviates all other distress, the regret, the love, and the lamentations of their friends, can yield no comfort to them, serves only to exasperate our sense of their misery’ (TMS I.i.1.13: 13).
It is significant that it bothers us that our sympathetic reactions provide no comfort to the dead. Smith's observation that we are troubled by this fact provides a key clue to one of the distinctive roles he supposes sympathy plays in our psychology and our social interactions. The unpleasantness associated with what we might call ‘unrequited sympathy’ has a powerful motivational force. Smith emphasizes the symmetry between our desire for others to sympathize with us and our wanting to sympathize with others:
As the person who is principally interested in any event is pleased with our sympathy and hurt by the want of it, so we, too, seem to be pleased when we are able to sympathize with him, and to be hurt when we are unable to do so. We run not only to congratulate the successful, but to condole with the afflicted; and the pleasure which we find in the conversation of one whom in all the passions of his heart we can entirely sympathize with, seems to do more than compensate the painfulness of that sorrow with which the view of his situation affects us. On the contrary, it is always disagreeable to feel that we cannot sympathize with him, and instead of being pleased with this exemption from sympathetic pain, it hurts us to find that we cannot share in his uneasiness. (TMS I.i.2.6: 15–16)
As before, for Smith this is a matter for which we should turn to empirical observation for confirmation. He thinks that a little reflection on one's own case will make it clear that we want to sympathize with others, and when such a desire for a sympathetic connection is thwarted, we do not like it.
In summary, we can see that Smithean sympathy is a psychological process through which we come to experience a version of the same emotion that we conceive another person as having. Very often, though by no means always, we have such a sympathetic emotion in the physical presence of the person with whom we are sympathizing. Importantly, although it is closely associated with three sorts of affective states, Smithean sympathy is not itself an emotion.
In order to get clear on the relationship between sympathy, in Smith's sense, and emotions, it will be useful to set out those three types of associated affective states. (Vivienne Brown argues that these three reactive states constitute Smith's ‘tripartite theory of moral judgment’ [Reference Brown, Fricke and Føllesdal2013: 256].) The first is the emotion we have as a matter of ‘fellow-feeling’: that is, when we feel an emotion in sympathy with another person (as we will see, in his objection, Hume labels this affective state that comes from fellow-feeling the ‘Sympathetic Passion’). The second type is the sense of satisfaction we have when, in the course of experiencing a sympathetic emotion, we notice the complete identity between the other person's emotion and our own. (In his response to Hume's objection, Smith calls this affective state that comes from noticing the agreement between the passions in the original person and the spectator who is observing her the ‘sentiment of approbation’.) The third type is closely related to the second and, depending on the details of the particular instance, can go one of two ways: that is, it is the unpleasant sensation we have when we realize that we are either (a) unable to sympathize with another person's emotional reaction to her or his situation, or (b) by an act of the imagination we do have an emotion in sympathy with the person, but we realize that our reaction is incapable of comforting that person in any way.
With this overview of Smith's account of sympathy in place, we are ready to introduce Hume's objections. The upcoming discussion will refer to both the overview and to the explication of the affective states that are related to Smithean sympathy.
2. Hume's Objections
Upon learning that Smith was preparing a second edition of TMS, Hume wrote the following in a letter to his friend Smith:
I wish you had more particularly and fully prov'd, that all kinds of Sympathy are necessarily Agreeable. This is the Hinge of your System, and yet you only mention the Matter cursorily in p. 20. Now it would appear that there is a disagreeable Sympathy, as well as an agreeable: And indeed, as the Sympathetic Passion is a reflex Image of the principal, it must partake of its Qualities, and be painful where that is so. Indeed, when we converse with a man with whom we can entirely sympathize, that is, where there is a warm and intimate Friendship, the cordial openness of such a Commerce overpowers the Pain of a disagreeable Sympathy, and renders the whole Movement agreeable. But in ordinary Cases, this cannot have place. An ill-humord Fellow; a man tir'd and disgusted with every thing, always ennuié; sickly, complaining, embarrass'd; such a one throws an evident Damp on Company, which I suppose wou'd be accounted for by Sympathy; and yet is disagreeable.
It is always thought a difficult Problem to account for the Pleasure, received from the Tears and Grief and Sympathy of Tragedy; which woud not be the Case, if all Sympathy was agreeable. An Hospital woud be a more entertaining Place than a Ball. I am afraid that in p. 99 and 111 this Proposition has escapd you, or rather is interwove with your Reasonings in that place. You say expressly, it is painful to go along with Grief and we always enter into it with Reluctance. It will probably be requisite for you to modify or explain this Sentiment, and reconcile it to your System. (Smith Reference Smith, Mossner and Ross1987: 43, Hume's letter to Smith of 28 July 1759; hereinafter: Corr.)
