It has been a long time since the Adam Smith scholarship started to correct the oversimplified and somewhat biased image of Smith as ‘the father of modern economics’. One key moment occurred around the completion of the Glasgow edition of Smith’s works in the 1970s–80s. The importance of The Theory of Moral Sentiments (TMS)1 has since been widely recognised. This is not only because it provides the fundamental basis for a thorough understanding of Smith’s economic thought in An Inquiry into the Nature and Causes of the Wealth of Nations, but also because many have begun to recognise Smith’s original ideas on ethics, which remain relevant to political and ethical discussions today.2

Although there are many approaches to TMS, one involves examining how Smith constructs his argument in response to the other thinkers of his time. If we concentrate on the specific approach to morality suggested in the title of his book, Francis Hutcheson and David Hume are particularly prominent, as all three share the core belief that morality cannot be explained without referencing emotions or sentiments.3 Consequently, considerable effort has been dedicated to exploring their connections. In line with this direction, the paper aims to deepen our understanding of Smith’s moral theory by examining his relationship with David Hume.

This paper argues that Smith’s criticism of Hume, especially that developed in Part IV of TMS, can be seen as Smith’s application of a specific view of human nature presented in Hume’s epistemology. More specifically, it concerns Smith’s reinterpretation of Hume’s argument that describes natural propensities in contrast to reflection, assigning a more dominant mental influence to the former. In light of this understanding of the relationship between Smith and Hume, I will show that the gist of Smith’s criticism of Hume is not about a simple question as to whether moral judgment is derived from sympathy or reflection on utility, but rather about determining the natural propensity of the mind that affects our moral judgment. Hence, this argument is broadly twofold: (a) to reveal Smith’s debts to Hume’s epistemology and (b) to revisit Smith’s criticism of Hume. Before turning to the main discussion, I would like to briefly touch on each point.

First, this paper discusses Smith’s connection to Hume; however, this approach is not entirely novel. Many Smith scholars have explored this issue. One of the examples notes that Smith adopts Hume’s concept of ‘The science of man’, implying it shaped Smith’s overall approach and methodology.4 Furthermore, because the primary focus of TMS is moral subjects, it is natural that considerable attention has been paid to Hume’s moral arguments, which seem directly relevant to TMS. Although the most obvious topic might be ‘sympathy’,5 much research has been dedicated to Smith’s critique of Hume’s utilitarian view of morality.6

Although the tendency to focus on Hume’s moral theory is understandable, it seems to downplay one of Hume’s most important contributions to philosophy, that is, epistemology. Indeed, compared to the studies mentioned above, few scholars explore the link between Smith’s moral theory and Hume’s epistemology.7 This raises questions as to whether their intellectual relationships have been properly assessed, given the historical records suggesting their close friendships.8 That is, if Smith did indeed read A Treatise of Human Nature (Treatise)9 in Balliol, which Hume considered to contain philosophically important implications, it seems strange that Smith learned nothing from Hume’s epistemology. Following ‘a prolegomenon to a more sustained and more fully developed reconsideration of Smith’s debts to Hume’,10 I will consider this intellectual relationship in detail and propose that Smith adopted a particular view of human nature from Hume.

Second, drawing on this connection between Hume and Smith, I argue that Smith’s critique of Hume can be understood differently than previous research. More specifically, I will show that Smith’s view of moral judgment is more nuanced than the common understanding, which tends to simplify the contrast between Hume and Smith as utility versus sympathy. My interpretation also has some implications for Smith’s famous statement on the punishment of a sentinel (TMS II.ii.3.11), suggesting that his remarks should be regarded as an integral part of his moral theory, rather than an exception.

