Olivia Bailey
Empathy, Care, and Understanding in Adam Smith’s Theory of Moral Sentiments
I. Introduction
When we observe and interact with other people, it is impossible for us to really
think that they are not also experiencing the world. Inevitably, we understand people to
be minded. And not only that: it also seems to us utterly obvious that we have some
acquaintance with the content and character of others’ experiences. I understand that
the store clerk is bored, and that the children long for ice cream. I understand how the
bitterly disappointed fourth-place finisher is feeling, and what the shame of the man
who accidentally knocked over the Ming vase must be like. When we encounter another
person whose inner world is a true mystery to us, it is a startling and disquieting
experience. These individuals are the unsettling exception, rather than the norm. At the
same time, it also seems obvious to us that we are interested in others’ passions,
thoughts, and attitudes, independent of their instrumental significance for us, and not
only out of intellectual curiosity. That is, we care about others’ inner lives. What others
think and feel matters to us in a way that can directly motivate us to act on their behalf.1
In The Theory of Moral Sentiments, one of Adam Smith’s aims is to defend these
commonsense convictions about care and understanding from solipsism and
psychological egoism. Why does Smith see these twin positions as posing a threat that
demands a response? At least part of the reason, I think, is that he endorses a particular
picture of the difference between self-directed understanding and concern and otherdirected understanding and concern. I will call this the ‘picture of egocentric primacy.’
This picture presents our self-care and self-understanding as uniquely primitive and selfevident. Our own minds are immediately transparent to us, but “we have no immediate
experience of what other men feel” (TMS, I.i.1.2). Similarly, our self-care is immediate,
but our care for others looks like the kind of thing that we can only come to through
considerable mental maneuvering. Once one admits that our understanding and care for
others needs to be explained, one had better follow up with an explanation. The skeptic
can press the point: if the origin and nature of our understanding and care for others is
1
Or, indeed, against their interests– care in the sense that I am concerned with is not always altruistically
oriented.
1
mysterious, and if we cannot dispel that impression of mystery with a plausible
explanation of how and why we come to care for and understand others, then our
everyday impression that we do in fact understand and care for others may not be good
enough reason to reject psychological egoism and solipsism.
One resource that has appealed to both early-modern and contemporary
philosophers faced with this conundrum is the psychological phenomenon now
generally known as empathy. What exactly empathy consists in has been the subject of
lively debate since the time of Hume and Smith’s exchanges about the nature of what
they called “sympathy, or…fellow feeling” (TMS, I.i.1.3; see also TMS, VII.iii.3.17).
For our purposes, though, the following loose but handily ecumenical definition of
empathy should suffice: empathy consists in feeling what another person feels, or at
least imagining feeling what another person feels, not because one is literally in the same
situation as she is, but because one has in some other way come to imaginatively engage
with something like the other person’s experience. Empathy tends to interest
philosophers operating with some version of the picture of egocentric primacy because
it seems like it could provide the means of ‘bootstrapping’ up from our intimate
understanding of and care for ourselves to our understanding of and care for others. In
this paper, I will consider Smith’s attempts to fend off skeptical worries about
understanding and care by appealing to our capacity for empathy. My aim here is partly
critical, and partly rehabilitative. Taken together, Smith’s explicit pronouncements on
empathy’s role in fostering understanding and care generate a problem. Like his friend
Hume, whose theory of empathy heavily influences Smith, Smith assigns a double duty
to the series of mental operations that issue in empathy, treating it as the source of both
our understanding of other people and our non-instrumental care for them. However, I
will argue that if Smith’s empathetic mechanism does generate an accurate
understanding of the other, that understanding will simply not be the right kind of
acquaintance to generate care. Only a seriously confused grasp of the attitudes and
passions of the other could give birth to a heretofore absent non-instrumental care for
others. As we will see, the fulfillment of either one of the empathetic mechanism’s
supposed functions requires conditions that will make it impossible for the other
function to be fulfilled.
2
Smith’s official account of empathy is seriously flawed. However, his theory of
human sociability also contains within it the seeds of an important improvement upon
the official account. Whether or not Smith actually recognizes it, one element of his
account actually demands the conclusion that an important form of care precedes
empathy and empathetically-derived understanding. Specifically, his explanation of why
we engage in imaginative projection actually entails (or at least comes close to entailing)
that concern for the other’s thoughts and feelings is often (if not always) a necessary
condition of our being motivated to empathize. This would mean that care is not
something we get to through empathetic ‘bootstrapping up’ from self-care; it is prior to
empathetic feeling. And if care is prior to empathetic feeling, then the problem that
emerges from his official account does not get off the ground: that problem arises only
because Smith officially treats both care and understanding as empathetically-derived. I
will aim to make it clear that given Smith’s conception of empathy, he should on pain of
inconsistency be committed to a very different conception of care’s relation to empathy
and understanding than the one he more explicitly endorses.
II. Smithian empathy: an overview, and a problem
i. care and understanding
Before entering into the details of Smith’s account, I want to identify more
precisely the nature of the care and understanding that interests Smith. Let us first
consider the range of attitudes we might classify under the heading of ‘care’. What
does it mean to care for or about someone? We tend to think of care as an attitude of
benevolence, one associated with a motivation to aid the object of our care.
However, we can also identify a broader sense of care, one that encompasses all of
our interest in others and their inner lives, benevolent or not. Of course, our interest
in others may be more or less motivated by our own self-concern. The kind of care
that Smith regards as most in need of explanation is a concern for others’
experiences, attitudes, emotions, and intentional actions that is not motivated by selfinterest. I will refer to this particular kind of care as basic care. Significantly, this kind
of care includes distinctly non-altruistic attitudes. For instance, I might care about my
enemy’s agony just because his unhappiness matters to me absolutely.