Previous commentators have generally discussed Hume's letter as presenting a single objection to Smith's account of sympathy. I think it helps bring some important issues into sharp relief if we distinguish the following four objections: First, Hume gives the example of an ‘ill-humord’ man who ‘throws an evident Damp on company’. He takes this to be a counterexample to Smith's putative claim that ‘all kinds of sympathy are necessarily agreeable’. Second, as a second kind of counterexample, he notes the traditional puzzle that people enjoy tragedies and melodramas. Third, Hume argues that if Smith's account of sympathy were correct, then ‘An Hospital woud be a more entertaining Place than a Ball’. Fourth, finally, Hume accuses Smith of ignoring the differences between sympathizing with a friend and sympathizing with someone who is not a friend. I shall now discuss and offer responses to each of these objections in turn.
3. The Responses to Hume's Objections
We have no record of a direct response from Smith to Hume. The two pieces of text to which scholars typically refer are a footnote to the second edition of TMS and a letter, not to Hume, but to Gilbert Elliot. This means that there is no decisive textual evidence indicating just what Smith thought of the various objections raised in Hume's letter. The one place Smith explicitly mentions Hume's objection is the letter to Elliot, in which he writes: ‘You will find too in the Papers I have sent you an answer to an objection of D. Humes. I think I have entirely discomfited him’ (Corr.: 49, Smith's letter to Gilbert Elliot of 10 October 1759). We will see shortly that several commentators think Smith is wrong about this, and one in particular goes so far as to claim, ‘I believe Smith's claim to have ‘entirely discomfited’ him is almost certainly false’ (Broadie Reference Broadie and Haakonssen2006: 173). In parts 3.1 through 3.4 of this essay I will seek to answer the question of whether Smith has reason to think he can reply to Hume's objections. But as these two sentences are all Smith says about that matter in his letter to Elliot, we shall now turn to the footnote:
It has been objected to me that as I found the sentiment of approbation, which is always agreeable, upon sympathy, it is inconsistent with my system to admit any disagreeable sympathy. I answer, that in the sentiment of approbation there are two things to be taken notice of; first, the sympathetic passion of the spectator; and, secondly, the emotion which arises from his observing the perfect coincidence between this sympathetic passion in himself, and the original passion in the person principally concerned. This last emotion, in which the sentiment of approbation properly consists, is always agreeable and delightful. The other may either be agreeable or disagreeable, according to the nature of the original passion, whose features it must always, in some measure, retain. (TMS I.iii.1.9: 46, note)
This note is appended to the sentence, ‘[It] is painful to go along with grief, and we always enter into it with reluctance’ (TMS I.iii.I.9: 46), which is the line Hume highlighted in his letter. It would appear, then, that Smith thought Hume had simply overlooked the distinction between ‘the sympathetic passion’ and ‘the sentiment of approbation’.
What is this ‘sentiment of approbation’, and how is it, unlike ‘the sympathetic passion of the spectator’, supposed to be agreeable in each and every one of its occurrences?Footnote 1 In the first instance, the sentiment of approbation is the pleasant feeling agent A gets when she notices that she is experiencing the very same emotion as agent B. A paradigmatic case would be if A, after the imaginative exercise of putting herself in B's situation, sympathetically feels the same emotion B feels and then notices the concurrence between B’s emotional state and her own. A sentiment of approbation is an instance of interpersonal emotional connectedness together with a kind of affective agreement.
For Smith, enjoying this kind of emotional accord, and being uncomfortable when there is discord, are fundamental parts of the human condition. These reactions appear in all sorts of situations, including seemingly trivial ones far removed from the ones fraught with moral weight that will dominate much of the ensuing discussion. We are simply wired up, as the saying goes, to notice and appreciate the levels of consonance and dissonance between our own emotions and those of the people around us:
A man is mortified when, after having endeavored to divert the company, he looks round and sees that nobody laughs at his jests but himself. On the contrary, the mirth of the company is highly agreeable to him, and he regards this correspondence of their sentiments with his own as the greatest applause. (TMS I.i.2.1: 14)
It is not, of course, only the desire to be thought charming and funny that generates pleasure when we find we are experiencing emotional states that are in tune with those of the people around us. Smith gives this further illustration:
When we have read a book or poem so often that we can no longer find any amusement in reading it by ourselves, we can still take pleasure in reading it to a companion. To him it has all the graces of novelty; we enter into the surprise and admiration which it naturally excites in him, but which it is no longer capable of exciting in us; we consider all the ideas which it presents rather in the light in which they appear to him, than in that in which they appear to ourselves, and we are amused by sympathy with his amusement which thus enlivens our own. (TMS I.i.2.2: 14)
In this case the agreement between the reader's emotion and her companion's is more a matter of consonance between aesthetic judgments than between moral ones. But the example nicely highlights a key aspect of sympathetic feeling, namely, the emotion the reader feels is essentially a matter of ‘fellow-feeling’ with her companion's emotions. And on Smith's account it is a basic fact that we seek out such instances of emotional agreement with other people; we enjoy them when they happen, and we are displeased when the opposite occurs. Smith considers the latter sort of case in the example of reading to a companion: ‘On the contrary, we should be vexed if he did not seem to be entertained with it, and we could no longer take any pleasure in reading it to him’ (TMS I.i.2.2: 14).
That is, Smith's ‘official’ response to Hume's challenge is the deployment of the distinction between the sentiment of approbation, on the one hand, and the sympathetic passion of the spectator, on the other hand. We shall encounter several reasons for thinking this move is insufficient. In parts 3.1 through 3.4 we will examine the four objections that constitute Hume's challenge and consider other aspects of Smith's account that can be brought to bear on this debate.