In summary, this paper aims to open a new perspective on Smith’s relationship with Hume by proposing an alternative approach to this subject. In so doing, it offers a fresh understanding of one of the most debated aspects of Smith’s moral theory: the sentinel example. To support these points, the argument is divided into three parts. In the next section, I will turn to Hume’s argument on the scepticism of the senses and highlight an aspect of Hume’s argument that I believe influenced Smith. In section two, I revisit Smith’s critique of Hume in TMS and determine exactly on which point Smith disagrees with Hume. In the final section, I argue that Smith’s critique of Hume can be interpreted as Smith’s application of a particular view on human nature presented by Hume and present some implications of my interpretation.

To begin with, it is necessary to clarify the framework that I argue Smith learned from Hume.11 It is not about the specific content of natural beliefs discussed by, for example, Norman Smith,12 but about a specific view of the human mind. This framework has two points. First, it recognises different mental states, which might lead us to arrive at different conclusions on a given issue depending on which thought process prevails. Second, one of those distinct processes is more dominant and influential than the other and is called natural. This is the general framework that I argue Smith reinterprets in TMS.

To locate the framework within Hume’s epistemology, let us examine his argument concerning scepticism about the senses in Book One, Part Four of Treatise, as an example. The primary focus of this section is our beliefs in external objects. Specifically, Hume aims to analyse ‘the causes which induce us to believe in the existence of body’ (T 1.4.2.2). I will concentrate here on his description of the process by which we form ‘the opinion of a continu’d or of a distinct existence’ (T 1.4.2.2), and divide it into three stages for analysis.

The first stage is where we come to believe ‘a continu’d as well as an invariable existence’ (T 1.4.2.37). To elucidate how we conceive this idea, Hume begins the argument by explaining how we attribute identity to perceptions or objects that are interrupted but similar. It should be noted that when explaining this stage, Hume explicitly assumes a careless attitude that does not distinguish between perceptions and objects, which he describes as ‘the opinions and belief of the vulgar with regard to the existence of body’ (T 1.4.2.31). How then do we believe distinct but similar perceptions identical? While Hume’s explanation is rather complicated, the resemblance is essential here, as he states:

Resemblance here includes not only the resemblance between each perception but also that between the dispositions we have when viewing an object with perfect identity and when seeing similar but separate perceptions (T 1.4.2.32). In other words, the resemblance between separate but similar perceptions causes a seamless flow of imagination, which can be mistaken for the experience of observing an object with a perfect identity without any interruption. Consequently, according to Hume, an association of ideas and a similar association of dispositions lead us to ascribe identity to distinct but similar perceptions (T 1.4.2.32).

However, in the second stage, we reflect on the idea of continued existence and become sceptical of it. Hume observes:

According to Hume, even though we believe in a continued and distinct existence in the first stage, our mental process does not stop there. We move on to reflect on this matter and realise that the idea cannot be sustained, because it is clear not only that, if we reason strictly, we cannot ascribe identity to separated but similar perceptions, but also that ‘our perceptions are not possest of any independent existence’ (T 1.4.2.45). If you were one of those reading Treatise for the first time, you might expect this reflection would persuade us not to fall into the same mistake; however, Hume tells a different story.

In the third stage, according to Hume, we attempt to resolve the detected contradiction by differentiating between perceptions and objects and regard the latter as a distinct continued existence, as he says:

Does this solution work? Hume says no because it lacks a reasonable foundation. That is, our causal inference can only operate among perceptions and will never extend beyond them to reach objects that are separate from perceptions (T 1.4.2.47). Then, in Hume’s account of our belief in external bodies, we end up with another unreasonable hypothesis to avoid the false idea of assuming continued existence. How does this seemingly unreasonable situation occur? According to Hume, it is due to the same principle that initially led us to believe in the mistaken idea of continuous existence at the first stage, as he claims that this ‘philosophical system acquires all its influence on the imagination from the vulgar one’ (T 1.4.2.49).

