Abstract
The inoffensive title of Section 1.4.7 of Hume’s Treatise of Human Nature, ‘Conclusion of this Book’, belies the convoluted treatment of scepticism contained within. It is notoriously difficult to decipher Hume’s considered response to scepticism in this section, or whether he even has one. In recent years, however, one line of interpretation has gained popularity in the literature. The ‘usefulness and agreeableness reading’ (henceforth U&A) interprets Hume as arguing in THN 1.4.7 that our beliefs and/or epistemic policies are justified via their usefulness and agreeableness to the self and others; proponents include Ardal (in Livingston & King (eds.) Hume: a re-evaluation, 1976), Kail (in: Frasca-Spada & Kail (eds.) Impressions of Hume, 2005), McCormick (Hume Stud 31:1, 2005), Owen (Hume’s reason, 1999), and Ridge (Hume Stud 29:2, 2003), while Schafer (Philosophers, forthcoming) also defends an interpretation along these lines. In this paper, I will argue that although U&A has textual merit, it struggles to maintain a substantive distinction between epistemic and moral justification—a distinction that Hume insists on. I then attempt to carve out the logical space for there being a distinctly epistemic notion of justification founded on usefulness and agreeableness. However, I find that such an account is problematic for two reasons: first, it cannot take advantage of the textual support for U&A; secondly, it is incompatible with other features of the text.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
In references to Hume’s texts throughout the paper, ‘THN’ refers to the Treatise of Human Nature, ‘EHU’ to the Enquiry Concerning Human Understanding, ‘EPM’ to the Enquiry Concerning the Principles of Morals, and ‘EMPL’ to Essays Moral, Political, and Literary. Arabic numerals refer to section and paragraph numbers (EHU and EPM), or to book, part, section, and paragraph numbers (THN). EMPL numbers refer to pages in the Miller revised edition of the Essays (Liberty Fund Inc., 1987).
There are of course differences among these accounts; for example, Ridge emphasises immediate agreeableness to oneself, while McCormick emphasises usefulness to society.
I do not wish to substantively define epistemic normativity in this paper on pain of begging the question against U&A. I will only roughly characterise it (hopefully uncontroversially) as the primary kind of evaluation to which beliefs are subject that determines their philosophical legitimacy.
Hume mentions philosophy more than he does science in THN 1.4.7, but his interest in the latter is confirmed by his claim that forbidding all abstruse reasoning will ‘cut off entirely all science and philosophy’ (THN 1.4.7.7). Indeed, it seems that for Hume, science encompasses philosophy; he describes his own Treatise (which he clearly considers philosophy) as an endeavour falling under the ‘science of man’ in his introduction to it.
This does not mean that we are to arrive at such an epistemic policy by means of reasoning about justification, as Garrett (2005) points out. Indeed, Hume seems quite insistent that reasoning cannot lead us to the correct epistemic policy, which explains why he is eventually led to the Title Principle via ‘a serious good-humour’d disposition’ rather than through ‘philosophy’ (THN 1.4.7.11). However, upon arriving at an epistemic policy, we should confirm its acceptability by reasoning about its justification.
Note that Hume typically takes ‘superstition’ to refer to a form of religion (EMPL 74); in thinking superstition dangerous, Hume probably had in mind religious wars.
These two criteria of ‘liveliness’ and ‘mixing with some propensity’ are related, as associated propensities can enliven reasoning. Going into the mechanics of such a process within Hume’s psychological framework is a complicated matter, and would take me too far afield.
Hume has already dismissed another alternative policy of assenting to reason except for where it is ‘refin’d or elaborate’ in THN 1.4.7.7.
Of course, these two positions are caricatures of Hume’s opponents. Sceptics will not always assent to reason, and superstitious people do of course sometimes assent to reason. But in attacking these caricatures, Hume is using exaggeration to highlight the problems with both these positions. Sceptics may not be left absolutely belief-free, but they will be left with far fewer beliefs than they should have. Superstitious folk will not so fully pursue their fancies as to completely avoid the judgments of reason, but they avoid it enough that they may be dangerous to society (for instance, they might avoid in-depth reasoning about whether a benevolent God would approve of Holy Wars).