3.1
Hume’s counterexample: ‘An ill-humord Fellow; a man tir’d and disgusted with every thing, always ennuié; sickly, complaining, embarras’d; such a one throws an evident Damp on Company, which I suppose wou’d be accounted for by Sympathy; and yet is disagreeable.’ (Corr.: 43)
According to the intellectual historian and Smith biographer, Nicholas Phillipson, the distinction between ‘the sympathetic passion’ and ‘the sentiment of approbation’ provides an adequate response to this objection: ‘While it might be easier to sympathize with a person's joy than with his resentment, the emotional bond that is created by a recognition that we are in sympathy with each other is necessarily pleasurable’ (Phillipson Reference Phillipson2010: 163). Such emotional bonds are indeed crucial, but more needs to be said for the neutral reader to be convinced that Smith wins this debate, as Phillipson's verdict is very much in the minority.
So what are the options for a response on Smith's behalf? Hume has presented a counterexample that he says undermines the claim that it is always pleasurable to sympathize with the passions of another person. The counterexample is a person I think of as the whinging Footnote 2man: the ‘ill-humor'd Fellow’ referenced above.
There are at least two possible replies; the second of which has two components. First, Smitheans can claim that the empirical evidence is not as Hume says it is and argue that it is in fact, on balance, a pleasant thing to interact and sympathize with such a person. Second, Smitheans can resist Hume's inference from the observation that the man's display of his discomforts ‘throws an evident Damp on company’ to the conclusion that this phenomenon is to ‘be accounted for by sympathy’. In other words, one can deny that we sympathize with the whinging man's emotions. In the scenario envisioned on this second option, two points are particularly noteworthy: first, the sentiment of approbation would not arise since there is no sympathetic emotion on which to pass judgment, so to speak, and second, there is a ready explanation for the unpleasantness of the exchange since, on Smith's account, it is always unpleasant to find that we do not sympathize with another's sentiments.
We saw that Phillipson, at least, is convinced that the first option can be made to work. The suggestion is that I sympathize entirely and with all of the whinging man's unhappy sentiments, but that the bond thus created renders the whole exchange pleasurable. That seems an unpromising strategy for a number of reasons. First, the experience of being made unhappy, or at least uncomfortable, by the complaints and grumbles of another person is, I take it, a phenomenon that is familiar to everyone. Second, Alexander Broadie has developed the counterexample into a compelling argument on Hume's behalf. Broadie reads the footnote Smith added to the second edition of TMS as embracing the option that the spectator sympathizes with the whinging man and claims that, as such, ‘Smith's rebuttal begs the question, namely whether observation of agreement gives rise to an agreeable feeling. Smith's reply simply states that it does’ (Broadie Reference Broadie and Haakonssen2006: 173–74). After making that charge of question-begging, Broadie sharpens the point by arguing that any adequate defense of Smith's position, ‘must say what is wrong with Hume's example’ (ibid.).
In light of this argument from Broadie on Hume's behalf, it behooves the Smithean to opt for the second reply set out above, paying close attention to its two components. First, Smith can and should deny Hume's claim that the spectator in question sympathizes with the whinging man's ‘ill-humored’ sentiments. Second, Smith can and should claim that the reason the whinging man ‘throws an evident Damp on company’ is precisely because the spectator does not sympathize with him. The plausibility of the first move is based on the fact that Smith's system includes significant normative constraints both on which emotions we sympathize with and on the circumstances in which we do so. Further, the case presented in Hume's counterexample is a fraught one in this regard, because the norms on sorrow and self-pity are particularly stringent. The plausibility of the second move is based on something about Smith's account that we have already noted, namely, that it causes us emotional pain and discomfort to note that we cannot sympathize with someone. In such a case we experience the very opposite of the sentiment of approbation:
[It] is always disagreeable to feel that we cannot sympathize with him, and instead of being pleased with this exemption of sympathetic pain, it hurts us to find that we cannot share his uneasiness. If we hear a person loudly lamenting his misfortunes, which, however, upon bringing the case home to ourselves, we feel, can produce no such violent effect upon us, we are shocked at his grief; and, because we cannot enter into it, call it pusillanimity and weakness. (TMS I.i.2.6: 16)
The fact that we do not sympathize with such a person's lamentations forms the basis of the judgment that the expression of them is unwarranted. It is important that such a judgment comes after, and at least partially as a result of, an attempt to sympathize with the emotion in question.