3
We can now turn to understanding. Understanding ranges along a scale of
depth. At the shallow end, we have the understanding that consists in an acquaintance
with another’s behavioral patterns. This is mere predictive understanding. Then there
are the varieties of understanding that are relevant only in those cases where we
recognize the individual in question as having a mind. Classification of the various
forms our understanding of other inner lives can take is inevitably contentious. One
particularly difficult question is this: if we are to understand another’s inner life, how
important is it that we appreciate the other’s experience? By experience, I mean the felt
character of inner life, that aspect of our mental existence that we tend to refer to as
‘what it is like.’ One might think that we can distinguish between two ways of
understanding another as minded, as follows: one the one hand, some understanding of
an individual’s feelings and beliefs can reproduced in propositional form without
remainder. And on the other hand, some understanding includes an ineffable sense of
‘what it is like’ for the other. But this way of carving up our understanding invites
difficult questions: Can I truly be said to understand that you are thinking of me if I
don’t know what thinking of me feels like for you? Can I truly be said to understand
that you are sad if I do not have an idea of how your sadness feels ‘from the inside’? Or,
to take a more extreme case: can I truly be said to understand that you are sad even if I
myself (through some happy stroke of luck) have never once been sad myself?
I cannot begin to answer these questions here. I will say, though, that Smith
generally takes understanding another’s thoughts and feelings to involve either 1)
drawing upon old knowledge about what it is like to be sad, or curious, or desirous of
fame, and so forth, 2) producing a new and vivid idea of what the episodes of another
person’s life must be like for them, or both. For him, and also for those contemporary
philosophers of mind who take inspiration from him, understanding another person’s
inner life critically involves acquaintance with that person’s experience.2 So, my
discussion of the relation between care and understanding will naturally focus on a
conception of understanding that emphasizes the importance of experience.
2
Philosophers of mind who adopt this position, and who also claim to draw upon Smith, include Gordon
(1995) and Goldman (1992), along with authors who have a more pronounced cognitive-scientific
orientation, such as Kiesling (2012).
4
ii. empathy as a ‘bootstrapping’ device: the official theory
On the very first page of The Theory of Moral Sentiments, Smith both gives us
his version of the picture of egocentric primacy and identifies empathy (which he
calls “sympathy”) as the means by which get outside of ourselves, and arrive at
understanding of and care for others. His identification of empathy as the means by
which we come to understand others is straightforward: “As we have no immediate
experience of what other men feel, we can form no idea of the manner in which
they are affected, but by conceiving what we ourselves should feel in the like
situation” (I.i.1.1). Clearly, Smith thinks our apprehension of others’ experience can
only come through a mental operation that exploits our familiarity with the only
inner life we directly apprehend, namely our own. His affirmation of the other half
of the picture of egocentric primacy requires a little more work to see clearly,
coming as it does in the first, rather infamously garbled sentences of the text. Smith
writes:
However selfish soever man may be supposed, there are evidently some
principles in his nature, which interest him in the fortune of others, and render
their happiness necessary to him, though he derives nothing from it except the
pleasure of seeing it. Of this kind is pity and compassion, the emotion which
we feel for the misery of others, when we either see it, or are made to conceive
it in a very lively manner. (TMS, I.i.1.2)
To see what is going on in these lines, we need to understand what Smith means by
“pity or compassion.” We tend to associate ‘pity’ and ‘compassion’ with feeling bad
about someone’s unhappy situation, and feeling motivated to ameliorate it. These
certainly seem to be phenomena distinct from what I have been calling empathy, which
involves feeling (or imagining feeling) something like what someone else feels.
However, Smith’s later uses of ‘compassion’ suggest that he may not have in mind a
notion of compassion that necessarily involves benevolence (see, e.g. TMS, I.i.1.10 and
VI.iii.15). Rather, it is likely that he has in mind the more archaic (now obscure) sense
5
of compassion that just denotes participation in someone else’s suffering.3 If that is
correct, then we can understand Smith as holding the view that at least compassion (and
possibly pity also) are a subset of the feelings generated through the process of
empathizing with another. This interpretation is supported by Smith’s later statement:
“Pity and compassion are words appropriated to signify our fellow-feeling with the
sorrow of others. Sympathy, though its meaning was, perhaps, originally the same, may
now, however, without much impropriety, be made use of to denote our fellow-feeling
with any passion whatever” (TMS, I.i.3.5).
Smith writes that there are “some principles in [man’s] nature, which interest
him in the fortunes of others, and render their happiness necessary to him,” and that
pity or compassion is “of this kind.” I take him to be using the phrase “of this kind” to
signal that pity or compassion is one of the principles in human nature that has the
effect of generating both interest in other’s fortunes and care for their happiness. Since
Smith also describes pity or compassion as an emotion, and since an emotion does not
seem to be the kind of thing that can be a principle, we can assume that what Smith
really means is that our tendency to feel pity or compassion is one of these principles of
our nature. At any rate, he seems to be endorsing the following general order of
explanation: our tendency to feel pity or compassion (by which, I have argued, he
means our tendency to empathize), is a principle of our nature that causes our basic care
for other people’s feelings and fortunes, and not the other way around.