That being said, one should not overstate the importance of utility to the passion of curiosity. As Gelfert (2013) points out, Hume thinks utility important to curiosity not because it enhances the passion, but merely because it provides an initial focus for our attention, which is then sustained by our enjoyment of the subsequent exercise of genius (THN 2.3.10.6).
Schafer (forthcoming) argues that curiosity and ambition are closely related; our intellectual ambition will only be fulfilled by reasoning that satisfies the curiosity of our ‘epistemic community’.
Hume’s essay ‘Of Superstition and Enthusiasm’ was written much later than the Treatise, of course; nevertheless, it seems plausible that Hume’s opinions on the nature of superstition did not change considerably in the intervening time.
This might be seen as begging the question against the sceptic, who will argue that ignorance (viz. a lack of knowledge) is inevitable whether or not reason is present. Clearly, Hume does not have the sceptic in mind in this passage, which appears in his essay ‘Of Superstition and Enthusiasm’; here, it should be uncontroversial to read Hume as thinking ignorance to result from a lack of reason. However, Hume holds that always assenting to reason also undermines knowledge, since it destroys our beliefs. The key is to find the right balance in assenting to reason, which is Hume’s concern in THN 1.4.7.
Given that Hume considers Catholicism to result from superstition, one might object that Catholic theologians mix their propensities with reason, and therefore superstition can arise when mixed with reason. Of course the claim that superstition arises from a lack of reason is an empirical generalisation, and admits of exceptions. However, the Catholic theologian need not be such an exception. We may distinguish two epistemic phases that the Catholic theologian undergoes: first, he abides by fear and melancholy untempered by reason (superstition) and comes to believe in Catholicism; then, he tries to use reason to justify this belief. His arriving at superstition is very much a product of his passions when unmixed with reason; obviously, his adopting reason later to attempt and justify his superstition does not produce his initial superstition, and so his superstition is partly a product of his lack of reason, though it may subsequently be mixed with reason.
Durland (2011, p. 81) and Winkler (1999, p. 211 fn. 21) argue that the Title Principle cannot recommend philosophy over superstition. As argued above, I think the Title Principle does recommend philosophy over superstition: since superstition is the product of a lack of reasoning, it cannot be an instance of lively reason mixing with a propensity. In correspondence, Don Garrett suggests that the Title Principle might be thought to rule out superstition in a different but similar way: superstition is not ultimately approved of by reason since it might, for example, violate reason’s rules by which to judge of causes and effects (THN 1.3.15), and so would not be an instance of lively (considered) reason mixed with a propensity.
Although philosophy and superstition both result from propensities, they result from different propensities: philosophy results from curiosity and ambition (along with reason), while superstition results from fear and melancholy (with weakness, in the absence of reason). This is because reason mixes most naturally with curiosity and ambition, therefore encouraging these propensities: a strong faculty of reason, frequently exercised, makes it easier to discover interesting and useful truths, and such successes encourage us to further pursue these propensities. As Hume notes: ‘curiosity… requires youth, leisure, education, genius, and example, to make it govern any person’ (EMPL 113); in other words, curiosity requires (among other factors) a strong and active faculty of reason (that is, genius) in order to be effective. The pleasure derived from the exercise of reason in accordance with curiosity and ambition dispels melancholy (as Hume narrates in THN 1.4.7), and it presumably dispels fear as well; moreover, the exercise of reason is capable of curing intellectual weakness.
Schafer (forthcoming) also ably defends a version of U&A that gives the Title Principle a central role, although the exact details differ from mine.