Alongside cases of ‘disagreeable’ emotions like that of the whinging man and the sorrow of the person in the quote above, we should note that, on Smith's account, failure to sympathize with positive or ‘agreeable’ emotions can also lead to the judgment that an expression of such an emotion is inappropriate. Consider the following example:
[If] we do not entirely enter into, and go along with, the joy of another, we have no sort of regard or fellow-feeling for it. The man who skips and dances about with that intemperate and senseless joy which we cannot accompany him in, is the object of our contempt and indignation. (TMS I.iii.1.2: 44)
One assumes that, here too, the Smithean psychological principle, ‘it is always disagreeable to feel that we cannot sympathize’ (TMS I.i.2.6: 16), is applicable. I say more about this in part 3.2 of this paper, but if Hume were correct that the ‘evident Damp on company’ that the whinging man manages to bring about were due to our sympathizing with his laments, then ‘the man who skips and dances about’ ought to spread his joy to us through the same mechanism. That is, if my perception of the outward expression of your emotion were enough to generate a sympathetic reaction in me, then both your ill-humor and your joy would be spread to me just by your displaying how you feel. But this, rightly, is not how it works on Smith's account. That is because of the role of the imagination in sympathy and the fact that there are norms influencing whether or not I sympathize with your emotions in any particular situation.
Consider the following passage in which Smith criticizes his philosophical contemporaries for insufficiently emphasizing the importance of the appropriateness of emotional reactions to the conditions that give rise to them:
Philosophers have, of late years, considered chiefly the tendency of affections, and have given little attention to the relation which they stand in to the cause which excites them. In common life, however, when we judge of any person's conduct, and of the sentiments which directed it, we constantly consider them under both these aspects. When we blame in another man the excesses of love, of grief, of resentment, we not only consider the ruinous effects which they tend to produce, but the little occasion which was given for them. The merit of his favourite, we say, is not so great, his misfortune is not so dreadful, his provocation is not so extraordinary, as to justify so violent a passion. We should have indulged, we say; perhaps, have approved of the violence of his emotion, had the cause been in any respect proportioned to it. (TMS I.i.3.8: 18)
The notion of a passionate reaction being justified or unjustified is absolutely crucial here. It is a fundamental feature of Smith's account of sympathy that our assessments of those emotional reactions that are candidates for our sympathy are guided by norms. Finding oneself in sympathy with another person's emotional reaction—both in terms of the type of emotion and the strength (or ‘violence’) of it—is to find oneself in emotional agreement with it. That is the essence of Smith's notion of the sentiment of approbation, and it is at the very foundation of our interactions with our fellow human beings.
In summary, one can successfully respond to Hume's counterexample by saying: first, we do not sympathize with the whinging man, and second, the extent to which he dampens our mood is to be accounted for by the fact that our interaction with him causes us to experience the sentiment of disapprobation.
3.2
Hume invokes the traditionally recognized problem of accounting for the fact that people enjoy tragedies: ‘It is always thought a difficult Problem to account for the Pleasure, received from the Tears and Grief and Sympathy of Tragedy; which woud not be the Case, if all Sympathy was agreeable.’ (Corr.: 43)
Instead of being a problem for Smith, this should be seen as a significant redeeming feature of his account. When it was published, TMS presented the most sustained and thorough development of the concept of sympathy to date. It would be surprising if all it included was a representation of the received view on that topic. When Smith identifies the sentiment of approbation with, ‘the emotion which arises from his observing the perfect coincidence between the sympathetic passion in himself, and the original passion in the person principally concerned’ (TMS I.iii.1.9: 46, note), he provides us with a very good explanation of why people enjoy reading, hearing, and watching tales that involve tragedy, melodrama, and even horror.
This can also be used to explain why a well-crafted tragic story is more enjoyable than one that is shoddily constructed. If a character is richly drawn and multidimensional, the audience member will find it relatively natural to identify with her and, as such, will be in a good position to do the sort of imaginative exercise that Smith tells us is a necessary condition of being able to have an emotional reaction in sympathy with that character. If, on the other hand, the author or director has failed to give the character depth and plausibility, we will find it relatively difficult to imagine what it would be like to be in that character's situation. Besides finding such a story or performance boring, we would, according to Smith's theory, also find it unpleasant in that the characters are doing all that emoting, and yet we find ourselves unable to sympathize. It seems that quite a bit about the experience of reading or viewing tragedy, melodrama, and the like can be very well explained using the resources of Smith's account.
To see how the principles of Smith's system can be used in such explanations, it will be helpful to consider examples where we observe someone having an emotional reaction, but our response falls short of fellow-feeling, such as an angry man with whom we find we cannot sympathize:
The furious behaviour of an angry man is more likely to exasperate us against himself than against his enemies. As we are unacquainted with his provocation, we cannot bring his case home to ourselves, nor conceive any thing like the passions which it excites. (TMS I.i.1.7: 11)
This provides insight not only into the general case of us having little or no patience for people who have trouble controlling their temper, but also into the specific case of a dramatic character whose motivation has not been made sufficiently clear. This helps us understand why Sidney Lumet's 12 Angry Men is a more compelling and satisfying film than John McKenzie's Act of Vengeance. Returning to the example mentioned earlier of the disdain we feel toward ‘the man who dances about with that intemperate and senseless joy’ will remind us that this sort of failure to sympathize can, and does, occur with more pleasant emotions as well (TMS I.iii.1.2: 44). It doesn't take too much effort to see how this helps explain the disparate impacts that slapstick and subtle comedies have on audiences.
