Is Hume a Perspectivalist?

1. Introduction

Hume notoriously pursues a constructive science of human nature in the Treatise while raising serious skeptical doubts about that project and leaving them apparently unanswered. On the perspectivalist reading, Hume endorses multiple incommensurable epistemic perspectives in the Treatise. This reading faces two significant objections: that it renders Hume’s epistemology inconsistent (or at least highly incoherent) and that it is ad hoc. In this paper, I propose a perspectivalist account of epistemic justification in the Treatise that addresses, to a significant degree, these concerns. Hume has available to him an account—what I will call epistemic dispositionalism—that is both internally consistent and allows for epistemic continuity between perspectives. In addition, far from being ad hoc, this account is well grounded in Hume’s cognitive psychology and coheres with the spirit of Hume’s naturalism. I will not be endorsing this view here, which would require arguing for its fittingness over the most plausible non-perspectivalist readings. I only seek to show that by answering the main charges against it, it can be raised to the level of a plausible contender. I will also add that here I am only concerned with the epistemology of the Treatise. I am moved by arguments that Hume’s commitments in the Treatise differ from those in the Enquiry or Dialogues.¹ So I will limit myself to claims about the Treatise and resist the temptation to draw support from these other works.

Perhaps the leading proponent of the perspectivalist reading is Robert Fogelin:

It seems, then, that Hume’s writings exhibit a radical form of epistemological, or better, doxastic perspectivism. What we believe and what we think it appropriate to believe is a function of the level of investigation we are indulging in… Hume does not simply describe these perspectives; he actually presents his views from within the confines of one of them. (1998: 164; see also 1993)

According to Fogelin, Hume traverses between three perspectives: gentlemanly Hume, wise Hume, and Pyrrhonian Hume. Fogelin answers the problem of reconciling Hume’s naturalistic project with his skeptical doubts by denying that it is a problem at all. The pantheon of great philosophers, of which Hume is a member, contains only those whose systems are rich enough that they must be “deeply impregnated with inconsistency” (1998: 161). To force consistency upon Hume on pretense of charity, therefore, would actually be to cheapen his philosophy and assail his legacy.

A more recent expression of the perspectivalist² account comes from Donald Baxter (2006; 2008; 2018), who is explicitly concerned with absolving Hume of charges of inconsistency. According to Baxter, Hume avoids inconsistency by distinguishing between passive assent and active belief, following Pyrrhonian skeptics. Though we cannot be alethically justified (i.e. have reason to think a proposition is true or probably true) in the belief in the products of reason and in external objects, we are instinctually, and nearly irresistibly, coerced into assenting passively. But unlike the Pyrrhonians, according to Baxter, Hume thinks that this natural assent can cover theoretical propositions, such as those that compose the science of human nature (2018: 390).

The perspectivalist interpretation has the advantage that it is uniquely licensed to take the text at face value. The passages that seem highly skeptical are highly skeptical, and likewise for those apparently naturalistic passages. Here, the skeptical interpreter³ is forced to account for why Hume continues philosophizing after (and even before) the skeptical arguments in Book 1, and to explain exactly what kind of normativity can be applied to Hume’s positive claims. The naturalistic interpreter,⁴ on the other hand, has the burden of reinterpreting these apparently skeptical arguments and explaining exactly what kind of epistemic justification is consistent with them. This latter question is common to all interpretations that read Hume as allowing the claims of the science of human nature to have epistemic justification, including perspectivalism, and so bears pausing on.

For any reading on which Hume is not an extreme skeptic, the commentator must accept the following unsavory fact. There can be no plausible Humean theory of epistemic justification insofar as such a theory connects some property of a belief or belief-forming mechanism to truth. If justification tracks or tends to track truth, it does so only incidentally. I will use this term ‘truth-tracking’ repeatedly. For justification of a belief or belief-forming process to be truth-tracking, it must be the case that a higher degree of justification corresponds to a higher probability of truth. Although Hume does discuss knowledge and perhaps thinks it is possible,⁵ Hume, following Locke, uses the term in a special sense. He defines ‘knowledge’ as “that evidence, which arises from the comparison of ideas” (T 1.3.11.2).⁶ But Hume isn’t interested in relations of ideas insofar as he is attempting to introduce the experimental method into moral subjects. Experimental philosophy employs causal reasoning (T 1.3.15.11). Although Hume claims that reason simpliciter is naturally oriented towards truth (T 1.4.1.1), causal reasoning, for Hume, requires a non-rational element provided by the imagination (“custom” or “habit”) that, as it were, breaks the truth-tracking chain.⁷ This is because the problem of induction has uncovered that it is not rational (in the strict sense of not being wholly determined by reason) to believe that past or observed regularities will hold in future or unobserved cases. And so, we have no reason to think that causal reasoning—reasoning concerning matters of fact that goes beyond what’s immediately present to the senses—tracks truth. Therefore, in order to even have causal beliefs, their source must have some non-rational element. And causal reasoning makes up the bulk of the science of human nature—including Books 2 and 3 of the Treatise. Thus, if knowledge in the non-technical sense in which we use it is (at least) justified, true belief, and the role of justification in a theory of knowledge for an internalist^8,9 is as an internally accessible mark that a belief is true or probably true, and Hume thinks causal reasoning can be justified despite our having no reason to think it is truth-tracking, then Hume must believe that, at least in the domain of causal belief, justification does not track truth (or it does so only incidentally).¹⁰ And so, knowledge of this kind is impossible.

The positive upshot of this for the perspectivalist is that she need not explain which of the set of incommensurable epistemic standards is the one that tracks truth, and so bypass the further question of why one is not the sole standard. The more general upshot for those interested in finding a theory of epistemic justification in Hume’s Treatise is this: a theory of epistemic justification in Hume functions to take stock of the set justified belief-forming principles Hume actually accepts,¹¹ and reconcile this with the skeptical doubts generated by following these principles themselves. One who searches for a way for Hume to ground that set of principles in truth will come up short.

This theme will reemerge in §4, but for now let’s return to our topic. We have seen that the great interpretive advantage of the perspectivalist reading is its ability to take Hume’s text at face value. But a price is paid for this in the currency of philosophical plausibility. Hsueh Qu says of these interpretations generally that “reading Hume as committed to irreconcilable viewpoints that are nevertheless equally valid seems significantly to compromise the overall coherence and systematicity of Hume’s project” (2020: 135). In Fogelin’s account, this comes in the form of straightforward inconsistency—a feature whose virtue, whatever Whitmanian romanticism it might add, we may justly dispute. In Baxter, as we will see in more detail below, this comes in the form of epistemic disjunctivism regarding grounds of assent. If skeptical Hume uses truth as his normative standard, and natural Hume uses the psychological features of instinct and irresistibility, then we are imposing upon Hume a quite radical form of epistemic pluralism.

There is the further though related problem that the interpretation appears ad hoc. If the considerations in favor of perspectivalism are really just that (i) it allows us to take the text at face value and (ii) it is an answer, of sorts, to what Phillip Cummins (1999) dubbed the integration problem: the problem of reconciling Hume’s constructive naturalism with his destructive skepticism; then maybe we really ought to reinterpret those problematic passages, or even accept the integration problem as an unfortunate flaw in Hume’s philosophy. Such a prima facie strange epistemology ought to be rooted in or bolstered by other parts of Hume’s system. It should not be a wall nailed in at an angle to make the house look straight.

I develop here a perspectivalist account that avoids these problems. On this interpretation, Hume describes at least three doxastic dispositions, as I will call them: the vulgar, the true philosopher, and the extreme skeptic. A doxastic disposition is constituted by (i) active belief-forming principles, (ii) higher-order corrective tendencies, and (iii) emotional disposition. Epistemic justification, for Hume, reduces to the psychological feature of believability. A belief is epistemically justified if and only if it is believable within the philosophical or skeptical doxastic disposition and one is in that disposition. A belief is epistemically unjustified in all other cases (e.g. when believed within the vulgar disposition). In the skeptical disposition, it may be the case that the only ideas that are believable—and so justified—are those about occurrent mental states. The philosophical disposition is characterized by dominant philosophical belief-forming principles, higher-order general rules that correct the more precipitous first-order judgments, and minimal influence of the passions. The skeptical disposition is constituted by extreme consistency in application of higher-order, corrective principles and faculty-justificatory reflection that lead to wide-ranging suspension of judgment. Shifts between dispositions are not directly under the control of the will, but can be precipitated by an increase or decrease in philosophical reflection.