So, according to the opening page of The Theory of Moral Sentiments, both
understanding and basic care are to come from empathy. This is what I will refer to as
Smith’s ‘official theory’ of empathy. In a moment, I will move on to the problem that
this dual role generates, but first we need to look more closely at the details of Smith’s
conception of empathy.
iii. empathy as imaginative projection
Smith’s description of this mental operation uses various spatial metaphors that
help to give a sense of what he has in mind: frequently, we are told that empathy
involves “enter[ing] into” another’s situation. 4 Smith also sometimes speaks of
3 This meaning of ‘compassion,’ now deemed “obscure” by the Oxford English Dictionary, first appeared in
English in the 1340’s (The OED Online entry for ‘compassion, n.’ June 2012, Oxford University Press).
4The phrase recurs a total of 73 times in The Theory of Moral Sentiments, including e.g. at I.i.1.4; I.i.1.8;
I.i.2.2; and I.i.2.6.
6
“bringing the case home” to oneself.5 Following the lead of these spatial descriptions, I
will call the operation in question ‘imaginative projection.’ 6 Properly engaging in
imaginative projection means mentally conjuring up as much of the detail of another’s
circumstances as possible, as though one were in the other’s position: “the spectator
must, first of all, endeavor, as much as he can, to put himself in the situation of the
other.” (TMS, I.i.4.6).
What, exactly, counts as another’s situation is a difficult question. Should it
include only a person’s material circumstances? How about their previous experiences,
or even their likes and dislikes? Smith wavers on this point. About empathizing with the
mother of a dead son, he writes: “I do not consider what I, a person of such a
character and profession, should suffer, if I had a son, and that son was unfortunately
to die: but I consider what I should suffer if I was really you, and I not only change
circumstances with you, but I change persons and characters” (TMS, VII.iii.1.4). Here,
Smith is conceiving of agents’ situations in what we can call a broad sense, one that
encompasses facts about our inner life, as well as our material circumstances. However,
he also writes: “Sympathy…does not arise so much from the view of the passion, as it
does from that of the situation which excites it. We sometimes feel for another, a
passion of which he himself seems to altogether incapable, because, when we put
ourselves in his case, that passion arises in our breast from the imagination, though it
does not from his in reality” (TMS, I.i.1.10). This suggests that Smith is thinking of the
situation with which we imaginatively engage as something more like the sum total of
the other person’s material circumstances, and less like the sum total of their material
circumstances, plus their history, their preferences, their fears, and so on. Overall,
though, it makes sense to think that insofar as our goal is to understand someone, the
more complete we can make our projection into her circumstances (broadly
understood), the better. To illustrate: I will better understand how a lost child feels if I
am able to imagine not just how I would feel if I, with my current education, size,
emotional maturity, and so forth were lost, but rather how I would feel if I could not
read, did not know how to ask for help, and so forth.
5
See e.g. I.i.1.4; I.i.2.6; I.i.3.9; and II.i.3.2.
Here I am following Charles Griswold, who describes Smith’s notion of imaginative engagement as
“projective imagination” (1999: 90).
6
7
When we imagine ourselves in another’s situation, Smith thinks, we actually
experience something like a weak version of the feelings we imagine the other to be
having. Smith calls this kind of feeling a “shadow”: “Every man feels his own pleasures
and pains more sensibly than those of other people. The former are the original
sensations, the latter the reflected or sympathetic images of those sensations. The
former may be said to be the substance; the latter the shadow” (TMS, VI.ii.1.2). The
comparison to a shadow helps to capture the quality of the feeling attained through
projective imagination: just as a shadow mimics the outer form of an object exactly, but
in a duller tone, an observer’s empathetic feeling has the same object and the same
phenomenal quality as the original emotion, except that it is less vivid. It is this less
vivid feeling that is, according to the opening of The Theory of Moral Sentiments, supposed
to be the source of our basic care for others. The feeling is, apparently, enough to make
us perceive others’ fortunes as worth caring about for their own sake. In particular, it
makes us see others’ happiness as “necessary to us” (TMS, I.i.1.1).
For Smith, the imaginative transportation into the other’s situation is typically
effortful. It is true that we have a natural capacity for empathy, but we are all affected
by various motivational and imaginative limitations that affect our ability to empathize,
and so we often “find it…difficult to sympathize entirely, and keep perfect time” with
other people’s feelings (I.iii.1.8). Especially when it comes to others’ sorrows, we must
battle against a “dull sensibility to the afflictions of others” in order to see things from
their points of view (I.iii.1.13). The task of imaginatively abstracting from our own
position is, Smith thinks, hard enough that we reserve some of our highest admiration
for the person who is able to feel nearly as much for others as he does for himself
(TMS, VI.iii.1.1 ff.).
iv. the problem with the official theory
The idea that a weak but actual feeling, rather than the mere conception of a
feeling, is needed in order for us to have basic care for other’s feelings makes a certain
sense given that Smith is, after all, a sentimentalist. Though his theory of sympathy is
ultimately quite different from Hume’s, he has no quarrel with Hume’s general
insistence upon the idea that only sentiment, and not reason, can ultimately motivate
8
us.7 However, the idea that we need to feel a shadow of another’s feeling in order to
care for him runs into a serious problem. In brief, the difficulty is this: it is hard to see
how the experience of a feeling could be anything but the experience of one’s own
feeling, and the experience of one’s own feeling is not a suitable source for basic care
for others.
To bring out the problem more clearly, let us consider a case in which I am to
empathize with another person’s suffering. Suppose I have turned my attention to Jan,
whose investments in the tulip market have fizzled spectacularly. Taking stock of Jan’s
circumstances and outward behavior, I attempt to imaginatively enter into his position.