Another application of these ideas can be found in Lectures on Rhetoric and Belles Lettres (Smith Reference Smith and Bryce1983, hereinafter: LRBL), where Smith ponders the fact that it is possible to read a well-crafted tragedy many times and still find it engaging, even though we think suspense is the key to the appeal of such a work:
Even in Tragedy, where it is reckoned an essentiall part to keep the plot in Suspense this is not so necessary as in Romance. A tragedy can bear to be read again and again, tho the incidents be not new to us they are new to the actors and by this means interest us as well as by their own importance….
Euripides often in his Prologues by means of a God or a Ghost makes us acquainted with the Events and puts us on our Guard that we may be free to attend to the Sentiments and Action of each Scene, some of which he has laboured greatly. (LRBL ii.30: 97)
Smith's explanation of this phenomenon is that since the events are new to the characters in the play, we use our imaginative capacity to put ourselves in their situation and have emotional reactions in sympathy with the emotions we imagine them having. And it is not the mere having of those sympathetically felt emotions that is the source of our enjoyment. It is rather our observation and recognition of the concurrence—both in type and intensity—of the emotions expressed by the dramatic character and that experienced by us, qua spectators, that gives us pleasure. Once again, when we remember that an inability to sympathize causes distress, we see that Smith's system can be used to explain why poorly made dramas are not enjoyable just as well as it can explain why well-crafted ones delight us.
The upshot of this discussion is that, far from counting as a point against Smith's theory of sympathy, the fact that Smith's account has less trouble than rival theories in explaining why we enjoy tragedies should count as a significant positive feature. Furthermore, the account's flexibility and applicability to many of the ways that emotions are expressed by authors and actors and conveyed to audiences gives us additional reason to think Smith identified some key truths in this area.
3.3
Hume claims that if Smith is correct that ‘all Sympathy was agreeable’, then ‘An Hospital woud be a more entertaining Place than a Ball’.
This can be read as a very brief, and thus perhaps extremely efficient, reductio ad absurdum argument against Smith's position on the putative pleasure brought about by the sentiment of approbation. As such, the first reply that might be made on Smith's behalf could be to simply rehearse and emphasize his distinction between what he calls ‘the sympathetic passion of the spectator’, on the one hand, and ‘the sentiment of approbation’, on the other hand. But there are reasons for thinking this strategy alone will not be enough. Consider the way David Raynor sharpens the issue between Hume and Smith:
The aim of Hume's objection is to show that Smith's assumption that the observation of corresponding sentiments must always be agreeable is false because untrue to experience: in cases where both the communicated passion and its prototype are disagreeable it is simply implausible that the observation of their correspondence should be agreeable. (Raynor Reference Raynor1984: 58)
As he is commenting on two eighteenth-century theorists who are committed to what they called ‘the experimental method’, Raynor is correct to frame this debate as one that should be decided by appealing to what is true (or untrue) to experience. But as we saw in part 1, pace Raynor, Smith does offer anecdotal evidence that we seek out and sympathize with those who are experiencing unpleasant emotions and that we find such experiences rewarding (TMS I.i.2.6: 15–16).
Thus, we have two competing and contradictory appeals to personal experience of sympathetically felt emotions. Smith says, on balance, sympathizing with the afflicted, if there is emotional agreement and thus a sentiment of approbation, is agreeable. For Hume and Raynor this claim is incredible.
In order to dig more deeply into this issue, it will be helpful to ask a seemingly odd question about Hume's example: Is a ball an entertaining place? When we recall Smith's psychological principle that failure to sympathize with the emotions of others is a source of distress, we realize that, for Smith, the answer to this question is: It depends. One line of response on Smith's behalf thus could begin with the observation that sometimes one does not enjoy being at a party. Sam Fleischacker pursues this strategy in a recent essay (Reference Fleischacker, Fricke and Føllesdal2013: 300–303). He asks us to imagine a hospital visit in which we feel very much ‘in synch’ with the pains and discomforts of the patients and then contrasts that with the experience of being at a party where we disapprove of the enjoyment all around us because, say, the fete is being held immediately after a catastrophe. Fleischacker poses the rhetorical question, ‘Where would you rather be?’, and then approvingly quotes Ecclesiastes: ‘It is better to go to house of mourning than to the house of feasting’ (7:2).
There is merit in this response because it is certainly possible to go to a party and have a bad time because you find yourself unable to join in the good cheer. And Smith's account helps us explain this phenomenon. But by focusing on such a nonstandard case, Fleischacker risks underestimating the force of Hume's objection. Parties are usually fun. And Hume is quite reasonable in assuming that sympathizing with the diseased and injured, and perhaps even dying, denizens of a hospital would involve a good deal of sadness and discomfort.
The key to a more comprehensive reply to Hume is acknowledging that since Smith is committed to the view that it is unpleasant to sympathize with painful sentiments, the response should appeal to both ‘the original sympathetic passions’ and ‘the sentiments of approbation’ I might feel in response to those sympathetic feelings. For the sake of argument, let's assume that I sympathize with the same number of people at the hospital as I do at the ball. The number of approvals (which are pleasant) would be equal on the two occasions, but the original sympathetic feelings would presumably be largely unpleasant at the hospital and largely pleasant at the ball. Thus, in the standard case, attending a party would overall be more entertaining than visiting a hospital; which is as it should be. Accordingly, we can respond to Hume's worry in a way that acknowledges what is right about his point while still maintaining that Smith's account holds up to the objection.