So far, we have reconstructed Hume’s argument concerning the scepticism of the senses. Hume appears to describe a series of mental processes in which we swing between two different mental states. First, he explains why we believe in the idea of continued existence. However, we do not stop there; we reflect on its plausibility and find it false. This implies that our beliefs in continued existence lack solid, rational grounds. Then, not being satisfied with this conclusion, we seek for another theory that could be compatible with the idea of continued existence, which had already been dismissed by reflection in the second stage; therefore, we distinguish between perceptions and objects, insisting that objects should be considered continued and distinct. In short, Hume’s description illustrates how we revert to the first principle after experiencing a reflective state.

It should be noted that the principle which induces people to believe in the idea of continued existence is characterised as natural by Hume; for instance, Hume argues that ‘we are led after this manner, by the natural propensity of the imagination, to ascribe a continu’d existence to those sensible objects or perceptions’ (T 1.4.2.44).

Furthermore, Hume writes how dominant and influential this natural principle is in our mind as follows:

From the examination above, it seems clear that Hume captures multiple principles at work, which would lead to forming different views on the given subject, while pinpointing the natural one as a prominent principle of the human mind. According to his analysis of scepticism about the senses, we can approach the same subject, namely external things, from different perspectives, and the natural principle is more likely to influence the way we approach the issue than the reflective attitude that doubts the truth of the idea of continued existence.

Before proceeding to the next section, two points should be noted. First, when I say Hume acknowledges multiple principles at work in the preceding paragraph, I follow Hume’s usage of the term, when he states that ‘[a]s long as our attention is bent upon the subject, the philosophical and study’d principle may prevail; but the moment we relax our thoughts, nature will display herself, and draw us back to our former opinion…. But tho’ our natural and obvious principles here prevail above our study’d reflections, ’tis certain there must be some struggle and opposition in the case; at least so long as these reflections retain any force or vivacity’14 (T 1.4.2.51–52). I do not understand the opposition between the philosophical and studied principle and the natural principle to be one between distinct mental faculties. Indeed, many scholars seem to accept the view that imagination is a key concept underlying Hume’s philosophy.15 For the present case, it can be taken as ‘the conflict between causal reasoning and the belief in “continu’d and distinct” existences’16, both of which can be considered to have imagination as their foundation. Still, Hume seems to admit different mental states or processes that can produce conflicting views on the same subject and to express them as different principles in some parts of Treatise.

Second, I have confined my analysis to Hume’s argument of scepticism about the senses. It is reasonable to ask whether the same framework can be found in the other arguments in Book One of Treatise. While the importance of the task cannot be denied, it would require at least a thorough examination of each section of Part 4 of Treatise, which I cannot provide within the space allowed in this paper. For the present, I would like to note that some scholars have pointed out the connection between the discussion of scepticism about the senses and the conclusion of Treatise,17 which would indirectly show the significance of that argument in his entire philosophical inquiry in Book One of Treatise.

The next section revisits Smith’s criticism of Hume, with the framework determined in this section in mind.

Since Part IV of TMS is widely recognised as a dedicated critique of Hume’s view on moral judgment, it has been heavily relied on when Smith students examine how his theory diverges from Hume’s; however, in examining that part, Hume’s epistemology has received little consideration.19 In this section, I reconsider Smith’s criticism of Hume to illustrate that this critique can be read as Smith’s application of Hume’s framework to his moral theory.

Regarding the arguments presented before Part IV of TMS, he clarifies the roles assigned to them: while Part I and II focus on our moral judgments of others, Part III centres on the moral judgment of oneself (TMS III.1.1). In these three parts, Smith largely employs the same principle (TMS III.1.2): sympathy. According to Smith, we make moral judgments based on our sympathetic emotions experienced through imaginary changes of situations (TMS I.i.1.2, I.i.3.1). Hence, it is fair to say that Smith’s arguments regarding moral judgment before Part IV of TMS rely on this principle of sympathy.