I imagine that I once was rich, but am suddenly poor, and all because of some bad bets.
Let us suppose (not I think, implausibly) that part of Jan’s inner state is a painful regret
occasioned by the enduring presence of the thought: “I have lost my whole fortune.” If
I am to successfully understand Jan’s inner state, I should, according to Smith,
experience a shadow of this painful regret. What, precisely, will be the nature of that
shadow? To answer that question, we need to think a bit about what work the indexical
‘I’ and the indexical possessive pronoun “my” are doing in the thought that is the focus
of Jan’s regret. What does Jan regret, exactly? In some sense, it would be correct to say
that Jan regrets the loss of Jan’s fortune. But unless he is highly unusual, it is not the
loss of Jan’s fortune qua Jan’s fortune that matters to Jan. It is, rather, the loss of his
fortune that he cares about. Jan’s thought, “I have lost my fortune, ” is not really
interchangeable with the thought “Jan has lost his fortune,” (or even “I have lost Jan’s
fortune”). A counterfactual will bring out the point: if by some odd fluke Jan had
forgotten that he was that man whom people call Jan, then the thought “I have lost
Jan’s fortune” would have a resonance very different from that of the thought “I have
lost my fortune.”8 It might prompt Jan to feel guilty for having ruined another person’s
future, but it would not cause him to long for his bygone salad days. Clearly, Jan’s regret
is not for the fortune of the man named Jan, who happens to be him. Rather, it is
essential to the nature of his regretful thought that the one who thinks it is the same
person who has lost the fortune, and knows that he is that person.
7
See the Treatise, II.iii.iii and following.
John Perry introduced the use of counterfactual analysis to show that indexicals are essential: swapping
third-personal designators in for first-personal designators inevitably alters the character of attitudes. See
Perry (1993: 3-53).
8
9
The character of Jan’s thought will not be accurately preserved if my shadow of
Jan’s regret includes the thought “Jan has lost his fortune,” or “I have lost Jan’s
fortune.” Jan’s thought “I have lost my fortune” refers to the person who is thinking
the thought as the one who has lost a fortune, and so my recreation of Jan’s regret must
have this same structural feature. The thought’s indexicals must be preserved ‘as is’ in
my shadowy imaginative reconstruction of Jan’s inner life. That means that when I
come to understand Jan’s regret through imaginative projection into his situation, I
myself will be pained (if only a little) by the thought, “I have lost my fortune.”
Otherwise, my experience of the putative shadow of his inner life will not be accurate.
But in that case, the only care that painful thought should generate is care for my own
fortune, and not care for Jan or for his fortune!
This conclusion needs a minor qualification: my regretful feeling could cause me
to be concerned about Jan’s regret, but only instrumentally. If empathy with painful
sentiments will itself be a bit painful, and if I am bound to empathize at least some of
the time, it would be natural for me to prefer that others not have such sentiments, and
so I might hope for the return of the man’s fortune. However, this is not basic care, as I
have defined it, and it is presumably not the kind of concern for other people that
Smith has in mind. After all, my preference in this case would be just as well served by
any circumstance that interrupts the empathetic mechanism, including my falling asleep
or simply forgetting about Jan. Such purely instrumental concern for others’ feelings is
still fundamentally selfish.
Effectively, Smith is caught in a bind. According to his picture of how empathy
works, we understand a person’s feeling insofar as we are able to accurately reproduce
the other’s passion in us though imaginative effort. When we succeed in understanding
the other, only a variation in liveliness distinguishes their passion from the shadow of it
that we experience. However, the development of understanding looks to be at odds
with the generation of care. My empathetic experience of someone else’s regret about
their situation can only result in my feeling regret about my own situation– a particularly
strange result, given that I may well be in an enviable situation, myself.
A confused grasp of Jan’s regret does look like it could generate care for Jan, of
a sort. If my imaginative reconstruction of Jan’s feelings included a bitter regret focused
on the thought “Jan has lost his fortune,” then my empathetic feeling would involve a
10
shadow of regret directed at Jan’s loss of his fortune, rather than my loss of my fortune.
If I come to regret Jan’s having lost his fortune, then perhaps it makes sense that I
would come to care about Jan and his situation (although I am not sure about that). But
getting to care in this way means sacrificing a real understanding of Jan’s inner life. Jan’s
regret does not focus on the thought “Jan has lost his fortune,” and if my reconstruction
of Jan’s regret includes this kind of focus, then I have misunderstood Jan’s inner life in
a significant way. Furthermore, there is something suspect about the kind of care we
could expect this kind of reconstruction to produce. Perhaps it would not seem so odd
for me to care about Jan’s loss in the way that a regretful feeling would demand. But
suppose that the feeling in question were François’ unhealthy self-hatred. It would be
utterly bizarre, and not at all consistent with our expectations of what basic care should
amount to, for my care for François to arise from empathetic self-hatred (or, depending
upon how we think of François’ feelings, from hatred of François!).
Schopenhauer described the question of how our sharing others’ feelings could
lead us to care for others as “the great mystery of Ethics, its original phenomenon, and
the boundary stone, past which only transcendental speculation may dare to take a step”
(170). Smith’s official attempt to dispel this mystery, whilst also accounting for our basic
understanding of others’ passions, does not succeed. His account draws on what strike
us as familiar thoughts: it seems reasonable enough to think that I understand you at
least in part by relying upon my ideas of my own experience, and that my feeling
something like what you are feeling has something to do with my caring for you
(witness the prevalence of injunctions to “walk a mile in my/their shoes”). However,
when we try to determine exactly how our empathy with others relates to our
understanding and care for them, difficulties like Smith’s rapidly crop up.