The fact that Smith's account provides resources for a relatively nuanced answer to our question about the entertainment factor of balls should draw our attention to an important point of contrast between Hume and Smith on sympathy. Hume invokes a mechanical model centered on a relatively immediate process of the contagion of observed emotions, whereas in Smith's account the process is very much mediated by the imagination. I note that Broadie acknowledges this difference only to set it aside as irrelevant to assessing Hume's objection:
Nor is Smith's case defended effectively by pointing out that Smith's concept of Sympathy is significantly different than the one Hume develops in the Treatise. First, it is highly unlikely that Hume, having read the Theory of Moral Sentiments, would then make the methodologically flawed move of attacking Smith's account of sympathy on the basis of a concept of sympathy that is not Smith's. Second, it is in any case difficult to see how Smith, even on his own terms, can deal plausibly with Hume's counterexample. (Broadie Reference Broadie and Haakonssen2006: 174)
In part 3.1 of this essay, I explained that Smith arguably does have resources to deal with Hume's counterexample ‘on his own terms’. But the matter currently at hand is that Hume's deployment of the contrasting pair of the hospital and the ball compels us to consider the possibility that Hume did not adequately appreciate the differences between Smith's account and his own.
As the noted Hume scholar Páll Árdal has emphasized, the function that the imagination plays in Smith's account is largely absent in Hume's work:
The whole process of sympathizing is conceived of by Hume in very mechanical terms, and the impression of self enters into the picture only as a source of vivacity or liveliness. There is no suggestion that in sympathizing one imaginatively puts oneself into the other person's place, which is characteristic of Adam Smith's account. (Árdal Reference Árdal1989: 45)
In what sense is the process ‘mechanical’ in Hume's account? Sympathy is depicted as mechanistic in the same sense as perception and the formation of ideas are depicted as mechanistic. There is a receptivity and passivity to the process as Hume conceives of it that seems quite different from what we see in Smith's explanation. (Stephen Darwall explores these differences by highlighting what he calls the ‘projective’ character of Smithean sympathy [Reference Darwall1999: 142–45].) Hume highlights these features in the section of his Treatise of Human Nature (Hume Reference Hume, Norton and Norton2000, hereinafter: THN) that is focused on the distinctive character of sympathy:
No quality of human nature is more remarkable, both in itself and in its consequences, than that propensity we have to sympathize with others, and to receive by communication their inclinations and sentiments, however different from, or even contrary to our own. (THN 2.1.11.2: 206)
Hume illustrates this point with some examples that shed light on how he might think about two cases discussed in this essay: namely, the unamused partygoer in Fleischacker's example (above) and the whinging man (part 3.1):
A good-natur'd man finds himself in an instant of the same humour with his company; and even the proudest and most surly take a tincture from their countrymen and acquaintance. A chearful countenance infuses a sensible complacency and serenity into my mind; as an angry or sorrowful one throws a sudden damp upon me. Hatred, resentment, esteem, love courage, mirth and melancholy; all these passions I feel more in communication than from my own natural temper and disposition. (THN 2.1.11.2: 206)
The reference to the sorrowful person who ‘throws a sudden damp upon me’ is, of course, highly reminiscent of the example from Hume's letter. And the poor mood of Fleischacker's disapproving party attendee stands in contrast to Hume's claim that ‘even the proudest and most surly take a tincture’ from the company of those with a ‘chearful countenance’.
Why is Hume so confident that sorrowful faces will dampen the general mood and that smiling ones will spread good cheer? Here it is important to remember that in Hume's view sympathy works by a process of emotional contagion via the observation of expressions of emotions (by which he means behavior such as crying, yelling, and laughing; autonomous responses, such as sweating, eyes dilating, and face reddening as well as actions such as running, talking, and fighting). In contrast to Smith's model of the spectator using her imagination to put herself in another person's place, Hume's account is focused on the spectator's sensory perceptions of the ‘external signs’ of the other person's emotional state:
When any affection is infus'd by sympathy, it is at first known only by its effects, and by those external signs in the countenance and conversation, which convey an idea of it. This idea is presently converted into an impression, and acquires such a degree of force and vivacity, as to become the very passion itself, and produce an equal emotion, as any original affection. (THN 2.1.11.3: 206)
This is the ‘mechanical’ process in Hume's philosophical psychology to which Árdal drew our attention (above) and, in a similar vein, Ryan Hanley emphasizes the relative passivity of the sympathetic agent in Hume's account. (Reference Hanley and Russel2014: n.p.). Also noteworthy is Hume's claim that a sympathetically felt emotion will be of an equal degree of force and vivacity as the original, which stands in contrast to Smith's assertion that such sympathetic feelings are ‘weaker in degree’ than those to which they are reactions (TMS I.i.1.2:. 9). In light of what he says in the letter, these observations lend some credence to the interpretation that Hume may indeed have made the methodological mistake of criticizing TMS, in part, on the assumption that Smith's account of sympathy was more like the one Hume developed in THN than it in fact was.