In Part IV, titled ‘The Effect of Utility upon the Sentiment of Approbation’, Smith responds to views that seem different to those presented in the first three parts of TMS. In Smith scholarship, this section is generally interpreted as Smith’s critique of Hume,20 with the implication that Smith refers to Hume when he states, ‘[t]he same ingenious and agreeable author who first explained why utility pleases, has been so struck with this view of things, as to resolve our whole approbation of virtue into a perception of this species of beauty which results from the appearance of utility’ (TMS IV.2.3). In fact, Smith’s description of the theory—that virtuous qualities, ‘according to this system, are originally valued as useful to ourselves’ or ‘esteemed on account of their usefulness to others’ (TMS IV.2.5)—appears to align with Hume’s idea of four distinct sources of pleasure and pain, which form the basis of moral judgment (T 3.3.1.30).

What Smith first questions is whether our moral approbation and disapprobation are, as Hume insists, derived from recognising the utility of each character (TMS IV.2.1–3). In other words, Smith asks whether we approve of a character because we reflect on its usefulness to society or individuals. There are two reasons that lead scholars to conclude that Smith entirely rejects Hume’s view. First, this explanation seems different from what Smith aimed to establish earlier in the discussion of TMS, namely, sympathy. Second, Smith himself seems to clarify this point, when he says:

Thus, it appears evident that Smith is trying to distance himself from Hume’s view. Smith’s illustrations about prudence (TMS IV.2.6–8) and generous qualities such as benevolence and justice (TMS IV.2.9–11), and Hume’s explanation of both qualities (T 3.3.3–4) are in stark contrast. Of course, the accuracy of Smith’s depiction of Hume’s moral theory can be questioned. This point was noted early on by Thomas Reid, who states, ‘Doctor Smith has given the most distinct and Accurate Account that is to be found any where of the Several Theories concerning Morals both Ancient & moderns that were before him, excepting Mr Humes [sic] System’.21 Nevertheless, it seems clear that Smith’s intention is to distance himself from Hume’s theory: The next question is, at exactly which point does Smith disagree with Hume?

The simplest answer would be that while Hume believes moral judgment is based on utility, Smith considers it based on sympathy. One example representing this view can be found in D. D. Raphael’s claim that ‘Hume is by and large a utilitarian. Adam Smith is an anti-utilitarian’.22 For another example, A. L. Macfie observes that ‘Smith then was not a utilitarian, even in Hume’s sense’.23 Although this contrast between utility and sympathy seems to be implicitly accepted since it is noted by the editors of the Glasgow edition of TMS, my view in this paper aligns more closely with that of Smith’s contemporary, Dugald Stewart. In ‘Account of the Life and Writings of Adam Smith, LL.D.’, he describes the relationship between Smith and Hume as follows:

Stewart’s testimony indicates that Smith’s theory does not exclude Hume’s view; instead, the former develops a moral theory that can comprehend the latter. This suggests that Smith’s critique of Hume cannot be seen as an exclusive opposition between utilitarian and sympathetic perspectives. There are several reasons to support this interpretation.

First, it aligns with Smith’s statement that we sometimes pass verdicts based solely on utility. One example appears in one of the most discussed passages in TMS, in which Smith mentions punishment for a sentinel who falls asleep during duty. He argues that ‘[u]pon some occasions, indeed, we both punish and approve of punishment, merely from a view to the general interest of society’25 (TMS II.ii.3.11). Although this has been considered one of the difficulties in TMS,26 it would not cause confusion if we understand that Smith does not exclude moral judgment based on utility. In my view, Craig Smith provides an accurate explanation. He argues that ‘[i]n Smith’s own account the long term utility of punishing the Centinel [sic] involves different sentiments and a different thought process to that involved in punishing the murderer. The centinel example demonstrates that there may be occasions where the demands of distinct moral sentiments that produce distinct moral judgment processes can come into conflict’.27 I agree with the view that Smith describes two distinct moral sentiments in the passage about the sentinel.