Later on in the The Theory of Moral Sentiments, Smith does qualify his ‘substance :
shadow :: original feeling : empathetic feeling’ analogy in a way that indicates he sees a
problem with his original take on the relations between empathy, care, and
understanding. Of those observed by another, Smith writes: “What they feel will,
indeed, always be, in some respects different from what he feels…because the secret
consciousness that the change of situations, from which the sympathetic sentiment
arises, is but imaginary, not only lowers it in degree, but, in some measure, varies it in
kind, and gives it a quite different modification” (TMS, I.i.4.7). Smith does not say what
11
this “quite different modification” is supposed to be, and he does not mention anything
like it again. Still, his appeal to it may signal a kind of recognition that an exact copy of
another’s emotion, taken by itself, is not a realistic source of care for the other. Perhaps
without realizing it, Smith seems to be grasping for a way of having his cake and eating
it too. That is, he is trying to ensure that the empathetic emotion is different enough
from the original that it can be a source of care for the other, whilst also holding on to
the idea that the one is enough like the other that it could be a source of true
understanding of the other’s inner life. The fact that Smith is only willing to talk about
the hypothetical modification in the most vague of terms does not bode well for this
strategy, however.
In the next and final section of this paper, I am going to turn from criticism to
an attempt at a constructive modification. I will argue that Smith should abandon his
commitment to the idea that empathy generates care for others. Doing so will dissolve
the problem I have discussed above, but this is not the only reason to reject the idea
that empathy generates care. Whether he realized it or not, Smith’s account of our
motivation for empathizing with others practically entails that basic care is at least
sometimes a precondition of empathetic imagination. So, giving up on his particular
strategy for countering psychological egoism will make his theory more coherent in two
respects.
III. Care and Understanding: the Implicit Smithian Alternative
i. sentimental harmony and the motivation to empathize
As I have mentioned, Smith describes imaginative projection primarily in terms
that make it look like it involves deliberate work. For Smith, perfect fellow feeling is a
goal that we struggle to meet. He writes of the spectator’s task: “the spectator must,
first of all, endeavor, as much as he can, to put himself in the situation of the other, and
to bring home to himself every little circumstance of distress which can possibly occur
to the sufferer. He must…strive to render as perfect as possible, that imaginary change
of situation upon which his sympathy is founded” (TMS, I.i.4.6). Smith’s talk of
“endeavoring” and “striving” makes it clear that this is an difficult undertaking; it may
be natural, but it is hardly automatic. Because Smith rejects the conception of
imaginative projection as automatic or effortless, he is faced with this question: why are
12
we motivated to perform the imaginative exercises that generate fellow-feeling in the
first place? In what follows, I will argue that Smith ought to treat basic care as a
precondition of a kind of imaginative projection that is central to human sociability, the
only kind that Smith pays any real attention to. Were it not for the attitude of basic care,
we would not be driven to engage in imaginative projection in a large and important set
of cases. As we have seen, for Smith, imaginative projection is the essential means by
which we acquire experiential understanding of others. Taking the two previous points
together, we will be able to conclude that in a large and important set of cases, we
would not come to understand the target agent if we did not already care about him.
We have no particular reason to assume that there is just one explanation for
why we engage in imaginative projection. It seems obvious that imaginative projection
could sometimes be helpful for securing benefits for, and avoiding harm to, ourselves.
For instance, if I want to avoid losing my lunch money, it could be helpful to project
myself into my bully’s position, and imagine just what kind of groveling I would find
most appealing if I were him. But Smith is quite clear that we do not only engage in
imaginative projection when knowing a person’s mind could be important to our
fortunes. It is hard to see how fellow-feeling with another person could be useful to us
in cases where the other is unlikely to ever interact with us, and yet Smith is adamant
that our fellow-feeling can and does extend to such people: we are liable to “bring home
to ourselves” the misery of “any innocent and sensible being,” not just the misery of
those in our own country (TMS, VI.ii.3.1). So, in cases where knowing the other’s mind
does not seem useful to us, why do we nevertheless engage in imaginative projection?
Smith does have something to say about why we are motivated to engage in
imaginative projection. And reflection upon the explicit reason he gives for our
motivation will reveal that Smith should not endorse the claim that empathy produces
caring. According to Smith, we “passionately desire” to attain a “harmony of hearts”:
we want our feelings to match others’ (TMS, I.i.4.7). This desire is at work in our
attempts to reconcile our opinions about things like art or philosophy with the opinions
of others. As Smith points out, sharing others’ opinions in these cases does not require
“sympathy, or that imaginary change of situations from which it arises, in order to
produce, with regard to these, the most perfect harmony of sentiments and affections”
(TMS, I.i.4.2). But Smith argues that the desire for harmony is still stronger in cases
13
where attaining a harmony of hearts means entering into another’s position, and
understanding things from that person’s point of view. When we cannot harmonize our
feelings with those that an agent feels about his surroundings, the result is “shock” and
“pain” (TMS, I.i.2.1).