Obviously, this survey of some of Hume's comments about the origin and psychological potency of sympathy does not constitute a full exposition and examination of his account. My purpose in presenting it is to give a context to his critique of Smith's account. The crux of the third objection presented in Hume's letter is that Smith must have made a mistake when he said there is pleasure in mutual sympathy even when we sympathize with disagreeable emotions. When we consider that objection in the context of the brief survey just presented, we should read Hume's claim in the letter that ‘as the Sympathetic Passion is a reflex image of the principle, it must partake of its Qualities, and be painful where it is so’ with renewed emphasis. The fact that Smith's account of sympathy is centered on a mediating role of the imagination that is absent from Hume's account gives us reason to think this first objection is less powerful than some scholars have taken it to be, and that it can be refuted effectively. (For examples of the opposing view that Hume's objection has no effective response, see Raynor [Reference Raynor1984: 56–61]; Blackburn [Reference Blackburn1998: 202–204]; and Broadie [Reference Broadie and Haakonssen2006: 172–74].)
3.4
Hume accuses Smith of treating all cases of sympathy as being like cases of sympathizing with a friend: ‘when we converse with a man with whom we can entirely sympathize, that is, where there is a warm and intimate Friendship, the cordial openness of such a Commerce overpowers the Pain of a disagreeable Sympathy, and renders the whole Movement agreeable. But in ordinary Cases, this cannot have place.’ (Corr.: 43)
Here Hume clearly has some of Smith's remarks from the TMS chapter ‘On the Pleasure of Mutual Sympathy’ in mind. As we saw above in part 3.1, Smith draws our attention to the behavioral evidence that we are moved to console the distressed as much as we are to celebrate with the successful and that ‘the pleasure which we find in the conversation of one whom in all the passions of his heart we can entirely sympathize with, seems to do more than compensate the painfulness of that sorrow with which the view of his situation affects us’ (TMS I.i.2.6: 15–16). This is precisely the point where Hume engages Smith to tell him his account has gone awry.
A key point of the controversy is the question of what the reference class is of ‘one with whom in all the passions of his heart we can entirely sympathize’. Now, there is nothing in the quote from Smith (above) to suggest that the reason the spectator can ‘entirely sympathize’ with the person's sorrow is that the person is the spectator's friend. But that does seem to be what Hume has in mind. Broadie spells out and amplifies Hume's take on the issue by emphasizing the possibility that ‘in the case of friendship, a good deal of baggage comes along within the relationship which might easily explain why sympathy is associated in a special way with pleasure’, and by contrasting such cases with ones in which we do not know or even dislike the person (Broadie Reference Broadie and Haakonssen2006: 174).
Smith does not address this specific accusation in the footnote he added to the second edition of TMS, but I think Smith's account has the resources to respond to this charge. One way to get at the issue is to ask just what sort of ‘baggage’, to use Broadie's term, would explain why sympathizing with a friend is ‘associated in a special way with pleasure’. On the face of it, there seems to be a very compelling case to be made for the claim that sympathizing with a friend who is sorrowful, angry, or grieving might be more disagreeable than sympathizing with such an emotion in a person we do not know very well. Imagine that my good friend has been estranged from her father for over two decades but has recently made overtures to him and received a positive response. This has raised the hope of a reconciliation. Alas, her father has a fatal stroke before there is a chance for the two of them to meet and try to bridge their differences. I am aware of all of this when I hear the news of his demise and try to comfort her in her loss. Wouldn't my sympathy with my friend's sorrow and grief be all the more intense and painful precisely because I have this intimate knowledge about her situation? Accordingly, would it not be much more painful than my sympathy for the sorrow of someone whose loss of a parent I learn of but whom I do not know very well?
This example highlights the fact that whether or not a person is my friend may well have significant bearing on the intensity of the ‘sympathetic passion’ (to use Smith's terminology) I feel along with that person. And, as Smith allows, that experience ‘may either be agreeable or disagreeable, according to the nature of the original passion, whose features it must always, in some measure, retain’ (TMS I.iii.1.9: 46, note). But for Smith this is, as it should be, a separate issue from whether we recognize a consonance (or dissonance) between our sympathetic reaction and what our friend is feeling.
All of this points to another, deeper, response to Hume and Broadie's objection that Smith has conflated all instances of sympathy with the case of sympathizing with a friend. The Smithean can agree that we are more likely to sympathize with the emotions of our friends, but insist that the reason is that our relatively extensive knowledge of their situation gives us a rich set of resources to draw upon to help us imagine being in their situation. This debate helps us see another way in which Smith's emphasis on the contributions of the imagination to the mechanism of sympathy lends a great deal of explanatory capacity and subtlety to the account.