THN 1.4.7.14 and EHU 1.3 are admittedly theoretically consistent with Hume’s accepting beliefs only partly on the basis of their agreeableness and/or usefulness, as well as on the basis of other factors; however, this seems to run contrary to the tone of THN 1.4.7.14 especially. Moreover, THN 2.3.2.3 and EHU 8.26 seem fairly clear that even partially judging beliefs on the basis of their dangerousness is unacceptable.
A quick objection that might be raised here is that Hume believes actions can be unreasonable when mixed with a false judgment (THN 3.1.1.12). However, it seems clear from this passage that Hume does not think that passions and actions can be genuinely reasonable or unreasonable: they can be at best unreasonable ‘in a figurative and improper way of speaking’ (ibid.). Hume repeats a similar claim about the passions being derivatively unreasonable in THN 2.3.3.6, and again emphasises that ‘even then’ tis not the passion, properly speaking, which is unreasonable, but the judgment.’
I take THN 2.3.3 to be highly relevant to the interpretation of THN 3.1.1, since Hume’s argument in THN 3.1.1 (that morals cannot be derived from reason, since reason alone cannot motivate while morals can), turns on the result of Hume’s argument in THN 2.3.3 (that reason alone cannot motivate); indeed, Hume refers back to THN 2.3.3 in a note in THN 3.1.1.8.
Of course, Hume does argue that reason cannot motivate as well, which also tells against the view that we should be guided by reason rather than the passions. My point here is merely that this normative aim of Hume’s makes it perfectly natural for him to also claim that reason and passions possess different forms of normativity, therefore explaining why he seems to make such a claim in the course of this argument.
There is an interpretive controversy about whether Hume uses ‘reason’ and ‘the understanding’ interchangeably. Millican (2002) argues for the equivalence thesis, but Garrett (1998) disagrees, thinking that the understanding encompasses reason along with intuition; note that Garrett reads ‘reason’ more narrowly than Millican does. In any case, on either reading, it is true that reason cannot justify or condemn the passions, on the plausible enough assumption that a subset of the understanding cannot justify or condemn something if the understanding itself cannot do so.
My treatment of this passage owes a great deal to very helpful discussions with Karl Schafer and Cian Dorr.
Hume goes on to claim that these errors in belief ‘extend not beyond a mistake of fact, which moralists have not generally suppos’d criminal, as being perfectly involuntary’ (THN 3.1.1.12). One should be careful not to read Hume as straightforwardly citing the involuntariness of belief as the reason for this innocence, as Ridge (2003, p. 172) does. Hume unreservedly endorses the claim that false beliefs are not morally blameable while being careful to put the ‘involuntariness’ basis for this claim in the mouths of ‘moralists’—note also the hedging terms ‘commonly’ and ‘generally suppos’d’. Hume wields this argument from involuntariness as an ad hominem attack against his rationalist opponents, who typically accept ‘Ought implies (voluntary) Can’ while maintaining that moral turpitude derives from false beliefs. Indeed, were Hume to agree with the moralists on this matter, this would create a tension with his claim in THN 3.3.4.3 that natural abilities are subject to moral appraisal despite being involuntary. I argue that Hume rejects ‘Ought implies (voluntary) Can’ in my “Hume’s Doxastic Involuntarism” (work in progress) (Qu 2013).
In correspondence, Cian Dorr and Karl Schafer independently object that Hume only states that errors of belief are ‘commonly’ very innocent, which leaves open the possibility that they are sometimes not innocent. They propose that in THN 3.1.1.12 Hume is merely pointing out the implausibility of errors in beliefs that are related to actions being the sole source of moral blameworthiness given that such beliefs are usually blameless, without committing himself to the thesis that beliefs are not subject to any moral appraisal whatsoever. However, Hume’s statement that ‘No one can ever regard such errors as a defect in my moral character’ is not qualified in the same way, and it seems that his point is the stronger one that beliefs are not subject to moral appraisal, as argued above.