In building an account that answers the two charges discussed above—inconsistency/incoherence and lack of textual support—it will be necessary to address the latter first. Then, once our perspectivalist account is grounded in Hume’s cognitive psychology, we will see how the account is internally consistent and coherent with the mechanics and spirit of Hume’s system.

2. Dispositions to Believe

On the perspectivalist account I will be presenting, the criterion of epistemic justification for Hume is believability—that an idea would become a belief, given a certain set of background contents, within certain psycho-epistemic dispositions of the believer. This will need to be qualified and explained. But first it is necessary to explain what I mean by believability as a psychological property. Then I will show how we can understand believability as epistemically-laden for Hume.

A belief is a certain kind of idea, for Hume. A belief¹² is a lively (or vivid, forceful, strong, solid)¹³ idea, paradigmatically associated with a present impression by a relation of experienced constant conjunction.¹⁴ There is a kind of economy of force or vivacity in Hume’s cognitive psychology.

[It is] a general maxim in the science of human nature, that when any impression becomes present to us, it not only transports the mind to such ideas as are related to it, but likewise communicates to them a share of its force and vivacity. (T 1.3.8.1)

The liveliness of the belief is conducted from the liveliness of the impression in accordance with certain principles of association—paradigmatically, cause and effect. Belief comes in degrees, and the level of credence in a proposition corresponds to the quantity of vivacity of the idea with that content. Degrees of vivacity—or levels of confidence—in an idea/proposition can be influenced by a variety of factors, including the emotional and cognitive dispositions of the agent. Hume mentions that a coward will give more credence to claims about dangers that might affect him than a non-coward, since these excite his fear, and the vivacity of this passion is naturally conducted to those ideas (T 1.3.10.4). For example, he will be more confident that he will be bit by a rattlesnake when told he is hiking in rattlesnake country than a non-coward, ceteris paribus.

Hume describes several different doxastic dispositions in the Treatise. A doxastic disposition is constituted by (i) active belief-forming principles, (ii) emotional disposition, and (iii) higher-order corrective tendencies. We have already seen how one’s emotional disposition can affect belief formation. Passions act as a source of input vivacity that, as it were, travels along the circuits of belief-forming principles to one’s ideas, augmenting their vivacity/credence. The other two constituents of doxastic dispositions are the belief-forming principles themselves, and the hierarchical way that these principles interact to enable or prevent the enlivening of ideas to belief. These features will be taken in turn.

The two most important types of belief-forming principles for Hume are principles of the imagination and general rules. Hume distinguishes between two senses of ‘imagination’—the narrower sense being the faculty that influences the formation of ideas that are not formed by sense experience, memory, or the understanding (T 1.3.9.19, fn 22). Although it is prima facie strange for Hume to epistemically endorse imaginative principles in this narrow sense, especially given that he seems to think that principles of the imagination at least partially determine the distribution of vivacity to all beliefs (T 1.3.13.11), this is one of the central results of his cognitive psychology in the Treatise. As we saw above, ‘habit’ or ‘custom,’ the enablers of all causal judgments, have their source in the fancy (or narrow-imagination) and so are non-rational. Although cause and effect reasoning is the paradigmatic form of probabilistic reasoning:

All probable reasoning is nothing but a species of sensation. ‘Tis not solely in poetry and music, we must follow our taste and sentiment, but likewise in philosophy. (T 1.3.8.12)

And:

All our reasonings concerning causes and effects are deriv’d from nothing but custom; and […] belief is more properly an act of the sensitive, than of the cogitative part of our natures. (T1.4.1.8, Hume’s emphasis)

This is not to say that the reason has no share in belief; only that reasoning must be mixed with (i) the input energy of the present impression and (ii) the non-rational forces of the imagination in order to produce belief.¹⁵ Recall that this is why justification cannot be truth-tracking, and so why Hume cannot distinguish between, say, scientific experimentation and palm-reading on the basis of their conduciveness to truth.

Recall the three prominent doxastic dispositions of the Treatise: the vulgar, the true philosopher, and the extreme skeptic.¹⁶ Despite the infamy of the extreme skeptical arguments in the Treatise, most of the epistemically normative content of Book 1 concerns distinguishing between vulgar and philosophical doxastic practices. Regarding the principles of the imagination, the bulk of this work comes in Part 3, “Of knowledge and probability,” specifically Hume’s distinction between philosophical and unphilosophical probability (i.e. principles for forming probabilistic beliefs). Philosophical probability is “receiv’d by philosophers, and allowed to be reasonable foundation of belief,” while unphilosophical probability has “not had the good fortune to obtain the same sanction” (T 1.3.13.1).¹⁷ An important example of a philosophical principle of the imagination concerns the way the mind responds when presented with a “contrariety of causes,” that is, when one experiences exceptions to an otherwise constant conjunction.

‘Tis evident, that when an object is attended with contrary effects, we judge of them only by our past experience, and always consider those as possible, which we have observ’d to follow from it. And as past experience regulates our judgment concerning the possibility of these effects, so it does that concerning their probability; and that effect, which has been the most common, we always esteem the most likely. (T 1.3.12.8)

We may consider this a subtype of the rule that one ought to proportion one’s belief to the evidence. When an A-type event has been followed by both B- and C-type events, we proportion our belief that a B will follow a new A according to its past frequency. From this general principle, Hume derives a specific form that is the engine for his most destructive skeptical argument in the Treatise—the infamous argument at T 1.4.1 that appears to undermine the epistemic justification for all beliefs formed on the basis of inference, demonstrative or probabilistic. Here is the specific form:

In every judgment, which we can form concerning probability, as well as concerning knowledge, we ought always to correct the first judgment, deriv’d from the nature of the object, by another judgment, deriv’d from the nature of the understanding. (T 1.4.1.5)

This can be reformulated as follows:

probabilistic control principle: for every inference P, demonstrative or probabilistic, we ought to form a judgment P* regarding our reliability in performing inferences of P-kind, and adjust our credence in P accordingly. (T 1.4.1.1, 1.4.1.5, 1.4.1.6, 1.4.1.9)

Because we have made errors in reasoning in the past, reason, considered as a cause (T 1.4.1.1), must be regarded as producing contrary effects: sometimes truths, sometimes falsehoods. So in every inference we ought to proportion our credence in the output to our reliability in performing that kind of inference. Hume thinks repeated application of this wide-scope principle (see “every judgment” above) would cause reason to completely subvert itself, were the mind capable of the focus necessary to carry it out. This is because our confidence in the second-order check on our first-order judgment ought also to be reduced by a third-order judgment about its (the second-order judgment’s) reliability. And this diffidence trickles down to the first-order judgment. Because there is no non-arbitrary stopping point to this iterative review process, Hume thinks that consistent, higher-order application of the probabilistic control principle ought to occur in infinitum (T 1.4.1.9), eventually sapping the first-order judgment of all credence.

It is important here to highlight the significant role Hume’s skeptical argument against reason at T 1.4.1 plays in the epistemology of the Treatise. Not only is it highly destructive on its own, but it also features in Hume’s dangerous dilemma—the skeptical concern that precipitates his famous panic attack in the conclusion of Book 1 (T 1.4.7.6-7). Here, Hume is faced with the dilemma of either letting the imagination or the understanding rule belief. If the imagination rules, all false philosophy and superstition is licensed. If the understanding rules, all belief is subverted (via the skeptical argument against reason at T 1.4.1). And if we try to avoid the skeptical regress by letting the understanding rule but abstaining from “refin’d or elaborate reasoning,” we cut off the possibility of science and philosophy (T 1.4.7.7). And so, melancholy and delirium ensue.

One might think that the difference between the vulgar and philosophical doxastic dispositions is that they are constituted by unphilosophical and philosophical principles of the imagination, respectively. But Hume’s position is more complex. First, Hume thinks that there are specific kinds of beliefs and belief-forming processes that are universal. For example, nature has made it impossible for us to fail to believe in the existence of bodies, and likewise impossible not to believe the products of reason (T 1.4.2.1). These features of the mind govern both the wise person and the fool. Second, Hume accepts that everyone is, to some degree, vulnerable to unphilosophical belief formation. But for the wise, unphilosophical principles have only a pro tanto influence on belief formation. Thus, with respect to belief-forming principles, the vulgar disposition is characterized by dominant unphilosophical principles of the imagination. This dominance exists either because philosophical principles of the imagination are absent, or if present, unable to oppose the force of the unphilosophical ones.