Second, Smith himself counts what he describes as utilitarian judgment to be among the sources of moral sentiments. Near the end of TMS, Smith summarises his theory as follows:

It is notable that Smith includes what, according to some scholars, he opposes in Part IV of TMS as a source of moral sentiments, using expressions similar to those he employed in his critique of Hume. While Smith argues ‘it seems impossible that the approbation of virtue should be a sentiment of the same kind with that by which we approve of a convenient and well-contrived building’29 (TMS IV.2.4), he does not deny a form of beauty, which is derived from utility and ‘which we ascribe to any well-contrived machine’,30 as a source of our moral sentiments.

Now, if we return to Smith’s initial reaction to Hume, it actually aligns with Smith’s comprehensive view of moral judgment. For example, when Smith claims that ‘it is not the view of this utility or hurtfulness which is either the first or principal source of our approbation and disapprobation’ (TMS IV.2.3), this remark merely excludes utility from being ‘the first or principal source’ of moral approbation, but not from any qualification as a moral sentiment. Accordingly, the situation might not be as straightforward as Leslie Stephen’s brief description that ‘Adam Smith tried to resolve all our moral sentiments into sympathy’.31

Suppose that Smith does not exclude utility as a source of moral sentiments. Smith’s critique of Hume should not be interpreted as a binary choice between utility and sympathy. What I would like to claim here is that Smith’s fundamental concern was not choosing between the two ways of moral judgment but comprehending their relationship. The clues to this understanding lie in his insistence that utilitarian reflection is not the first or principal source of moral sentiment. What does Smith mean by this? Two things can be considered. Firstly, considering utility requires a reflection that demands a stretch of imagination, which explains why utility is not, according to Smith, ‘the first’ source of our moral sentiments. For example, Smith argues:

Thus, according to Smith, recognising utility involves a more complex thought process than sympathy; therefore, for most people, it is not the first thing that comes to mind.

Secondly, Smith observes that people tend to be influenced by the sympathetic process rather than through utilitarian reflection. For example, Smith explains why people tend to become obedient to the powerful and wealthy, referring to ‘a peculiar sympathy’, which causes us to enter into them and ‘favour all their inclinations, and forward all their wishes’ (TMS I.iii.2.2). In this discussion, he contrasts ‘the doctrine of reason and philosophy’, which is characterised by ‘a regard to the utility of such submission’, with ‘the doctrine of Nature’, which concerns sympathy with the rich and powerful (TMS I.iii.2.3). Smith thus emphasises how much more dominant and influential the latter principle is compared to the former:

So, it can be argued that, for Smith, utility is not the ‘principal’ source of moral judgment either. Smith observes that we might sometimes reflectively consider political issues, such as the right of resistance, but such a reflection can be easily overridden by another mental process: sympathy.

To summarise our analysis, Smith views utility and sympathy as distinct sources of moral judgment, indicating that he recognises distinct mental processes that could produce different moral judgments. Also, Smith observes that considering utility is more reflective or philosophical and less prominent in the mind, whereas sympathy is less reflective or philosophical and more dominant. It should be added that, by emphasising these aspects of sympathy, Smith seems to characterise sympathy as the natural process for moral judgment, as his contrast between ‘the doctrine of reason and philosophy’ and ‘the doctrine of Nature’ suggests.

With these points in mind, let us consider another key passage in the final part of TMS, where Smith is considered to discuss the difference between his theory and Hume’s. Smith says:

The phrase ‘the natural and original measure’ was added in the fourth edition of TMS, but it is crucial for understanding Smith’s critique of Hume because it seems to highlight the exact point where Smith disagrees with him. That is, the essential problem for Smith was what should be regarded as the natural principle of moral judgment. To recap what we have established so far, Smith admits that in some cases we can pass moral judgments based solely on utilitarian reflection, but he also notes that it is not the most influential or dominant mental process in the mind; for Smith, the natural principle is sympathy, not reflection on utility. Smith’s disagreement with Hume therefore primarily concerns how to understand the relationship between different mental processes within human nature. Consequently, when we return to the question ‘exactly on which point does Smith disagree with Hume?’, the answer I offer is that Hume fails to understand the natural principle of moral judgment correctly.