Smith identifies the harmony of hearts as the source of great pleasure for us:
“[N]othing pleases us more to than to observe in other men a fellow-feeling with all of
the emotions of our own breast” (TMS, I.i.2.1). When we fail in our efforts to bring our
own feelings in line with those of another person, conversely, the result is a kind of
pain: “it hurts us to find we cannot share his uneasiness” (TMS, I.i.2.6). Smith is careful
to clarify, in response to an objection from Hume, that even when we feel another’s
pain, which is at least partly an unpleasant experience, the fact of our feelings’
correspondence is itself an unfailing source of happiness.9 Of the silence and mirth of
our companions, when we ourselves are mirthful, Smith writes: “this correspondence of
the sentiments of others with our own appears to be a cause of pleasure, and the want
of it a cause of pain, which cannot be accounted for” except by the thought that the
correspondence is an independent source of pleasure (TMS, I.i.2.2). Note that Smith is
not arguing that we find sentimental harmony pleasurable because it is good for us in
some other way. We can of course imagine cases where this would be true: I am pleased
that we both enjoy scary movies, because it means I don’t have to watch Alien alone.
Or, referring back to the case of the bully, feeling something like what the bully feels
could be pleasurable, because I take pleasure in having the resources to avoid the bully’s
extortion. But on Smith’s view there is something pleasurable about the sharing of
sentiment regardless of any consequences. Sentimental harmony, it seems, is pleasurable
in and of itself.
Smith does not pin down the relationship between the pleasurableness of
sentimental harmony and its desirability as precisely as we might like. Pleasure makes its
first appearance in the opening lines of The Theory of Moral Sentiments: “How selfish
soever man may be supposed, there are evidently some principles in his nature, which
interest him in the fortune of others, and render their happiness necessary to him,
though he derives nothing from it except the pleasure of seeing it” (TMS, I.i.1). We
might wonder, here, what explanatory role (if any) pleasure is playing: should we think
9
See Hume’s letter to Smith dated 28 July 1759 (letter 36), in Smith (1987).
14
of the “principles” as directly motivating our interest in and care for others, whilst also
making it the case that we take pleasure in sharing others’ pleasures? Or should we
instead think that the “principles” are the facts of our nature that make it the case that
sentimental harmony is pleasurable for us, and that the prospect of pleasure is in fact
what directly motivates us to empathize? In other words: do we want sentimental
harmony because it feels good, or is the fact that it feels good just an indication of its
inherent desirability? Smith’s attempts to characterize our desire for sentimental
harmony do frequently invoke pleasure. Furthermore, Smith explicitly adopts the
position that “Pleasure and pain are the great objects of desire and aversion” (TMS,
VII.iii.2). These considerations encourage the thought that for Smith, it is the
pleasurableness of sentimental harmony that directly motivates us to pursue it.10 However,
there is simply not enough textual evidence to draw that conclusion with any real
confidence, and one could argue that reading Smith in that way makes it look as though
he accords pleasure an unrealistically prominent role in the explanation of empathy. For
our purposes, however, this interpretive question is not critically important. The
important point is that Smith regards sentimental harmony as an end that is either
inherently desirable and thus capable of directly motivating us to empathize, or as an
end that is desirable because it is pleasurable.
Before moving ahead, let us sum up what we have so far established. According
to Smith, all experiential understanding is the product of (successful) imaginative
projection. Smithian imaginative projection is typically effortful, so it makes sense to ask
why we are motivated to engage in it. One reason to engage in it is that it provides an
understanding of others that could be useful to ourselves. However, we have reason to
think that not all imaginative projection is motivated by such benefits. Smith himself
emphasizes the fact that we are motivated by a desire for sentimental harmony, which is
itself pleasurable. Now, why do people find sentimental harmony desirable and
pleasurable, even in cases where there is no instrumental benefit to be accrued from the
harmony? Smith does not give us an explicit answer. My purpose here is just to show
that on Smith’s view, basic care must precede the experiential understanding empathy
provides in a large and important number of cases. Therefore, I am not going to
attempt to give a full account of why we find sentimental harmony desirable and
10
Or, alternatively, we might say that it is the pain of disharmony that drives us to empathize with others.
15
pleasurable. Rather, I am going to argue that we can only make sense of the distinct
pleasure that we take in sentimental harmony in and of itself, a pleasure that in turn
appears to motivate our imaginative projection, if we conceive of ourselves as feeling
basic care for others prior to imaginative projection.
ii. care’s role in the motivation to empathize
What would our attitude toward others have to be like in order for us to take
pleasure in sharing their feelings, in cases where no instrumental benefits are at stake?
At this point, I see two possibilities. The first is that there is nothing in particular that
our attitude toward others would have to be like. On this view, the pleasure we take in
the correspondence of sentiments would not in any way be contingent upon our
regarding others’ sentiments as independently valuable or worth knowing about. The
other possibility is, of course, that the pleasure we take in the correspondence of
sentiments is contingent upon our regarding others’ sentiments as independently
valuable or worth knowing about. How can we decide between these two possibilities?
Smith is all for psychological realism. Therefore, I propose that we take a cue from him,
and begin by considering whether the first view seems like a good fit with our pretheoretical familiarity with human motivation.
One might argue that harmony just is a pleasing state to be in, regardless of
whether we care about others’ opinions as such. It just feels good to have the same
feelings that someone else does about the new blockbuster, or the politician’s behavior,
or the German football team. It may not be immediately obvious that sharing
sentiments couldn’t feel pleasurable in itself in the absence of a pre-existing basic
interest in others’ feelings. Upon reflection, however, it seems hard to imagine how it
would be psychologically possible to be motivated to seek sentimental harmony just
because it is inherently pleasurable, without also being interested in or concerned with
others’ feelings.11 It would be something like being motivated to keep a promise by the
thought that breaking a promise would make one feel bad, even if one does not care
11 Dick Moran has suggested that a ‘negative’ version of this possibility might be more psychologically
plausible. That is, it might be psychologically accurate to claim that we could find sentimental disharmony
painful even if we did not have pre-existing care for others. I am unsure what to say about this suggestion.