The philosophical doxastic disposition, on the other hand, is characterized by dominant philosophical principles of the imagination. Philosophical principles of the imagination can be employed either on their own at the object-level, or, sometimes, deployed as second-order, reflective principles. For example, when faced with a prima facie exception to a causal pattern, the vulgar will rashly conclude that there is objective uncertainty (randomness) in events, while the philosopher will, reflecting on past cases in which previously unknown causes were later found, think it’s possible that there exist, in this case, too, unknown causes rather than “contingency in the cause” (T 1.3.12.4-5). Hume’s “rules by which to judge of causes and effects” (T 1.3.15) are just these kinds of second-order principles discovered by reflection—rules by which we can know when something really is the cause of another (T 1.3.15.2), i.e., rules to distinguish accidental circumstances from efficacious causes (T 1.3.13.11).

This leads us to the third constituent of doxastic dispositions. Recall that a doxastic disposition is constituted by (i) active belief-forming principles, (ii) emotional disposition, and (iii) higher-order corrective tendencies. These corrective tendencies determine if, when, and how the believer applies reflective scrutiny to her lower-order beliefs.

Here it is necessary to introduce Hume’s notion of general rules. He uses this term throughout the three books of the Treatise. The term encompasses various kinds of principles at work in the mind. The important use to which Hume puts ‘general rules’ for our purposes is any belief-forming principle.

General rules vary in degree of intellectualization. There are those general rules that “precede reflection and which cannot be prevented by it” (T 1.3.13.8). Hume seems to think these can become built into the senses so that our perception is actually structured by these principles (T 2.2.8.5). At this level, general rules operate without the consent of our will, at the level of what we now call the subpersonal. The most basic causal reasoning happens at this level (Hume calls it “the first influence of general rules” [T 1.3.13.12]). Then there are the higher-order, intellectualized general rules that are both the products of reflection and applied via reflection on the application event of first-influence general rules. Sometimes the pre-reflective and reflective rules conflict, that is, give different answers to the question of which belief (if any) will be formed given some input. Of this kind of scenario, Hume writes:

When we take review of [the first influence of general rules], and compare it with the more general and authentic operations of the understanding, we find it to be of an irregular nature, and destructive of all the most establish’d principles of reasoning; which is the cause of our rejecting it. This is a second influence of general rules, and implies the condemnation of the former. Sometimes the one, sometimes the other prevails, according to the disposition and character of the person. The vulgar are commonly guided by the first, and wise men by the second. (T 1.3.13.12)

Here, Hume is identifying an important difference between the vulgar and philosophical doxastic dispositions. The vulgar’s beliefs are typically determined by first-order, unreflective general rules, while the philosopher submits the application of these rules to interrogation by certain general rules of reflection—i.e. “the more general and authentic operations of the understanding.” Hume uses the example of the connection between causes and effects (T 1.4.3.9). The vulgar, in their “common and careless way of thinking,” believe there is a connection between things constantly conjoined since custom has made it difficult to separate them in their minds. But philosophers, “who abstract from the effects of custom,” are able to separate elements of experience in the mind and discover that there is no such connection. He makes a similar distinction between the vulgar and the philosopher on the continued existence of unperceived objects (T 1.4.2.14).

While in the philosophical disposition moderate higher-order reflective activity protects against rashness, superstition, and other forms of epistemic vulgarity, high reflective activity can destroy the philosophical disposition altogether, effecting a transition into the skeptical disposition. Recall the general philosophical principle regarding the contrariety of causes. We proportion our belief that a certain event will follow another according to its frequency of following the first kind in past experience. Hume says this can function as a higher-order regulative principle (T 1.3.12.7). Recall the specific version in T 1.4.1:

probabilistic control principle: for every inference P, demonstrative or probabilistic, we ought to form a judgment P* regarding our reliability in performing inferences of the P-kind, and adjust our credence in P accordingly. (T 1.4.1.1, 1.4.1.5, 1.4.1.6, 1.4.1.9)

For the scientist, or even the average person doing her taxes, one or two applications of this principle results in healthy epistemic caution. But repeated, iterative application of this principle (Hume thinks) would sap inferentially-formed beliefs of all credence/vivacity. The important point is that Hume’s skeptical arguments, and the skeptical doxastic disposition that can temporarily result, are constituted by the same belief-forming norms as the philosophical disposition. These are just the standards of the scientist of human nature. The difference is consistency and domain of application. Excessive application of philosophical principles of the imagination in epistemological inquiry can effect a transition to the skeptical disposition.

The skeptical doxastic disposition is characterized by a short-term, involuntary suspension of believability within some very general doxastic domain as a result of epistemological reflection. Examples of general doxastic domains are those beliefs that are the products of probabilistic reasoning, demonstrative reasoning, the senses, and memory. In cognitive psychological terms, to suspend believability is to block the channels between impressions and the ideas that might be enlivened by their vivacity.

Hume writes of the cause of the skeptical disposition with regard to the reason and the senses:

‘Tis impossible upon any system to defend either our understanding or senses; and we but expose them farther when we endeavor to justify them in that manner. As the sceptical doubt arises naturally from a profound and intense reflection on those subjects, it always increases, the farther we carry out our reflections, whether in opposition or conformity to it. (T 1.4.2.57)

According to Hume, the epistemological reflection that precipitates the skeptical disposition can occur (always occurs?) whether one is criticizing or attempting to justify that faculty. So just as the shift from the vulgar to the philosophical disposition is caused by reflection, so too is the shift from the philosophical to the skeptical.

In T 1.4.2, “Scepticism with regard to the senses,” Hume gives a nice illustration of how one might travel through the doxastic dispositions, from vulgar to skeptic, by way of successive waves of reflection. Hume thinks that it is a fact of human nature that we believe that (i) the objects of our senses have a permanent and continued existence. But the philosopher notices that (ii) we never experience an object continuing to exist while unexperienced, and (iii) there is never a strict identity between successive perceptions, only resemblances. While these conflicts may never become salient to the vulgar, in the philosophical disposition one reflects upon this incongruity between belief and experience. And in accordance with general rules of coherence, one tries to reconcile (i), (ii), and (iii). Some philosophers (e.g. Descartes and Locke) attempted to do this by positing another realm of objects, independent of perceptions, that are steady and permanent and that systematically cause our perceptions. This makes (i), (ii), and (iii) consistent, but it comes into conflict with part of Hume’s theory of causation, expressed by (iv): a causal judgment can only be formed if it is in principle possible to experience a constant conjunction. But as these purported objects that cause our perceptions are in principle not directly observable, we cannot “form a just conclusion” from the perceptions to the existence of the objects (1.4.2.54). So in attempting to reconcile (i)–(iii), the philosopher has proposed a solution that conflicts with (iv), which is a central philosophical principle of the imagination.

This last wave of reflection causes a shift in doxastic disposition from the philosophical to the skeptical.

Having thus given an account of all the systems both popular and philosophical, with regard to external existences, I cannot forbear giving vent to a certain sentiment, which arises upon reviewing those systems. I begun this subject premising, that we ought to have an implicit faith in our senses, and that this would be the conclusion, I shou’d draw from the whole of my reasoning. But to be ingenuous, I feel myself at present of a quite contrary sentiment, and am more inclin’d to repose no faith at all in my senses, or rather imagination, than to place in it such an implicit confidence. I cannot conceive how such trivial qualities of the fancy, conducted by such false suppositions, can ever lead to any solid and rational system. (T 1.4.2.56)

In the skeptical disposition, one doubts that there is an external world of stable, independent objects. And one doesn’t merely doubt; Hume seems to think that one can actually cease to be persuaded (i.e. cease to believe) that there is anything beyond our impressions and ideas (T 1.4.2.57). “Carelessness and inattention alone” can effect a transition back to the philosophical disposition and the world of independent objects (1.4.2.57).