What enables Smith’s criticism of Hume is his recognition of the multiple mental processes that produce moral judgment. I think this is one of the most notable features of TMS among his contemporary moralists, who tended to build moral theories on a single principle. For instance, Francis Hutcheson sought to identify a specific sense that functions in moral judgment, and Joseph Butler clearly acknowledged the central role of conscience in his analysis of human nature. In my view, neither of them seriously addresses the case in which different mental processes lead to conflicts among moral sentiments, as in the sentinel example, nor do they pay attention to determining the relationship between different mental processes concerning moral judgment.

In the final section, I argue that Smith developed this unique moral theory by applying the framework within Hume’s epistemology, and I also consider some implications of my interpretations.

Although TMS can be seen as unique compared to other moral theories of its time in treating different mental processes that concern moral judgment, Hume appears to have developed his epistemology with a similar framework. Like Smith, Hume’s epistemology addresses multiple mental processes that can influence our judgment, while also recognising that we tend to view things from a particular perspective, which he describes as natural.

Based on the analyses so far, I present the following interpretation: Smith’s critique of Hume in TMS is his application of Hume’s framework within his epistemological argument. The framework shared by Hume and Smith implies multiple thought processes and a specific relationship between them. In Hume’s case, while we are often led to believe in the existence of external objects, we can sometimes reflect on this idea and detect its falsehood. Similarly, Smith argues that moral judgments are sometimes derived from utilitarian reflection, but in many cases, they are derived from sympathy. Smith’s criticism of Hume is that Hume mistakenly considered reflection on utility as the natural principle of moral judgment and misunderstood the relationship between utilitarian reflection and sympathy, reversing their positions.

It is true that, since Smith himself did not present any proper epistemological arguments apart from ‘Of the External Senses’, he might be considered less interested in that subject. However, Smith is said to have enthusiastically studied Treatises, especially while he was in Oxford, and we cannot overlook the view that Smith admired the idea of ‘The science of man’ and that the series of lectures he gave in Edinburgh from 1748–51 were partly inspired by it.35 If this is the case, it does not sound implausible that Smith, one of Hume’s closest friends, was impressed and inspired by the way the latter treated human nature.

A question that might arise is whether we cannot find the same element in Hume’s moral theory as discussed in Book III of Treatise. In other words, one may ask whether Smith borrowed the framework from Hume’s moral theory. To this, I would respond that although Hume appears to have a view that different thought processes affect moral judgment, it is uncertain whether this view plays a central role in his moral theory, as it does in his epistemology.

In some places, Hume seems to observe that multiple mental processes could influence us in moral matters. For instance, Hume argues that we may hold different attitudes towards those who have pride, depending on whether the principle of sympathy or that of comparison is dominant (T 3.3.2.2–7). Furthermore, at some points Hume appears to endorse a view similar to Smith’s; for example, he writes that ‘men, in the ordinary conduct of life, look not so far as the public interest, when they pay their creditors, perform their promises, and abstain from theft, and robbery, and injustice of every kind. That is a motive too remote and too sublime to affect the generality of mankind, and operate with any force in actions so contrary to private interest as are frequently those of justice and common honesty’ (T 3.2.1.11).

However, there is a significant difference between Books I and III. Hume does not seem to grant a fundamental role to the pluralism of mental processes in his moral theory. Instead of emphasising the diversity of mental principles, he appears to focus on reducing his arguments to a single principle. Thus, while claiming that pleasure can be derived from actions that tend to benefit society through sympathy (T 3.3.1.11), Hume also asserts that the same principle of sympathy explains our distinction between virtue and vice, which ‘may seem to flow from the immediate pleasure or uneasiness’ (T 3.3.1.29). Hume thus explains the four well-known sources of pleasure and pain underpinning our moral judgment from the single principle of sympathy. Here, we see a clear contrast between the conclusion of Book I, where Hume cannot help but confess his discomfort at facing contradictions among different mental processes, and the conclusion of Book III, where he, without hesitation, affirms that ‘sympathy is a very powerful principle in human nature’ (T 3.3.6.1). One might interpret Hume as a reductionist regarding the human mind in Book I, explaining every phenomenon from imagination; still, he had to face serious conflict between the natural and the philosophical states, which I think is missing in Book III.