It might be objected that in judging mistaken beliefs to be ‘innocent’ and to ‘draw no manner of guilt’, Hume is passing a moral verdict on them, and so beliefs are subject to moral appraisal. Against this, it may be replied that the adjective ‘innocent’ is commonly used to indicate that a creature is not subject to moral appraisal, for instance when we describe small children and animals as ‘innocent’. A toddler may perform mischievous acts; when we describe her as innocent we do not mean that she is a paragon of virtue, but rather that, being undeveloped in many morally relevant ways, she is not properly an object of moral judgments.
It may be objected that Hume denies a distinction between epistemic and moral normativity in THN 3.3.4, wherein he argues against there being a substantive distinction between natural abilities and moral virtues. I will address this objection later in Section 6b of this paper; indeed, rather than hindering my position, this passage ultimately reinforces it.
Hume uses ‘sentiments’ frequently in his philosophical works in a variety of ways. He often uses it to indicate emotions, but he also often uses it to indicate beliefs or opinions (e.g. THN 1.2.2.3; THN 1.3.13.14; THN 1.3.14.12; THN 1.4.2.14; THN 1.4.2.50, etc.) as well. The context of THN 1.4.7.11 clearly clarifies that Hume means by ‘sentiments of my spleen and indolence’ his belief that he ‘cannot satisfy [himself] concerning the reasonableness of so painful an application [of reasoning and philosophy], nor have any tolerable prospect of arriving by its means at truth and certainty’ (THN 1.4.7.10).
One possibility is that Hume is merely being disingenuous when he speaks of beliefs being justified, or when he continues on his positive naturalistic project. Broughton (2004) suggests a view in this spirit, arguing that Hume continues post-THN 1.4.7 in a detached or ironic manner. A full discussion of this interpretation would take me too far afield, but briefly, I find that interpretation along these lines come with tremendous interpretive costs in treating Hume as speaking somewhat insincerely in such a wholesale manner.
Schafer argues that THN 1.4.7 provides epistemic justification, but suggests that for Hume, epistemic justification is a subset of moral justification. He says: ‘An intellectual trait is an epistemic virtue just in case it receives the approval of the “moral sense”… because it tends to satisfy the curiosity and ambition of the believer and those in his "epistemic community" under normal circumstances.’ Insofar as receiving approval from the moral sense seems to entail moral justification, it seems that intellectual traits are epistemically justified (that is, virtuous) insofar as they are morally justified in a particular way, which points to the subset view. I think the subset view is problematic in light of THN 3.1.1.12, as I argued previously in this section.
Loeb (2002) argues that Hume is fundamentally committed to a dispositional account of beliefs, although his official theory treats them as occurrent. I agree with Marusic (2010) that this is unsatisfactory because it fails to make sense of the causal role that Hume thinks beliefs play in his psychology.
Hume’s taking actions to fall under the domain of moral normativity despite beliefs also contributing to the production of actions doubtless stems from his hyperbolic dismissal of reason’s importance in the production of actions: ‘Reason is, and ought only to be the slave of the passions, and can never pretend to any other office than to serve and obey them’ (THN 2.3.3.4). Given this characterisation of motivation, it is unsurprising that Hume takes the production of actions as falling under the jurisdiction of the passions.
In correspondence, Don Garrett and Lisa Downing independently point out the possibility of the primary object of epistemic evaluation being beliefs rather than the agents who possess them. I think that this possibility seems unlikely for a few reasons. First, beliefs are too transient to provoke the sentimental response that Hume thinks essential to normativity, and so it seems that belief-forming dispositions must be involved; once belief-forming dispositions are involved, it is natural to think that the persons who possess them should be the primary objects of epistemic evaluation, for the reason given above. Moreover, epistemic evaluations depend on factors beyond the beliefs themselves, e.g. the grounds for holding them. No contingent beliefs are in themselves unreasonable (one can always construct some complicated backstory that makes any non-necessary belief more or less reasonable); in thinking someone unreasonable for holding a particular belief, we are really thinking him unreasonable for holding this belief on some grounds or the other. Holding a belief on grounds is something only agents can do, and so it seems that the primary object of epistemic evaluation has to be believers rather than beliefs. Arguably, the grounds on which we hold a belief are only informatively significant insofar as they reveal one’s underlying epistemic character, in the same way in which the grounds on which we perform a virtuous action (viz. our intentions) are important (THN 2.2.3.3), but only insofar as they signify underlying moral character. Ridge (2003, p. 180) agrees that for Hume, epistemic appraisals concern durable dispositions: he argues that ‘if Hume is making a moral argument in “Conclusion of this book” it must be an argument in favor of some character trait’, and furthermore takes the ‘senses and understanding’ (THN 1.4.7.10) as ‘an adequate characterization of the durable principle of mind [Hume] aims to defend’.