Common to all of Hume’s skeptical arguments in the Treatise is the realization that the ‘fancy’—i.e. imagination narrowly-construed—is required for some domain of prima facie respectable beliefs. In his skeptical moods, Hume is inclined to refer to belief-enabling contributions of fancy as “trivial” (T 1.4.2.56, 1.4.7.3). It is clear that by “trivial” he does not mean that they have no use; they are highly important for survival and practical success. By “trivial,” Hume means that there is no indication that they are truth-tracking. The trivial propensity could provide a positive contribution like habit, which compels us to expect a B-type event when we encounter an A-type event if As and Bs have been constantly conjoined in past experience. But it could also be a negative contribution, as in the case of the skeptical argument against reason at T 1.4.1. After the first or second review, the process becomes “forc’d and unnatural” (T 1.4.1.10), and the mind is psychologically blocked by its own weakness and inconstancy from emptying the belief of all confidence.¹⁸

Now we have enough of the perspectivalist picture to look at an important objection to the claim that for Hume justification does not track truth. The objection is that there are passages in the Treatise that seem to suggest that truth is attainable.¹⁹ First, Hume sometimes appears to connect degree of justification to likeliness of truth. For example:

’Tis very happy, in our philosophical researches, when we find the same phænomenon diversify’d by a variety of circumstances; and by discovering what is common among them, can the better assure ourselves of the truth of any hypothesis we may make use of to explain that phænomenon. (T 3.3.1.25)

Second, Hume sometimes claims to have arrived at specific truths. There is, of course, the famous caveat of the final paragraph of Book 1, where Hume writes that when he uses expressions such as “’tis evident, ‘tis certain, ‘tis undeniable,” these “were exorted from me by the present view of the object, and imply no dogmatical spirit” and that they are “sentiments that…can become no body, and a sceptic still less than any other” (T 1.4.7.15). But in the Appendix, which Hume published along with the separately-published Book 3 (and with the aim of correcting mistakes in Book 1), he writes that it is “regarded as an undoubted truth, that belief is nothing but a peculiar feeling, different from the simple conception” (T App 3, my emphasis). Finally, Hume seems to think that belief that philosophical inquiry will lead to truth is a motivational precondition for this inquiry. He writes, “Love of truth [is] the first source of all our enquiries” (T 2.3.10.1). Relatedly, in his return to philosophy after the skeptical panic attack in the conclusion of Book 1, he cites curiosity as a motivating factor, which seems to presuppose the attainability of truth (T 1.4.7.12).

On the perspectivalist reading, this apparent conflict is not an actual conflict—and for the same reasons that apply to apparently conflicting standards of justification.²⁰ It is constitutive of the philosophical disposition that properly conducted inquiry will lead to truth. It is not until the science of human nature reveals that, in certain domains, mechanisms that produce belief cannot be shown to be connected to truth—they fall under philosophically irremediable suspicion—resulting in the draining of belief from ideas within those domains. Beliefs, like reason (T 1.4.1.1), are naturally oriented to truth (or, what is the same, the existence of the states of affairs they represent [T 1.2.6.2, 1.3.7.2-6]). And propensity for belief, often produced without the consent of the conscious mind, is certainly part of human nature. When Hume writes that “carelessness and in-attention alone can afford us a remedy” to the skeptical doubts that arise when we attempt to justify reason and the senses (T 1.4.2.57), it is carelessness regarding, and attention diverted from, the impossibility of connecting the proper use the faculties to truth. Once that is forgotten—and this is easily brought about, since the skeptical disposition is a highly unnatural, higher-order stance that is difficult to sustain²¹—one returns to the natural disposition of mind which presupposes that the world’s states of affairs correspond to our lively ideas.

The philosophical disposition is stable so long as one does not engage in faculty-justificatory epistemology. In Books 2 and 3 (and the majority of Book 1) this destabilizing inquiry is absent, and so Hume remains in the first-order scientist disposition in which it is presupposed that proper use of our faculties is likely to produce true beliefs. In the passage above quoted from T 3.3.1.25, Hume is endorsing the truth-conduciveness of an epistemic principle, but it is not deep epistemology—rather he is discussing diversity as a virtue of abductive inference, a virtue which itself has been inferred via induction from its scientific track record. This is paradigmatic first-order scientific methodology. The justification of induction itself (second-order, deep epistemology) is completely absent since it is irrelevant to the epistemic context. We see the conclusion and distillation of Hume’s investigation into scientific methodology in his “Rules by which to judge of causes and effects” (T 1.3.15). It is because he is securely in the philosophical disposition that Hume may write with confidence that he has discovered general rules for determining when causation has “really” occurred (T 1.3.15.2).

This also explains how it is possible to be motivated by truth in science and philosophy, even after having once inhabited the skeptical disposition. The cognitive-behavioral indolence of the skeptical disposition (T 1.4.7.8) is replaced by the relative epistemic indolence of the philosophical disposition (T 1.4.7.10; see fn 21), by way of which the motivational force of the passion for truth is restored.

I would like to briefly mention a fourth potential doxastic disposition: that of the false philosopher. These are philosophers that illicitly introduce metaphysical posits and jargon in order to solve problems that arise from reflection on vulgar attitudes.²² This includes the Peripatetics and rationalists who, for example, posit an unchanging substance to reconcile (i) the vulgar belief in the diachronic identity of objects and (ii) perceived changes in sensible qualities. In the progression of dispositions caused by philosophical reflection, the false philosophical disposition would fall between the vulgar and the true philosopher. Hume writes:

We may consider a gradation of three opinions, that rise above each other, according as the persons who form them, acquire new degrees of reason and knowledge. These opinions are that of the vulgar, that of a false philosophy, and that of the true. (T 1.4.3.9)

He follows this with a curious complication: “We shall find upon enquiry, that the true philosophy approaches nearer to the sentiments of the vulgar, than to those of a mistaken knowledge [i.e. the false philosopher]” (T 1.4.3.9). To square these claims, we need to understand Hume’s problem with the false philosopher.

Put generally, the false philosopher’s mistake is that they engage in metaphysical reflection (theorizing about metaphysical conflicts and contradictions within the vulgar worldview) without first engaging in the science of human nature. Hume writes in the introduction that “the only solid foundation we can give to this science itself must be laid on experience and observation” (T Intro 7). The solutions developed by false philosophers to the vulgar conflicts do not pass the test of experience. In the dialectic of reflection and successive levels of dispositions, the false philosophy is a way station between the vulgar and the true. By positing the theory of substance, accident, and occult quality, the

[false] philosophers set themselves at ease, and arrive at last, by an illusion, at the same indifference, which the people [i.e. vulgar] attain by their stupidity, and true philosophers by their moderate scepticism. They need only say, that any phenomenon which puzzles them, arises from a faculty or an occult quality, and there is an end of all dispute and enquiry upon the matter. (T 1.4.3.10, my emphasis)

What exposes this illusion? The science of human nature. The survey of the faculties that begins with the copy principle and ends by exposing the meaninglessness of Peripatetic metaphysical jargon. This is proper reflection on the metaphysical reflective correctives of the false philosopher upon the vulgar worldview. Conceived in this way, Hume’s experimental philosophy is a third-order inquiry. Let’s bring in the skeptical disposition. The skeptical disposition is produced by a certain kind of epistemic reflection within the science of human nature itself (genuine epistemic inquiry within the philosophical disposition). The “indifference” mentioned in this passage sheds light on the claim that “the true philosophy approaches nearer to the sentiments of the vulgar, than to those of [the false philosopher]” (T 1.4.3.9). Both the vulgar and the true philosopher attain a state of “indolence and indifference” (T 1.4.3.9)—the vulgar through lack of reflection, the true philosopher through moderate skepticism. The false philosopher, on the other hand, continues to seek metaphysical answers that the true philosopher has concluded to be beyond reach. So when Hume writes that “the true philosophy approaches nearer to the sentiments of the vulgar” he uses ‘sentiments’ quite intentionally. Far from indifferent, the false philosopher restlessly seeks an answer “in a place, where ‘tis impossible it can ever exist” (T 1.4.3.9).

In the next section, we will look at exactly how the epistemic normativity is infused into the framework of doxastic dispositions. But I would like to make a concluding remark about these skeptical episodes that will help make some space for perspectivalism. As we have seen, extreme skepticism arises from consistent application of the very belief-forming practices constitutive of the positive science of human nature. Therefore, on pain of inconsistency, Hume must accept extreme skepticism if he accepts these practices. But because of the detachment of epistemic justification from truth, he need not only accept skepticism. A properly qualified endorsement of scientific norms is still available to him. It may not be irrelevant to note that Hume thinks that the fact that these norms lead to skepticism ought to give us pause about the norms themselves, and thus about the skepticism (T 1.4.1.12, 1.4.7.14). The general point here can be put another way. In assigning a view to a historical philosopher, one should balance interpretive and philosophical plausibility. Both naturalistic interpreters such as Garrett and perspectivalists such as Fogelin are burdened with assigning a non-truth-tracking account of epistemic justification to Hume. But the benefit received from releasing justification from truth is a license to apply it to multiple, incommensurable (though, I will argue consistent) sets of epistemic norms. Thus, the perspectivalist can retain her advantage over the naturalistic commentator in interpretive plausibility (in the currency of textual fidelity), without taking on the severe philosophical implausibility of multiple, incommensurable—but still truth-tracking—perspectives.

