1. The Revival of W. D. Ross

It may seem hyperbolic to speak of a revival of Ross, but although he was not forgotten by leading twentieth-century moral philosophers writing after he published The Right and the Good (1930), until the 1990s most of them considered him minor. This is partly because he wrote under the shadow of Moore,Footnote ¹ but a more important element was the influence of philosophical naturalism. That worldview took many forms, but in ethical theory two main strands were influential for much of the past century. One was reductivist, the other eliminativist. A controlling naturalistic idea was that natural properties are the only properties we should countenance. Mill's utilitarianism could be conceived as meeting this standard, but Moore's Principia Ethica (Reference Moore1903) convinced most ethical theorists that ethical naturalism in any reductive form is unsustainable. Moral philosophers who accepted that conclusion had to choose between conceiving ethics as committed to non-natural properties and framing an eliminativist view that accounts for moral discourse without positing them.

One influential response to this choice came with the positivism of A. J. Ayer. On his naturalistic view, moral predicates—say, ‘right’ and ‘wrong’—are emotive and do not represent genuine properties. In the aftermath of positivism, this eliminativist stance was developed by C. L. Stevenson, R. M. Hare, and many others (see Stevenson [Reference Stevenson1944], Hare [Reference Hare1952], and for a more recent version—‘norm expressivism’—Gibbard [Reference Gibbard1990]). In their hands, ethical naturalism survived the rejection of both positivism and naturalistic reductivism in ethics. The power of progressively refined eliminativist and naturalistic ethical theories reduced Moore's overall influence in ethics, but it also kept him in view since his anti-naturalistic arguments were taken as indirect support for the noncognitivist position that moral predicates do not express genuine properties, and, correspondingly, assertive moral sentences do not express propositions. Moore is important, then, for both strands of naturalism: his anti-reductivist arguments forced ethical naturalists either to refine reductivism in ethics or to devise a non-reductive naturalism aimed at meeting his challenge.Footnote ²

Twentieth-century ethics exhibited two pervasive tensions, both still unresolved. One, just described, is between naturalism, whether cognitivist or not, and non-naturalism, championed above all by Moore; the other is between realism and anti-realism. Here Ross is crucial: he is both a non-naturalist and a realist, and, beyond that, his ethical intuitionism strikingly combines a commonsense normative core and a sometimes far from commonsensical epistemology concerning that core. The core is a list of moral principles well known among moral philosophers and widely respected in applied ethics.Footnote ³ The principles express obligations of justice and non-injury, veracity and fidelity to promises, beneficence and self-improvement, and reparation and gratitude. Ross considered these principles self-evident. For instance, he thought it self-evident that if one promises to do something, one has a ‘prima facie duty’ to do it. The same kind of obligation applies to avoiding lies and injustices and, on the positive side, to aiding others in need.

Like Moore, Prichard, and many earlier philosophers, Ross took the self-evident to be unprovable but ‘just as certainly’ not in need of proof. He represented the self-evidence of the everyday moral principles he formulated as like that of a mathematical axiom or a form of inference.Footnote ⁴ Pronouncements like these led many to believe that for intuitionists, one can ‘just see’ the truth of the principles, as with logical truths corresponding to elementary inferences. One might call this the just-see (or luminosity) stereotype regarding self-evident propositions—given a good look at them, which was considered implicit in any comprehending consideration, one just sees their truth. This conception of self-evidence partly accounts for the view that intuitionism is dogmatic and its claims of self-evidence ‘conversation stoppers’. It also partly accounts for skepticism about claims of self-evidence, about the justificatory value of intuition, and about the substantive power of a priori reason.

Ross avoided claiming that only the self-evident is an object of intuition, and he emphasized that seeing the truth of his principles required ‘a certain mental maturity’. But he did not explicate that notion. Much work in recent decades has done so, and there is now wide agreement that a new intuitionism is possible by preserving much of the Rossian core, clarifying intuition, and providing a more plausible moral epistemology.Footnote ⁵ The new intuitionism has, among other things, far more to say about intuition, a different conception of the self-evident, and a more modest ontology. All of these topics are important for understanding any field of philosophy.

2. Six Dimensions of Intuitional Discourse

In the history of philosophy, intuition has played diverse roles. Keeping to the tradition of ethical intuitionism, which includes a major epistemological element, we find at least six cases essential for fully understanding intuition. First, there are cognitive intuitions—intuitions that p. These have propositions as objects and are in that sense propositional intuitions. These are the most common kind of intuition traditionally considered in ethics. Suppose I have the intuition that accepting a bottle of wine as a gift from an undergraduate student would be wrong. I find this proposition credible in its own right though I can also argue for it. The second case we must take into account is intuitiveness. For me, the proposition about such gifting has this property; that is, in its own right, taken in itself, it appears true: it evokes what might be called the sense of non-inferential credibility.