In correspondence, Don Garrett suggests that the proponent of U&A could distinguish moral normativity from epistemic normativity by the differing sentiments they evoke in the assessor: moral normativity evokes moral approbation or blame, while epistemic normativity evokes epistemic approbation or blame; there need not be any other distinguishing fact of the matter. This seems implausible to me. Say that for U&A, X, Y and Z really are the same in moral and epistemic judgments (and I later argue that this is in fact the case). In that case, there would be no basis for moral and epistemic judgments involving different sentiments where X, Y and Z are identical in both cases. Surely such a distinction in our sentiments evoked, when so widespread and systematic, should correspond to some external difference. Even if such a systematic distinction in our sentiments should be entirely baseless, it remains mysterious why Hume would insist on this distinction (as argued in the previous section), rather than merely dismissing it as groundless, or ignoring it in his philosophy. It thus seems likely that Hume envisions some external difference between moral and epistemic normativity, and therefore it seems that one of X, Y or Z must differ for U&A to adequately distinguish moral from epistemic normativity. Drawing a parallel to the impression of necessary connection might prove enlightening here: although this impression does not correspond to anything external to us (because it is an internal impression, like normative sentiments), it is nevertheless responsive to external factors; this is made evident by Hume pointing out in his ‘Rules by which to judge of causes and effects’ (THN 1.3.15) the external criteria that a veridical impression of necessary connection should accord with. Similarly, it seems that any distinction between specifically moral and epistemic sentiments should correspond with a distinction between some external moral and epistemic factors, i.e. the underlying disposition that is evaluated, the normatively relevant properties of this disposition, or the possessor of this disposition.
Although some qualities may be immediately agreeable to the individual even without causing any actions (cheerfulness, for example), these are not the norm.
This is more controversial than assuming a backdrop of true beliefs, and those with more pessimistic outlooks on human nature will probably disagree with this presupposition, but Hume, with his sunny view of human morality as driven by natural sympathy, would probably be unmoved by such pessimism.
I discuss the justificatory role of custom with respect to Hume’s scepticism about induction in the Enquiry in more detail in my “Hume’s Positive Argument on Induction” (forthcoming).
As mentioned in footnote 22, there is a controversy about whether for Hume, ‘reason’ is equivalent to ‘the understanding’. On either reading, it is true that philosophy involves reason while superstition does not.
In correspondence, Cian Dorr argues that there is a narrower sense of belief-forming dispositions according to which philosophy and excessive scepticism do not share the same belief-forming dispositions: philosophy will include a belief-forming disposition to form beliefs such as the belief that there is an oncoming bus, while excessive scepticism (the extinction of all our beliefs) will not. Nevertheless, Hume individuates belief-forming dispositions in a wider sense in THN 1.4.7 in terms of the belief-forming disposition of reason, which both philosophy and excessive scepticism share. Moreover, it seems clear from Hume’s dialectic that the point of contrast Hume wishes to draw is between the passions of curiosity and ambition in the case of philosophy (which subsequently enliven beliefs) and their absence in excessive scepticism, and so it seems that the justificatory role is played by the passions rather than by belief-forming dispositions construed in this narrow way.