3. From Believability to Normativity

Up to this point, we have mostly been engaged in the descriptive (to the degree possible) phase of identifying and analyzing the three major doxastic dispositions in the Treatise. Now it is time to bring in the normativity. On the perspectivalist account being developed here—epistemic dispositionalism—a belief is epistemically justified for Hume if and only if it is believable within the philosophical or skeptical doxastic dispositions and one is in that disposition. In the following two sections, we will answer two main questions. First, where does the normativity originate? Second why/how does the normativity get divided between these two incommensurable standards? In answering ‘why’ part of the second question, we will show how Hume’s view is coherent, and in doing so show how this perspectivalist view differs from that of Baxter. In answering the ‘how’ part of the second question, we will see, in precise terms, how Hume’s view is consistent, and in doing so show how this perspectivalist view differs from that of Fogelin.

The origin question divides into two: (i) where does the normativity originate in the logical structure of the Treatise? and (ii) where is its source in Hume’s epistemology as a finished product? These will be addressed in turn.

In the logical structure of the Treatise, epistemic normativity has its source in an initial assumption. Hume assumes—and keeps roughly fixed throughout the majority of the Treatise—the extension of ‘justified belief-forming principle’ as understood by his scientifically-minded contemporaries. When proposing a methodology in the introduction, he writes that the only “solid foundation” for the science of man is “experience and observation.” (T Intro. 7) But he gives no philosophical justification for this claim. He merely says that we shouldn’t be surprised that it took so long for experimental philosophy (i.e. Newtonian methodology), so celebrated in natural philosophy, to be applied to moral subjects. One attractive way to read Hume is that he recognized the successes of the experimental method in the physical sciences and decided to apply these same successful methods to moral subjects. Thus, the success of the sciences would justify the scientific method, and so, most fundamentally, observation and experience. But while this could be the genealogy of the project,²³ it cannot be the justification for it. The science of human nature, Hume thinks, is epistemically antecedent to natural philosophy (T Intro. 4–5). The Treatise starts in medias res; the experimental method is assumed at the outset, and then cleaned-up and precisified throughout Treatise 1.1–1.3, in part by way of consistent application of the copy principle.

One thing uncovered by this application of the assumed standards of experimental philosophy is that we have no reason to think that these standards track truth. This leads to question (ii): where is the source of normativity in Hume’s epistemology as a finished product (if not in the connection between these standards and truth)? On at least the general answer to this question, the perspectivalist and the naturalistic interpreter ought to agree. There are certain principles that have been found to be more fundamental to human nature than others. These principles have certain psychological properties that are characteristic of belief-formation in humans, and can be distinguished from the more contingent and accidental principles.²⁴ There are more “general,” “authentic,” and “established” (T 1.3.13.12) belief-forming principles in human nature. From these flow all epistemic normativity. The ideal epistemic posture is, as it were, encoded in human nature. This source coheres well with Humean naturalism as these psychological features are available to empirical study, while correspondence between what is empirically available and a separate world of objects is beyond the scope of Hume’s empiricism. Rational convergence (agreement based upon application of these principles) is all that can be said for a science.

When a warm imagination is allow’d to enter into philosophy, and hypotheses embraced merely for being specious and agreeable, we can never have any steady principles, nor any sentiments, which will suit with common practice and experience. But were these hypotheses once remov’d we might hope to establish a system or set of opinions, which if not true (for that, perhaps, is too much to be hop’d for) might at least be satisfactory to the human mind, and might stand the test of the most critical examination.²⁵ (T 1.4.7.14, my emphasis)

We may worry about circularity here. Perhaps it is no surprise that Hume should conclude that, roughly, the same set of belief-forming principles he assumed were justified at the outset, are in fact justified. But this may not be viciously circular. After all, he admits along the way that this set of principles cannot be proven to be connected to truth; this set, like all other candidate sets of epistemic norms, is in part determined by the imagination, by non-rational forces in human nature. So Hume is forced to distinguish among imagination-caused principles. Hume’s most explicit attempt to do this occurs at (T 1.4.4.1-2). Here, Hume is responding to an objection that he has been inconsistent. He has just criticized the ancient philosophers for being guided by “every trivial propensity of the imagination,” while at the same time holding that “the imagination, according to my own confession, [is] the ultimate judge of all systems of philosophy” (T 1.4.3.11-1.4.4.1). To resolve this apparent inconsistency he makes a distinction. He distinguishes between two kinds of principles of the imagination: those that are “received by philosophy”: permanent, irresistible, universal, solid; and those that are rejected by philosophy: changeable, weak, irregular, avoidable (T 1.4.4.1).

As always, the paradigm philosophical principle of imagination is the customary transition from cause to effect (T 1.4.4.1). Though this transition is determined by custom, one whose beliefs are enlivened this way, Hume claims, “reasons justly and naturally.” He gives the example of someone who receives the impression of a voice in the dark, which enlivens the idea of a human creature nearby. This is natural and just, according to Hume. He compares this person with one who believes there are “spectres in the dark.” Although this person reasons naturally, too, Hume says that his reasoning is natural in the sense that a disease is natural for a body, i.e. “arises from natural causes.” The first person, whose beliefs are formed by the philosophical principles of the imagination, reasons healthily, i.e. in “the most “agreeable and natural situation of man.” So in this section, part of what Hume is doing is sectioning off the more general and authentic belief-forming principles of human nature. It is these principles that are the source of epistemic normativity in Hume’s epistemology.

Distinguishing between sets of norms naturalistically (i.e. on the basis of apparently non-epistemic, psychological features) is common to both the naturalistic interpretation and the perspectivalist interpretation being developed here. But I would like to note one more thing about the T 1.4.4.1-2 passage that will help to identify a decision point for choosing between the two. I believe that when Hume writes “reasons justly and naturally,” he is picking out two different forms of doxastic normativity. ‘Justly’ refers to epistemic normativity, while ‘naturally’ refers to belief-formation that is psychologically healthy. It is necessary to briefly explain this because it is important for showing that the perspectivalist can plausibly claim that Hume endorses both philosophical and skeptical norms qua epistemic normativity, despite evidence that Hume all-things-considered thinks that the philosophical doxastic disposition is preferable. Along the way I will give some considerations for rejecting the so-called “title principle” as the master epistemic principle in the Treatise.²⁶

Don Garrett famously argues that Hume offers a considered, higher-order epistemic principle in the conclusion of Book 1 that circumscribes the domain of epistemically justified belief-formation to keep in scientific theorizing and keep out extreme skepticism. In the conclusion of Book 1, after describing the emotional character of his infamous skeptical panic attack, Hume writes:

These are the sentiments of my spleen and indolence; and indeed I must confess, that philosophy has nothing to oppose to them, and expects a victory more from the returns of a serious good-humour’d disposition, than from the force of reason and conviction. (T 1.4.7.11, my emphasis)

Hume then gives what Garrett (1997: 234) considers to be just this kind of higher-order epistemic principle:

In all the incidents of life we ought still to preserve our scepticism. If we believe, that fire warms, or water refreshes, ‘tis only because it costs us too much pains to think otherwise. Nay if we are philosophers, it ought only to be upon sceptical principles, and from an inclination which we feel to the employing ourselves after that manner. Where reason is lively, and mixes itself with some propensity, it ought to be assented to. Where it does not, it never can have any title to operate on us. (T 1.4.7.11)

Here, Garrett refers back to Hume’s discussion in T 1.4.4.1 of the principles of the imagination that are “permanent, irresistible, and universal.” Garrett thinks that in both of these passages Hume is saying something about what makes a belief/belief-forming process epistemically justified. He says that the above higher-order epistemic principle (expressed in the last two sentences of the passage), which he calls the “title principle,” can sanction these principles of the imagination because “even lively reason that mixes with our propensities cannot ultimately destroy their force” (1997: 234).