Intuitiveness is also predicable of concepts, arguments, and other intellectual phenomena, but their intuitiveness is largely explicable in terms of the intuitiveness of certain propositions in which they figure. A third case, then, is related to the first two. Propositions may be called intuitions when they are considered intuitive or are regarded as intuitively known. You might say, for instance, that p (some proposition) is an intuition we share, and in building a philosophical account of, say, reasons, we might speak of what intuitions—roughly, what intuitive propositions—about reasons should guide us.Footnote ⁶

A fourth case is that of objectual intuition: a kind of non-propositional intellectual perception: a direct apprehension of either a concept, such as that of obligation, or of a property or relation, such as the property of being justified or the relation of entailment. These direct apprehensions (‘apprehension’ is the more common term for objectual intuition) are, when occurrent, significantly akin to sensory perceptions: as with perceptions, their object is present in consciousness. An occurrent intuition of obligation, for instance, may embody apprehensions of agency and of being ‘bound’. The same cognitive content is dispositionally instantiable. If I have, occurrently, an intuition of the fittingness of doing a deed to my having promised to do it, I need not cease to have it when my mind is wholly occupied with other matters.

A fifth case is indicated where writers on intuition view it as a kind of non-inferential judgment: a sort of intellectual verdict on a proposition or situation.Footnote ⁷ In ethics, certain moral judgments—judgmental intuitions—are often so viewed, as where, on considering an accusation against a friend, one thinks, ‘That is unfair’. Judgmental intuitions seem to entail a kind of assent to the proposition in question: the sort of affirmational response normally taken to underlie sincere judgments.

These five cases are connected with a sixth: just as we have perceptual faculties such as vision, we have intellectual faculties, including intuition—reason in one sense of that wide-ranging term. The faculty of intuition is a kind of apprehensional capacity: a non-inferential capacity through which we know what we intuitively do know. It is not a modular part of the mind but a general capacity needed for philosophical reflection and manifested in relation to each of the other five cases. It is appropriately called non-inferential because of the way it responds to content in context: above all, to the appearance of truth in propositional objects considered in their own terms rather than as supported by premises. Intuition is like perception in the directness of its cognitive deliverances, but it does not require unresponsiveness to premises. Intuitions (as deliverances of this capacity) may not come fast, any more than an aesthetic response to a complex painting need come at first look, but intuitions are not inferentially dependent on premises.

Consider Ross's promissory principle, accepted in much ethical literature before and after him: If one promises to A, one has a prima facie obligation to A. I take this to be roughly the claim that promising entails a moral reason to do the promised deed: a reason that is prima facie in being strong enough to justify doing the deed in the absence of stronger opposing reasons but weak enough to be overridable by those. Ross and most other intuitionists would say, using the six locutions just described, that through our intuitive capacity that we can see the truth of the principle; that when the principle is considered with adequate understanding, it is intuitive; that one has an intuition of the obligation-making power of the relation, promising; that one may have an intuition whose content is this principle; that the principle itself is an intuition (a kind of intuitive datum) an ethical theory should accommodate; and that judgmental intuitions may have such principles as their propositional contents.

Intuitions as cognitive attitudes—propositional as opposed to objectual intuitions—need special attention. Ross and many other philosophers have spoken of intuitions mainly as a kind of belief. With intuitions in mind, Ross repeatedly referred to the ‘convictions’ of ordinary people.Footnote ⁸ This manifests a doxastic (belief-entailing) conception. Moreover, it remains common for philosophers and others to report their intuitions about hypothetical cases with a clear implication that they believe the propositions in question. In recent work in epistemology, however, intuitions are often treated as intellectual seemings—a phenomenal, attentional sense of the truth of the proposition in question. This sense often reflects, though it does not entail, the proposition's being intuitive: having a truth-appearance of a kind analogous to but distinct from a sensory appearance, such as one has with perceptions of colors and shapes.Footnote ⁹ These seemings—call them episodic intuitions—are commonly taken to entail inclinations to believe, but they are not beliefs and, like visual impressions whose accuracy one doubts, do not entail forming beliefs.

Ethical intuitionists have paid too little attention to the phenomenology of intuitions; intuitionistic epistemologists have tended to focus predominantly on the episodic, phenomenal conception of intuition and (in recent years) to stress that episodic intuition is not belief-entailing. But how the episodic and doxastic conceptions are related still needs explanation. Discerning that relation requires exploring phenomenological aspects of intuitions conceived in both ways. A central question here is whether, although intuitive seemings do not entail believing their propositional objects, doxastic intuitions nonetheless depend on intuitive seemings.Footnote ¹⁰ Answering this should help us to explain a difference between moral philosophers and epistemologists in their main conceptions of intuition—or at least their main appeals to it.