Going into the full details of Hume's alternative picture would take us too far afield to serve present purposes, but it will be useful to take note of the importance that the relation of resemblance has in his thinking about how, first, we select our friends:
‘Tis obvious, that people associate together according to their particular tempers and dispositions, and that men of gay tempers naturally love the gay; as the serious bear an affection to the serious. This not only happens, where they remark this resemblance betwixt themselves and others, but also by the natural course of the disposition, and by a certain sympathy, which always arises betwixt certain characters. Where they remark the resemblance, it operates after the manner of a relation, by producing a connexion of ideas. (THN 2.3.4.6: 229)
And how, second, those similarities between our own characters and those of our friends might dispose us to sympathize more readily:
[We] find, that where, beside the general resemblance of our natures, there is any peculiar similarity in our manners, or character, or country, or language, it facilitates the sympathy. The stronger the relation between ourselves and any object, the more easily does the imagination make the transition, and convey to the related idea the vivacity of conception, with which we always form the idea of our own person. (THN 2.1.11.5 207)
It is plausible that these ideas are behind a key assumption implicit in this second objection of Hume's about the interconnection of friendship and sympathy. As the foregoing discussion makes clear, Smith's account has advantages in this area that have not been adequately appreciated by commentators and that, arguably, make it better than Hume's account.
To recap, the Smithean spectator readily sympathizes with his friend's emotional reactions not because of some special ‘baggage’ or because the operations of the relation of resemblance, but because his awareness of her situation increases his capacity to imagine what it would be like to be in her shoes. And the lack of an appeal to the relation of resemblance means that the possibility is left open that this spectator might not sympathize if, after the imaginative exercise, he finds that he does not share her emotional reaction. Those are both compelling features of Smith's account and help to provide an effective response to Hume's fourth objection.
4. A Note about the Difference between the Discussions of the First Two Objections and the Second Two
Before closing this section on Hume's four objections, I shall emphasize an important difference between the context of my discussions of the first and second objections, on the one hand, and the third and fourth objections, on the other hand. In considering the whinging man (the first objection) and the case of tragedies (the second objection), the responses I have presented are freestanding, as it were. By contrast, in the course of discussing the hospital/ball comparison (the third objection) and treating all cases like friendships (the fourth objection), I have suggested that Hume may have misread TMS by assuming that the account of sympathy presented therein is more like his own account than it in fact is. Therefore, the discussions of the third and forth objections depend on some claims about both Hume's own account and about how it contrasts with Smith's account. The key to the contrast between Smith and Hume is cashed out in terms of the imagination.
The interpretation of the difference between the role of the imagination in Smith and Hume's respective accounts of sympathy that I have outlined here is consistent with discussions in Árdal (Reference Árdal1989: 139–47), Darwall (Reference Darwall1999: 142–45), and Hanley (Reference Hanley and Russel2014: n. p.). There is room for disagreement, however. Griswold sees a great deal of continuity between the two accounts of imagination and writes of Smith's ‘creatively developed Humean view of the imagination's constructive and unifying power’ (1999: 178). However support for my interpretation can be found in Griswold, for alongside his emphasis on the influence of Humean sympathy on Smith's account, he also sees the latter as a development and extension of the former (1999: 340, n. 38).
The discussion of Hume's account of sympathy that is perhaps most in conflict with my own is that of Gerald Postema (Reference Postema2005). Postema focuses on the interplay between the Humean psychological mechanisms of sympathy and comparison. By starting with that interaction and by emphasizing that Hume thinks sympathy often engages with the general point of view, Postema presents Hume's account of sympathy as much more intertwined with the imagination than writers like Árdal, Darwall, Hanley, and I do. While acknowledging that Hume indeed frequently depicts sympathy as a mechanism through which we take on sentiments that resemble what others around us are feeling, Postema also notes that Hume thinks we can use the imagination to sympathize with a broader range of cases than the perception-based model would suggest.
Postema pays particular attention to a passage in THN in which Hume begins: ‘There remains only to take notice of a pretty remarkable phaenomenon of this passion: which is, that the communicated passion of sympathy sometimes acquires strength from the weakness of its original, and even arises by a transition from affections, which have no existence’ (THN 2.2.7.5: 239). In that passage, Hume considers a very similar case to one we saw in Smith—namely, a case in which we feel ashamed on behalf of people who behave foolishly but ‘show no sense of shame, nor seem in the least conscious of their folly’ (ibid.).
I think it is important that Hume closes this passage by characterizing such phenomena as cases of ‘partial’ sympathy and emphasizing the contrast between those and the standard cases: ‘All this proceeds from sympathy; but ‘tis of a partial kind, and views its objects only from one side, without considering the other, which has a contrary effect, and wou'd entirely destroy that emotion, which arises from the first appearance’ (ibid.). How one understands the relationship between Smithean and Humean sympathy vis-à-vis the imagination will depend on just what place one thinks these cases of partial or ‘one-sided’ sympathy should be given in the explication of Hume's account.
Attempting to settle that interpretive debate would take me beyond the scope of this essay on Hume's objections and Smith's responses to them. I am forced to content myself, first, with acknowledging that there is a controversy to be had on this front and, second, with noting that my claim that the imagination plays a more central role in sympathy for Smith than it does for Hume is relevant only to the third and fourth responses I am discussing and not to the first two.
5. Conclusion
Hume's letter to Smith in which he identified a putative problem with the ‘Hinge’ of Smith's system in TMS contains four objections to Smith's account of sympathy. In this essay, I have shown that there are resources in Smith's account to respond to each of those four objections. In so doing, I have demonstrated that Smith's system is more robust than many commentators have thought. As such, it may well prove to be a comparatively stable basis for building a sentiment-based system of ethics.