But I believe that in both passages (T 1.4.7.11 and 1.4.4.1) Hume is employing a non-epistemic evaluative standard of belief. When Hume says that “if we believe, that fire warms, or water refreshes, ‘tis only because it costs us too much pains to think otherwise,” he is using “pains” in at least two senses. It pains us to try to remain in the forced state of skeptical suspense because our nature is constantly pushing us back to belief and action. It also pains us to be skeptical because skepticism (of the extreme form) is attended with very unpleasant emotions. A better way to interpret this passage is to attribute to Hume a doxastic standard in addition to the epistemic that is based on what is “natural” in the sense of agreeable or psychologically healthy. On this reading, the “ought” in the title principle is expressing this latter kind of normativity, and should be interpreted as a rival principle to the epistemic. The subject of the title principle is reasoning; Hume is saying of reasoning (and, presumably, epistemically good reasoning) that it should only be assented to only if it is “lively and mixes with some propensity,” that is, is doxastically healthy/natural.

By distinguishing between these two forms of doxastic normativity, the epistemic and the natural, we are able to make sense of why Hume thinks philosophy is not necessarily for everyone. In T 1.4.7.14, he discusses the “honest gentlemen” of England, who spend their time employed in domestic affairs and amusements. Of them he writes:

They do well to keep themselves in their present situation; and instead of refining them into philosophers, I wish we cou’d communicate to our founders of systems, a share of this gross earthy mixture… (T 1.4.7.14)

According to Loeb, who thinks that justification in Hume amounts to stability of belief under the agent’s actual level of reflectivity (2002: 92–96), Hume is committed to the view that the honest gentlemen’s beliefs are justified because their vulgar level of reflectivity is insufficient to destabilize their beliefs. Loeb accepts that this can be formulated into an objection: that on his interpretation, “justification comes too easy” (2002: 93). But he notes, quite correctly, that Hume does not condemn the honest gentlemen. Loeb concludes:

The beliefs of the unreflective person occupy a preferred epistemic status. I believe that securing this paradoxical result was among Hume’s intentions in the Treatise. Hume seeks to show that an epistemic preference for reflection is a prejudice. As a matter of temperament, I suspect, Hume took delight in disparaging intellectual reflection. (2002: 97–98)

I believe this is a bit coarse-grained. It is true that Hume does not all things considered endorse maximal reflectivity. But he does repeatedly criticize the vulgar on epistemic grounds.²⁷ The “honest gentlemen,” though doxastically approved by Hume, are only epistemically virtuous insofar as they abstain from inquiry that goes beyond the concerns of everyday life. According to Loeb’s interpretation, if the honest gentlemen did go beyond these concerns (e.g. engaging in superstition), and these unphilosophically-generated beliefs were actually stable (versus, for example, being counterfactually stable) then they would be justified. I think this is both an implausible view—as the most ignorant and least reflective will fare best epistemically—and incompatible with Hume’s attitude toward the vulgar²⁸ in the Treatise.

The reason why Hume doxastically condones the honest gentlemen is because they are living natural and agreeable doxastic lives—that is, they are naturally successful. All-things-considered, this is the more important of the two doxastic norms for Hume. If doing philosophy would jeopardize one’s doxastic health, then it ought to be avoided. See this passage near the end of Book 1:

Human nature is the only science of man; and yet has been hitherto the most neglected. ‘Twill be sufficient for me, if I can bring it a little more into fashion; and the hope of this serves to compose my temper from that spleen, and invigorate it from that indolence, which sometimes prevail upon me. If the reader finds himself in the same easy disposition, let him follow me in my future speculations. If not, let him follow his inclination, and wait the returns of application and good humour. (T 1.4.7.14)

And so, antecedent to any preference for perspectivalism, there are good reasons to think that Hume endorses these two distinct forms of doxastic normativity. The upshot for the perspectivalist interpretation is that it gives Hume a non-arbitrary reason to prefer the moderate higher-order epistemic activity characteristic of the philosophical doxastic disposition to the extreme—perhaps neurotic—reflectivity of the skeptic. Moderate use of the probabilistic control principle, for example, keeps the accountant from making mistakes; consistent use of the same principle—while no objection can be made against it on epistemic grounds—leads to a doxastic disposition unfavorable to psychological health and practical success. This distinction also enables Hume to recommend or proscribe certain domains of enquiry. According to the perspectivalist, faculty-evaluative epistemology is a perfectly legitimate domain of enquiry²⁹—after all Hume does it! But the desire for well-being ought to inhibit us from engaging in it (more than is salutary).

4. Epistemic Pluralism

Thus ends our discussion of the origins of epistemic normativity. Let’s move on to the second question posed at the beginning of §3, again split into two: (i) why and (ii) how does the normativity get divided between the two incommensurable perspectives?

First, the normativity must be split between the philosophical and skeptical dispositions because they are epistemically continuous. These dispositions share with each other, but not with the vulgar disposition, the belief-forming norms assumed by Hume to be justified at the outset of the Treatise. Along the way, epistemic normativity becomes naturalized in the sense that he realizes (using these norms) that (a) we have no reason to think they are truth-conducive, and yet (b) they are the general and established doxastic principles of human nature, and so carry a kind of normativity that is distinguishable from the practical,³⁰ moral, aesthetic, etc. Hume also realizes that there is no non-arbitrary way to epistemically distinguish between philosophical and skeptical use of these principles.

If there were a way to epistemically and non-arbitrarily distinguish between the philosophical and skeptical dispositions, this would weaken the perspectivalist account by making it vulnerable to the objection that it renders Hume’s epistemology incoherent. This is a problem for the perspectivalist account offered by Baxter (2006; 2008; 2018). Baxter proposes a Pyrrhonian solution³¹ to the integration problem. Like with Sextus, he believes Hume distinguishes between “active endorsement” which is “an act of the will based on appreciation of reasons” and “passive acquiescence” which is “a causal effect of appearances” (2018: 380). He associates these with the cogitative and sensitive parts of our natures, respectively, claiming that for Hume, passive beliefs are produced because belief is a manner of conceiving (i.e. a forceful and vivacious idea) (2008: 9). Because, in various domains, we cannot show our beliefs to be truth-tracking, reason—our cogitative nature—cannot produce belief. He likens this to the argument that reason alone is unable to motivate action in Book 2 (2008: 9). Instead, belief results from natural, irresistible processes, and he concurs with many naturalistic interpreters and myself that we should look to Hume’s distinction between principles of reasoning that are “permanent, irresistible, and universal” versus those that are “changeable, weak, and irregular” (T 1.4.4.1) to show how we can distinguish between proper and improper principles of natural belief formation (2008: 11).

One plausible way to read the normativity of Baxter’s distinction is that (i) active assent has alethic epistemic standards (i.e. is concerned with truth) while (ii) passive assent has naturalistic epistemic³² standards (i.e. that justification reduces to some psychological feature of a belief or belief-forming mechanism, and does not necessarily track truth). But this would open Baxter to the difficult question of why Hume doesn’t just endorse one or the other—apart from the consideration that nearly all interpreters accept: that Hume really seems to endorse both. After all, the skeptical interpreter just focuses on the fact that Hume argues that our natural beliefs cannot be grounded in truth. Meanwhile, the naturalistic interpreter convincingly argues that Hume endorses naturalized epistemic standards (Garrett’s lively reasoning with propensity, Loeb’s stable beliefs, etc.), to which the skeptical reflections do not live up and so are left abandoned in Book 1. Endorsing both saddles Hume with a radically disjunctive form of epistemic pluralism: not only do the standards of justification yield different verdicts on the same proposition, the standards themselves are completely different in kind. This would significantly increase incoherence in Hume’s epistemology.

Another concern about Baxter’s view is that I find no such distinction in Hume between active and passive assent. To see this, we must pull apart the justification of a belief from the psychological/causal mechanism that produces it. In the T 1.4.1 argument, which Baxter takes as a core text for his interpretation, it is not just the fact that belief is a lively idea that makes it an act of our sensitive nature; it is the fact that the liveliness (doxastically: the credence) cannot be sustained through the iterative review process. The same goes for causal beliefs, which require the non-rational link of habit to sustain the input vivacity. Hume’s vivacity theory of belief is in principle consistent with these kinds of beliefs being cogitative (i.e. alethically rational), but the actual processes that act on the input vivacity are contingently non-rational. But on Baxter’s view, active assent is in principle impossible for Hume, both because reason cannot endorse belief but also because belief is not an act of the will. We should note that first-order science can be very active in the sense of carefully assenting in accordance with rational principles (by the standards that constitute rationality in the philosophical disposition). The difference between this and the skeptical disposition lies in the content on which the principles are applied and in the order of their application (how many steps away from objects and toward the faculty itself).