Whether Smith’s representation of Hume’s moral theory is accurate remains debatable, especially regarding his description of the theory of utility in Part IV of TMS, which may be misleading as it suggests that reflection on utility is the only significant aspect of Hume’s moral theory. However, at least it can be safely said that Smith would not agree with Hume’s writing that ‘[m]ost people will readily allow, that the useful qualities of the mind are virtuous, because of their utility. This way of thinking is so natural, and occurs on so many occasions, that few will make any scruple of admitting it’36 (T 3.3.6.2).37

Let us assume that the argument made thus far has some validity. This not only opens up a new possibility for understanding the relationship between Smith and Hume but also has some implications for one of the most controversial passages in TMS: the sentinel example. As we have seen, Smith concedes that we sometimes make moral judgment ‘merely from a view to the general interest of society’ (TMS II.ii.3.11). An example is the punishment of a sentinel:

This part has not been viewed favourably by Smith scholars and is sometimes regarded as an exception in his theory.38 One of the most important statements on this may be Tom Campbell’s. Treating the sentinel example as an exception to Smith’s theory, he argues that ‘Smith admits more than he need do in this discussion of the laws of military discipline. There is no necessity for him to say that the sentinel’s punishment is just because men approve it’.39

In light of the foregoing considerations, it follows that the passage in question should be regarded not as an exception but rather as an essential element in Smith’s moral theory. Smith’s main concern, as read from his criticism of Hume, is to determine the natural principle of moral judgment. Since this view assumes multiple mental principles and is primarily concerned with their relationship, the principle of utility should also be regarded as an essential component of TMS.40 This is the same as Hume’s epistemological argument, in which the philosophical and reflective principle is as important as the natural one.

If we pursue this line of reading further, we could suggest a more radical and contentious interpretation of Smith’s scholarship. That is, rather than the sentinel example, another passage in TMS often cited for reading Smith’s intention of criticising Hume should be viewed as an exception. In Part IV of TMS, Smith writes ‘it will be found, upon examination, that the usefulness of any disposition of mind is seldom the first ground of our approbation; and that the sentiment of approbation always involves in it a sense of propriety quite distinct from the perception of utility’41 (TMS IV.2.5). This passage appears to contradict Smith’s recognition of the four sources of moral sentiments and the sentinel example. Lacking an interpretation to reconcile these, and considering the examination so far, I would like to propose a view that TMS IV.2.5 is an exception, which Smith overemphasises to distance himself from Hume. I adopt this view because Smith must have been familiar with Hume’s view on human nature, as discussed above, and also partly because I would like to give more weight to Smith’s description of the four sources of moral sentiment found at the end of TMS, agreeing with Charles L. Griswold’s observation that ‘The Theory of Moral Sentiments might be read as a story that unfolds in steps. Views that are propounded at the beginning may be refined substantially by the end, as would be natural in a literary dialogue or novel…’42.

The main aim of this paper is to argue that Smith’s critique of Hume in TMS is based on a particular view of human nature presented in Hume’s epistemology. This framework is to describe the human mind as composed of multiple principles, with one dominant principle called ‘natural’. In light of this understanding, I present an alternative interpretation of Smith’s treatment of utility. These analyses seem to support Nicholas Phillipson’s statement that ‘Smith became at Oxford as he remained, the perfect Humean’43. As this paper considers the intellectual connection between Hume and Smith in terms of the framework of human nature underlying their theories rather than specific ideas, it shows that Smith was influenced by Hume at a deeper level. For Smith, who opposed a kind of reductionism (TMS VII.2.14), Hume’s perspective on the diversity of mental processes might have seemed appealing and persuasive as a foundation for a moral theory that aims to explain the complex reality of our moral life.