Personally, this possibility does not strike me as any more realistic than did the corresponding claim about
the pleasurableness of harmony, but I recognize that my intuitions on this point may be idiosyncratic.
16
about promises themselves. Why would one feel bad about breaking the promise if one
did not care about promises as such?
It is easy enough to think of cases in which we might value a correspondence of
our attitudes with those of others without considering others’ attitudes to be
independently valuable or interesting. However, in every case I have been able to
conjure up where we take pleasure in correspondence without considering others’
attitudes to be independently valuable or interesting, the pleasurableness of the
correspondence seems to be due to the instrumental benefits it affords. Take this
example: I am pleased to share your positive feelings about the German football team.
Now, it seems possible that I could be pleased by this without having any antecedent
interest in your opinions. It could be that I just want the cheers for the German side to
be loud, in order to increase their chances of winning. But suppose there are no
instrumental benefits to our correspondence of sentiments in this case. Could I still
desire that our sentiments correspond, without being interested in or concerned with
your opinion as such? Such a desire just might be psychologically possible, but it strikes
me as improbable and strange. It seems to me that the pleasure I take in our shared
feelings about the German team must either be covertly instrumental, or dependent
upon my pre-existing interest in others’ attitudes. And in some cases, even a pleasure in
sentimental harmony that is instrumental will nevertheless be dependent upon a preexisting care for others’ sentiments as such. For instance, I may take pleasure in our
having the same sentiments about the German team at least in part because it means I
will have someone to cheer along with. But it could very well be that I would not enjoy
cheering along with you unless I regarded your sentiments as independently interesting,
something to be cared about.
I have argued that since Smith treats imaginative projection as effortful, it makes
sense to ask why we are motivated to engage in it. On occasion, the instrumental
benefits to be accrued from sentimental harmony, and the pleasure derived from those
instrumental benefits, may motivate imaginative projection. However, Smith does not
discuss such cases. He is interested in cases where we are motivated to imaginatively
project just by the prospective pleasurableness of sentimental harmony itself. And when
it comes to these cases, an attitude of interest or concern directed at the other– that is,
an attitude of care– seems to be a precondition of taking pleasure in sentimental
17
harmony. Since we are talking about cases in which instrumental interests are not at
play, we can further specify that the care in question is basic care. To be sure: this line
of reasoning does not amount to a deduction that basic care is sometimes a
precondition of imaginative projection. There is no abstract, conceptual consideration
that dictates this conclusion. Rather, the conclusion’s warrant is based upon an
observation about human psychology, one that is admittedly not drawn from a
complete and thorough survey of our motivational profiles. Still, it seems to me that it
hits on a true fact about the ways we take interest in other people’s lives.
If Smith were to conceive of the relation between basic care and imaginative
projection (and, by extension, the relation between basic care and understanding) in the
way that I suggest he should, given his explanation of our motivation to empathize,
then he would not in fact end up in the bind in which his official theory is entangled. It
may already be obvious why the alternative conception of the relation between care and
understanding avoids this problem, but let us make it explicit by employing one of
Smith’s own examples of an effort to achieve sentimental harmony.
iii. an illustration of the amended theory
Late in The Theory of Moral Sentiments, Smith describes the operation of
imaginative projection in the case where one is relating to a widow or widower:
When I sympathize with your sorrow or indignation, it may be pretended,
indeed, that my emotion is founded in self-love, because it arises from
bringing your case home to myself, from putting myself in your situation, and
thence conceiving what I should feel in the like circumstances. But though
sympathy is very properly said to arise from an imaginary change of situations
with the person principally concerned, yet this imaginary change is not
supposed to happen to me in my own person and character, but in that of the
person with whom I sympathize. When I condole with you for the loss of
your only son, in order to enter into your grief I do not consider what I, a
person of such a character and profession, should suffer, if I had a son, and
that son was unfortunately to die: but I consider what I should suffer if I was
really you, and I not only change circumstances with you, but I change
persons and characters. My grief, therefore, is entirely upon your account, and
not the least upon my own. (TMS, VII.iii.1.4)
Now, let us consider the example of sympathetic grief in the light of what we
have said about the relation between care and understanding in Smith. In an effort to
attain an understanding of your grief, to reach sentimental harmony with you, I put
myself in your position by imagining being you. If the effort is successful, I will imagine
feeling what you are feeling. I will feel devastated “as” you. I will feel grief at having lost
18
my child. This feeling could in fact be described as selfish or self-directed, in the sense
that it concerns me and what was in some sense mine. Smith’s claim that I “enter into
your grief” by imagining being “really you” suggests that this is what has in mind.
However, this feeling of mine can serve an informational function in the effort to arrive
at a feeling that is not self-directed. As I have argued, it cannot spontaneously generate
basic care for you. But what if I already care about you? What if I am already interested
in your experiences? What if your feelings matter to me, even before I have experienced
an echo of them? In that case, it makes sense that I would be motivated to put the
feelings I derive from imaginative projection to use. In particular, I can infer that you
are experiencing something like the grief that I am experiencing (faintly) in virtue of my
imaginative engagement with your situation. And now that I have some idea of what
your experience must be like, even though I cannot directly experience it myself, my
general care for you is better informed. This will help me to demonstrate my
understanding of your plight and otherwise condole you.