On the epistemic dispositionalism interpretation, contra Baxter, the skeptical disposition, too, is thoroughly naturalized, and so there is no room for either the skeptical or naturalistic commentator to drive a wedge between the dispositions. To see this, we need to distinguish between natural beliefs and natural belief-forming principles. It is true that Hume thinks that there are natural beliefs that constitute the philosophical disposition (e.g. the belief in continued, mind-independent external objects). And it is true that the skeptical disposition lacks these beliefs, both in the inferences that lead to the disposition and in the disposition itself. But natural belief-forming principles constitute a much greater portion of the philosophical disposition than natural beliefs. These principles are critiqued in the transition from the philosophical to skeptical disposition, but they are also the very principles utilized in this critique. For example, both demonstrative and probabilistic reasoning are natural in the sense that they rely on an imaginative contribution. And yet they are utilized in the arguments for, and ultimate conclusion that, these contributions (e.g. habit, custom, the vivacity of ideas) are trivial (see especially T 1.4.7.3). There is absolutely no point of view outside of natural belief-forming principles—except perhaps for gods and angels—from which to even attempt to ground these principles.

On this, I agree with Jani Hakkarainen, who sketches a perspectivalist view that contains two Humean “domains”: the everyday domain and the philosophical domain (2012: 302). Domains, here, are “spheres of belief- formation or cognitive commitment with different epistemic standards” that are different in degree but not kind (2012: 303). But where Hakkarainen finds two perspectives, I find three, and in fact none of his wholly map onto mine. For this distinction, he cites Hume’s claim in the Enquiry that “philosophical decisions are nothing but the reflections of common life, methodized and corrected” (2012: 301, citing EHU 12.25). But for Hakkarainen, the philosophical sphere is skeptical to the extent that it involves suspension of belief in external objects. Yet in this passage Hume is referring to decisions within (my) philosophical disposition; it is not true of the skeptical disposition, where nearly none of the reflections of common life hold. The problem of Hakkarainen’s lack of the tripartite distinction manifests in his answer to the question: if Hume the philosopher affirms or denies the existence of external objects (or, we may add, suspends judgment), why not say it explicitly? Hakkarainen responds that “silence fits quite nicely with the suspension of judgment” (2012: 305). Academic Hume (Hume’s mitigated skeptical position in the Enquiry) is silent, and this silence “supports the skeptical rather than dogmatic reading of his position” (2012: 305). But, in fact, the Academic (or, in the terminology of the Treatise, ‘moderately skeptical’) Hume believes in a modest but positive science of human nature. On my account, Hume is silent on the existence of external objects within the philosophical disposition not because he does not believe in them, but because there is no position at all from which Hume can affirm (or deny) the existence of external objects in general. To do so would assume their existence has become a question, and that means one is in, or on the way to, the skeptical perspective. In this perspective, there is no believability one way or the other. So if one believes in external objects (as one does in the vulgar and philosophical dispositions), one is disposed to only question the existence of this or that particular object, while holding the realm of objects fixed. This is quite interesting: for Hume some kinds of propositions—as a psychological and epistemic matter—may be presupposed but not thoughtfully endorsed.

But even if the normativity of the perspectives is the same in kind, one might worry about another source of incoherence: that the transitions between perspectives are unphilosophical, i.e. that one is not led from one to the other by way of justified inference. This is at least false with respect to the transition from the philosophical to the skeptical disposition. We have already seen the clearest case of this. At T 1.4.1, the philosophical principle that governs how we ought epistemically to react when faced with a contrariety of causes (T 1.3.13.8) is applied to our mixed track-record performing inferences. Hume thinks that if we were capable of applying this principle consistently to our inferential judgments,³³ it would result in the suspension of judgment on all products of inference. But this also occurs in Hume’s skepticism regarding the senses (see §2). Again, it is higher-order, consistent application of philosophical principles of the imagination that precipitates the skeptical disposition and its characteristic suspense.

It is fairly easy to see how the transition from the philosophical to the skeptical disposition is philosophically mediated. But it is very difficult to see how Hume can claim that the transition back to the philosophical disposition can be philosophically mediated. In fact, it appears that Hume explicitly denies this. In the middle of the conclusion of Book 1, Hume rhetorically takes up the skeptical disposition, in which he “can look upon no opinion even as more probable or likely than another” (T 1.4.7.8). Then he writes:

Most fortunately it happens, that since reason is incapable of dispelling these clouds, nature herself suffices to that purpose, and cures me of this philosophical melancholy and delirium. (T 1.4.7.9, my emphasis)

After he has come down from the skeptical panic attack, he still feels:

such remains of my former disposition, that I am ready to throw all my books and papers into the fire, and resolve never more to renounce the pleasures of life for the sake of reasoning and philosophy. For these are my sentiments in that splenetic humour, which governs me at presents… These are the sentiments of my spleen and indolence; and indeed I must confess, that philosophy has nothing to oppose to them, and expects a victory more from the returns of a serious good-humour’d disposition, than from the force of reason and conviction. (T 1.4.7.10-11, my emphasis)

It is distraction and forgetting that effects the transition from the skeptical to the philosophical disposition. Backgammon and dinner parties do what no philosophical argument could have. So in the transition from the skeptical to the philosophical perspective, it seems there is no philosophical mediation, though this transition is certainly justified by the other form of doxastic normativity, discussed above: what I have called natural justification. By moving from the skeptical to the philosophical, we are moving to a more healthy and agreeable doxastic state.

This leads to another question. One might object that in establishing continuity between perspectives, we make them vulnerable to epistemic comparison. For example, the skeptical disposition seems to fare better than the philosophical on consistency. With respect to reason, this shows up in consistent application of the probabilistic control principle; with respect to the senses, consistent application of Hume’s theory of causation; with respect to probability at T 1.3.6 (i.e. the problem of induction), consistent application of the norms of circularity. If the philosophical disposition runs afoul of its very own consistency norms, then perhaps this would be sufficient reason to reject it as a legitimate set of epistemic standards. So even by the lights of the naturalized standards of the Treatise, the skeptical interpreter has room for her wedge.

This objection of intra-dispositional consistency leads us to some very strange territory. In offering a response, it is necessary to engage in some rational reconstruction.

The philosophical doxastic disposition contains consistency norms, but, as the objector notes, these norms appear to be narrow in scope. This narrow scope, though, is not enforced by the letter of the norms themselves, but rather by non-rational forces of human nature that put psychological limits on our ability to apply them. But, again, non-rational influences are essential to all or nearly all cognitive activity, including justified belief-formation. It is also important to add that the scope of the consistency norm is only narrow with respect to higher-order judgments (i.e. judgments about our judgments, rather than judgments about objects). Inconsistencies, for example, in the physical sciences ought always to be reconciled. But Hume notices that it is a feature of human nature that iterative, nested judgments about judgments strain the understanding and obstruct the flow of vivacity from impressions to ideas, even if the same form of inference would be quite natural and easy if applied to more terrestrial topics (T 1.4.1.11). This is not to say that Hume wants to put strict limits on theorizing about perceptions qua perceptions. Much of the Treatise contains theorizing of this kind, from the very first distinction between impressions and ideas, which is repeated after the skeptical panic attack at the outset of Book 2. Theorizing about the contents of our minds as mental entities with their own realm of laws makes up a good portion of the cognitive psychology of not only Book 1, but also Book 2 (and, to some extent, Book 3). But the higher the reflexivity of the contents, the less natural the inference. This is not ‘natural’ in the sense of ‘psychologically healthy’ from above; rather it is in the sense that habit and custom are natural, i.e. as non-rational enablers of belief that constitute the more general and authentic principles of the imagination. And recall that according to both this perspectivalist interpretation and naturalistic interpretations, epistemic justification reduces to non-rational psychological features of beliefs. So the limiting of the application of consistency norms by natural, fundamental psychological features is countenanced by nearly every interpretation that grants epistemic justification to the science of human nature.

But these natural limits on the application of consistency norms can be overstepped. Of course, this makes the philosophical disposition unstable in the face of epistemic reflection. Both endorsed dispositions (the philosophical and skeptical) are unstable to some degree; this is part of the motivation for perspectivalism. The non-rational forces in human nature are not absent from the skeptical disposition either, so one cannot judge the skeptic pure, following reason faithfully wherever it has led her. Nor, perhaps, once she has arrived there can she consistently hold her ground.³⁴ This overstepping of the natural limits on the application of consistency norms is one gateway to the skeptical disposition. The question here is whether application of the consistency norms to the consistency norms themselves is acceptable within the philosophical disposition. If so, the philosophical disposition would be internally inconsistent. Perhaps not; perhaps when one engages in this activity one is already en route to the skeptical disposition.