References
Ardal, P. (1976). Some implications of the virtue of reasonableness in Hume’s treatise. In D. Livingston & J. King (Eds.), Hume: A re-evaluation. New York: Fordham University Press.
Broughton, J. (2004). The inquiry in the Treatise. The Philosophical Review, 113, 4.
Durland, K. (2011). Extreme skepticism and commitment in the Treatise. Hume Studies, 37, 1.
Garrett, D. (1997). Cognition and commitment in Hume’s philosophy. Oxford/New York: Oxford University Press.
Garrett, D. (1998). Ideas, reasons, and scepticism: Response to my critics. Hume Studies, 24(1), 171–194.
Garrett, D. (2005). Hume’s conclusions in ‘Conclusion of this Book’. In Saul. Traiger (Ed.), The Blackwell guide to Hume’s treatise. Malden, MA: Blackwell Publishing.
Gelfert, A. (2013). Hume on curiosity. British Journal for the History of Philosophy, 21(4).
Hume, D. (1739–1740). In L. A. Selby-Bigge (Ed.), A treatise of human nature (2nd ed., revised by P. H. Nidditch, 1975). Oxford: Clarendon Press.
Hume, D. (1748). In L. A. Selby-Bigge (Eds.), Enquiries concerning human understanding and concerning the principles of morals (3rd ed., revised by P. H. Nidditch, 1975). Oxford: Clarendon Press, 1975.
Hume, D. (1777). Essays moral, political and literary (revised ed., E. Miller, 1987). Indianapolis: Liberty Fund Inc.
Kail, P. (2005). Hume’s ethical conclusion. In M. Frasca-Spada & P. Kail (Eds.), Impressions of Hume. Oxford: Oxford University Press.
Loeb, L. E. (2002). Stability and justification in Hume’s treatise. Oxford: Oxford University Press.
Marusic, J. S. (2010). Does Hume hold a dispositional account of belief? Canadian Journal of Philosophy, 40, 2.
McCormick, M. (2005). Why should We be wise? Hume Studies, 31, 1.
Millican, P. (2002). Hume’s sceptical doubts concerning induction. In P. Millican (Ed.), Reading Hume on human understanding. Oxford: Clarendon Press.
Owen, D. (1999). Hume’s reason. Oxford: Oxford University Press.
Qu, H. (2013). Hume’s doxastic involuntarism. Unpublished manuscript.
Qu, H. (Forthcoming). Hume’s positive argument on induction. Nous. Available at http://onlinelibrary.wiley.com/doi/10.1111/nous.12015/full.
Ridge, M. (2003). Epistemology moralized: David Hume’s practical epistemology. Hume Studies, 29, 2.
Schafer, K. (Forthcoming). Curious virtues in Hume’s epistemology. Philosophers’ Imprint.
Winkler, K. (1999). Hume’s inductive scepticism. In M. Atherton (Ed.), The empiricists: Locke, Berkeley, and Hume. Lanham: Rowman and Littlefield Publishers Inc.
Acknowledgments
An overwhelming debt of gratitude is owed to Don Garrett, whose careful comments on multiple drafts of this paper were tremendously useful and agreeable, and to Karl Schafer, for excellent comments and discussions. I am also very grateful to Peter Millican and Axel Gelfert for very helpful comments on earlier drafts. Thanks are also owed to Amyas Merivale, Lisa Downing, Cian Dorr, Martín Abreu Zavaleta, Ian Grubb, Erica Shumener, Zee Perry, Max Barkhausen, Camil Golub, and Asya Passinsky for helpful discussion. For excellent questions and comments, I am also indebted to an audience at the South Central Seminar in Early Modern Philosophy, held at Texas A&M. Thanks also to Alex Worsnip for help with the title of this paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Qu, H. Hume’s practically epistemic conclusions?. Philos Stud 170, 501–524 (2014). https://doi.org/10.1007/s11098-013-0260-1
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-013-0260-1