IV. Conclusion
From Smith’s remarks about our desire for sentimental harmony, I have drawn
out an implication that is in tension with some of Smith’s more explicit
pronouncements concerning the relation between care, empathy, and understanding.
Smith officially identifies the empathetic mechanism as the source of both our interest
in others and our concern for them. However, Smith also maintains that we engage in
imaginative projection because we desire sentimental harmony (either because it is
inherently pleasurable or because it is a good in and of itself). And reflection upon our
own psychological make-up supports the conclusion that an orientation of basic care is
in fact a condition of finding sentimental harmony non-instrumentally desirable.
Therefore, if Smith 1) wishes his theory to be coherent and 2) does not want to give up
his account of why we are motivated to empathize, he should treat basic care not as the
product of imaginative projection, but as a something that often precedes it. In many
cases, basic care might even be regarded as the driving force behind the imaginative
effort in question.
19
As we have seen, Smith’s official account faced a major problem. The two tasks
he assigns to the empathetic mechanism are at odds. For Smith, the key to
understanding another is to arrive at conception of their passion that is as close as
possible to the original, an achievement possible thanks to the mind’s imaginative
capacity. This same mental process is also supposed to generate care for the other. As
the case of Jan’s regret made clear, a truly accurate ‘shadow’ of another person’s feeling
is not the kind of thing that can generate the right kind of care for the other. It might be
misleading to say that the approach I have suggested on Smith’s behalf would solve this
problem. It would be more apt to say instead that it dissolves the problem. In effect, it
prevents the problem of accounting for both care and understanding through empathy
from arising, by scaling back on the explanatory ambitions for empathy.
As I hinted above, empathy could be relevant to how we direct and manifest
our basic care. I might be generally interested in Jan, and think of his inner life as
mattering to me, but it is only when I come have a sense of his regretful feelings that
my care is properly informed. If my care is of the good-willed variety, I might use my
understanding to offer Jan support; I might, for instance, let him know that I once
suffered from similar feelings of regret, but was able to overcome them by refocusing
my energies on work (other than tulip speculation!). That said, if Smith were to opt for
the approach to care, understanding, and empathy that I have claimed would be
consistent with his story about why we imaginatively project, then he would have to
give up on the idea that basic care actually emerges out of empathy.
In the first section of this paper, I suggested that Smith’s response to the twin
skeptical threats of psychological egoism and solipsism is shaped by his commitment to
the picture of egocentric primacy. This picture, which treats our own self-knowledge
and self-care as uniquely primitive and not in need of explanation, has a prima facie
plausibility that is reflected in its considerable philosophical staying power. Many of
Smith’s philosophical descendants endorse this picture to some degree. Indeed, the
current wave of interest in empathy seems to be largely propelled by the thought that
we somehow, as Nancy Sherman puts it, “step beyond the egocentric point of view”
(1998: 83). Claims like Sherman’s take it for granted that we begin with selfunderstanding and self-care, and must somehow turn outwards in order to relate to
others as minded beings whose inner lives matter. If Smith were to adopt the view that
20
basic care precedes empathy, as I have suggested he should, would he need to rescind
his commitment to the perennially popular picture of egocentric primacy?
His commitment to the epistemic component of the picture of egocentric
primacy, the thought that we cannot feel others’ feelings directly, need not be affected.
For better or for worse, the improvement upon Smith’s explicit account that I have
proposed retains the assumption that our understanding of others’ minds must come
though bootstrapping up from our primitive self-understanding. When it comes to the
other half of the picture, the story is more complicated. I have argued that Smith should
recognize basic care as a necessary precondition for much of our empathetic
engagement with others, and for the understanding that this engagement produces.
Strictly speaking, this recognition would not undermine the claim that our care for
ourselves is primitive and needs no explaining, whereas our care for others is not
similarly primitive and does need to be explained. It does undermine Smith’s official
explanation of our basic care for others, however. And so we are left with two
possibilities: Smith would either need to supply a new explanation of why we feel basic
care for others, one that does not rely on empathy to move from care for ourselves to
care for others, or he would need to forego the search for such an explanation and
accept basic care for others as a primitive psychological disposition on a par with our
self-care. With empathy eliminated as an explanation for basic care, though, it is very
difficult to see what realistic alternative explanations might be available to Smith if he
were to pursue the former option. Therefore, although the claim that basic care
precedes empathy does not strictly require that we abandon the care component of the
picture of egocentric primacy, it does place would-be defenders of the care component
of the picture in a difficult position.
This paper has been focused on the particular ways that Smith appeals to
empathetic mechanisms in his theory of mind and morals, but the problems that Smith
encounters are not just the result of minor quirks in his descriptions of human
psychology. His problems begin with the picture of egocentric primacy. When we posit
such a dramatic gap between our how we relate to ourselves and how we relate to
others, it is entirely natural that we should attempt to bridge that gap by appeal to selfcare and self-understanding, phenomena with which we are so intimate that they seem
to require no explanation. However, empathy-based solutions to the problem of
21
accounting for care for others run into the problem of appropriately acknowledging the
distinction between self and other. Reproducing something like others’ feelings ‘in’
ourselves does not seem to actually get us to care for the other. And, while I did not
discuss the question in this paper, we might also ask whether this kind of reproduction
actually gets us to a real grasp of others’ passions, or if it only gets us to an idea of our
own passions that we mistake for an idea of others’ passions. Ultimately, I think that
these and related problems give us a reason not just to look for alternative ways of
bridging the gaps between self and other, but to look at the picture of egocentric
primacy, which posits these gaps in the first place, with a specially critical eye.
22