One might worry that a narrow-scope consistency principle is just irrational, or absurd. Isn’t a rule that we should only sometimes be consistent immediately repugnant? But again, it is not the principle that is narrow in scope, but the application. Selective application of consistency, while apparently absurd, may only appear so because of our antecedent preference for truth-tracking. That is, it may not necessarily be epistemically vicious if epistemic virtue is detached from truth. These questions are interesting, and I will leave it to the reader to decide how epistemic dispositionalism fares with respect to this Escherian puzzle into which we have stumbled.

There is a deeper, though less puzzling, point to be made here. For the perspectivalist interpreter, the epistemic standards that constitute the dispositions cannot be epistemically compared. If they could, it would have be against some master epistemic criterion. But if such an extra-dispositional, overawing criterion existed, it would be the sole epistemic criterion for Hume. It is true that the dispositions can be compared by the standard of natural normativity (i.e., conducive to health and success). As mentioned above, this is what allows Hume to say “’Tis happy… that nature breaks the force of all sceptical arguments in time” (T 1.4.1.12; see also T 1.4.7.9, 1.4.2.57). Distraction and forgetting effect a transition from the skeptical and the philosophical disposition; because they are extra-dispositional, they cannot be evaluated epistemically (and luckily, too; by any reasonable standard they would do poorly). But by natural normativity they are celebrated.

So because the dispositions cannot be epistemically compared, we need not worry about that species of inter-dispositional inconsistency. But Fogelin argues that Hume’s perspectivalism contains a different species of inter-dispositional inconsistency. He says that perspectivalism is straightforwardly inconsistent because the assertions made from the different perspectives are inconsistent (1998: 165–166). This, Fogelin claims, distinguishes perspectivalism from relativism. Relativistic assertions have the form from perspective A, x seems F; from perspective B, x seems not F. The ‘seemings’ insulate the judgments from coming into contradiction. But perspectivalist assertions, Fogelin argues, come in the form x is F (asserted from perspective A); x is not F (asserted from perspective B). Because the assertions themselves are unqualified they can come into contradiction.

Fogelin is right that judgments from the different doxastic dispositions are not mere seemings, and so they are the sorts of things that could come into contradiction. But they would only be inconsistent in the epistemically-loaded sense if held by a single person, in a single disposition, at a single time. These qualifications make all the difference. Beliefs within the philosophical and skeptical dispositions may be inconsistent, but this is acceptable because their justificatory statuses are indexed to different background principles and beliefs, and different sets of higher-order corrective tendencies. The contexts in which the belief is evaluated are so different that it looks much closer to changing one’s mind (e.g. because one gains new evidence, or because one recognizes the belief was based on an irrational epistemic principle). On this view, justification claims are always indexed to a doxastic disposition. This feature of epistemic dispositionalism is similar to typical internalist theories in which justification claims regarding a single belief are always indexed to a set of mental contents and inferential relations. Epistemic dispositionalism just extends the set of mental features necessary to determine whether a belief is justified.

To review, on the perspectivalist reading developed here Hume can answer both charges of inter-dispositional inconsistency. The dispositions themselves are not inconsistent because there are no extra-dispositional criteria with which to compare them. The assertions made within the dispositions are not inconsistent (at least in the relevant evaluative sense) because the justificatory features of the judgments are indexed to a disposition, and so cannot come into conflict.

5. Overview and the Problem of the Vulgar

Let’s step back. On this interpretation, epistemic justification, for Hume, reduces to the psychological feature of believability. A belief is epistemically justified if and only if it is believable within the philosophical or skeptical doxastic disposition and one is in that disposition. A belief is epistemically unjustified in all other cases. In the skeptical disposition, it may be the case that the only ideas that are believable—and so justified—are those about occurrent mental states (perhaps, also, those relations of ideas “discoverable at first sight”). The doxastic dispositions are constituted by (i) active belief-forming principles, (ii) higher-order corrective tendencies, and (iii) emotional disposition. The philosophical disposition is characterized by dominant philosophical principles of the imagination, higher-order general rules that correct the more precipitous first-order judgments, and minimal influence of the passions. The skeptical disposition is constituted by extreme consistency in application of higher-order, corrective principles and faculty-justificatory reflection that lead to wide-ranging suspension of judgment. Shifts between dispositions are not directly under the control of the will, but can be precipitated by an increase or decrease in philosophical reflection. This grounds Hume’s epistemology in the vicissitudes of our mental lives, which is very much in the Humean spirit. The psychological winds blow not only our passions but also the very standards of our judgments.

I believe that we shouldn’t think it strange that this interpretation commits Hume to epistemic pluralism. If one is going to reduce justification to a psychological property—as I and many naturalistic interpreters argue Hume does—one would need to give a principled reason why some conditions under which this property arises are appropriate and some inappropriate.³⁵ The fact that Hume uses normative epistemic language in two distinct and individually coherent standards and endorses both of these standards in his own voice gives us a reason to ascribe to him an epistemic pluralism. I have attempted to show that this pluralism is not only consistent with Hume’s science of human nature, but is very much in the spirit of it. One might wonder how this view compares to contemporary pluralistic epistemic theories, like contextualism.³⁶ Unlike typical contextualist theories on which epistemic standards are determined by social conventions (e.g. norms of discourse), for Hume these standards are characteristically naturalistic. They are determined by our psycho-epistemic disposition, which is not under our direct control. Likewise, shifts in standards are not, as the contemporary contextualist might argue, caused directly by intentions of interlocutors. We must submit to psychological forces beyond the direct control of the will, but that can be influenced indirectly by increases or decreases in philosophical reflection.

One might also wonder why, according to this interpretation, the vulgar’s vulgar beliefs are not justified (or how Hume can claim that any belief is unjustified!). If justification just amounts to believability within a doxastic disposition, isn’t it arbitrary to give epistemic approval to the philosophical and skeptical dispositions but not the vulgar? What principled reason could Hume have to draw this boundary? I would like to pair this with a different but related objection: if on this perspectivalist interpretation justification is only possible within the philosophical or skeptical dispositions, this seems to entail that none of the vulgar’s beliefs are justified. I take it that it is almost equally absurd to claim that all of the vulgar’s beliefs are justified as it is to claim that none of them are.

To answer the first objection, that this interpretation commits Hume to an arbitrary distinction between the vulgar and other, more charmed dispositions, I would like to return to the question of the origins of epistemic normativity in the Treatise. It is not, I argued, derived from a non-naturalistic epistemic standard like truth-tracking. Hume begins with standards of justification that would be endorsed by his scientifically-minded contemporaries and proceeds to refine them so that they are internally coherent and consistent with his empiricism. But it is clear by the conclusion of Book 1 that he does not think a Cartesian—or any other—grounding project can succeed. If the faculties can’t be grounded, then (barring an externalist theory) the epistemologist is left to (i) identify the more general and authentic belief-forming principles that constitute human nature and (ii) an empirical study of the norms governing the use of justificatory language. Regarding the latter, the scientist in the realm of the theoretical, as well as the cool, sober agent in the realm of the everyday, have roughly fixed the extension of ‘justified belief-forming process,’ and Hume is committed to the view that there is just not much more for philosophers to say about it.³⁷

The vulgars’ vulgar belief that there are, for example, “spectres in the dark” is not justified by the standards of the scientist, or even by sober, reflective common sense. And Hume is content to operate within the epistemic standards of this community (albeit cleaned-up with his help), since he thinks there is no epistemically superior vantage point from which to evaluate them.

This will help us answer the objection that epistemic dispositionalism seems to commit Hume to the absurd view that none of the vulgar’s beliefs are justified, since for one’s beliefs to be justified one must be within the philosophical or skeptical doxastic disposition. First, I will respond by noting that this picture of three wholly distinct doxastic dispositions is merely heuristic. In practice, a single human at any given time can be disposed to make judgments in accordance with radically different epistemic principles, especially with respect to different domains of belief. For example, a person could be philosophically disposed with regard to beliefs about cars and their effects on human bodies when they collide at high speeds, but vulgarly disposed with regard to beliefs about the long-term health effects of her diet, or about the properties of gods. This makes the full account of doxastic dispositions much more complicated. A total report of one’s doxastic makeup at any given time would identify those domains about which one is vulgar, philosophical, and skeptical, and one’s beliefs would be judged accordingly.