3. The Phenomenology of Intuition

Doxastic intuitions are not occurrences, and their possession does not entail manifestations in consciousness. As a kind of belief, they are dispositional mental properties, and these can be possessed without such manifestations. When doxastic intuitions do manifest themselves in consciousness—say, by one's having the thought that p—they do have a phenomenal character, but beliefs need not manifest themselves phenomenally even when doing causal work, as they do in sustaining other beliefs. Still, doxastic intuitions are not merely non-inferential beliefs with intuitive content. To see this, consider whether we can have an intuition that p if p has never come before the mind. Can we have doxastic intuitions without ever having episodic intuitions—intuitional experiences—with the same content?

Consider the proposition that if one encounters a wandering toddler on a sidewalk adjoining a traffic lane, one should protect the child. The proposition is before my mind and seems true on its own account. I have an episodic intuition that it is true. I need no premises for it, think of none at the time, and feel inclined to affirm it. My sense of its truth is enhanced by my vividly imagining a child wandering near fast-moving traffic; finding this proposition intuitive does not require such a mental picture, and people may differ in the kinds of routes by which they arrive at intuitions—say, ratiocinative, pictorial, or emotional—and, depending on content and context, we may all differ in these. As it happens, I also believe the proposition about the toddler. But this belief need not be what I report if my self-ascription of the intuition is about my cognitive relation to the proposition when it is before my mind. What if I see a woman who appears to be the child's mother some dozen feet away from the child looking in its direction but also animatedly speaking with someone? I might still have the intuition that I should protect the child, but this time I might not believe that. I would tend to believe it, but the presence of the apparent mother changes the pattern I respond to and makes the need for intervention unclear. An episodic intuition, even of an intuitive proposition, need not lead to a belief.

These cases may also suggest that a general intuition, such as the conditional one about wandering toddlers, arises from a singular intuition, of the sort I would have if I encountered such a child apparently alone near a traffic lane. This kind of situational genesis of intuitions with general content seems common, but not strictly necessary, in their formation. Our intuitive capacity may yield intuitional experiences whose objects are general—even abstract—without our having had parallel experiences of what might be called supporting singular phenomena. Indeed, perhaps intuitions as such need have no particular history.

There is, however, some basis for considering the doxastic notion of an intuition historical in implying that we cannot have such dispositional intuitions that p without ever having an intuitional experience that p. In avowing a doxastic (dispositional) intuition that p—say, that one should protect endangered children—I normally presuppose that it has an appropriate connection with an episodic intuition. But imagine a perfect psychological duplicate of me. Might my duplicate, at the very moment of creation, have the same relevant beliefs, including doxastic intuitions, that I do, at least given plausible assumptions about the supervenience of the mental on the physical? I see no conclusive reason to answer negatively, and I leave open whether doxastic intuitions must trace back to or be based on episodic ones, that is, intuitive seemings. In any event, for the normal cases, we might plausibly conceive episodic intuitions as the basic kind. How might we do this?

A unified account centered on episodic intuition might, in rough outline, go as follows. Doxastic intuitions embody a disposition to have episodic intuitions with the same content and normally have a basis in the latter, whether that basis is contemporaneous or developmental. Intuitiveness in a proposition is a truth-appearance of the kind manifested in an episodic intuition with the same propositional content: broadly, evoking a sense of non-inferential credibility. Objectual intuition, when occurrent, is an apprehensional consciousness of the concept, property, or other object in question with a focus that makes possible episodic intuitions about one or more aspects of it and, when dispositional, related to the episodic case essentially as dispositional propositional intuitions are related to their episodic counterparts. Judgmental intuition may be conceived as an assenting cognitive response to an episodic intuition (and, when dispositional, as where it is stored in memory, as having the same kind of relation to such assentings as doxastic intuitions have to episodic intuitions). The faculty of intuition is the non-inferential rational capacity whose central exercise manifests itself in the formation of episodic intuitions and in the retention of an appropriate range of dispositional cognitions having the same propositional or other objects.

This account partly explains why self-ascription of intuitions typically indicates belief: we tend to believe what is intuitive for us. It also partly explains the point, implicit in the concept of having a doxastic intuition, that if one considers its propositional object, one tends to find the proposition intuitive. Suppose the basis of a doxastic intuition is episodic intuition. Then it is expectable that, upon comprehendingly considering the proposition—unless under defeating conditions that eliminate its appearance of truth—one will have an episodic intuition with that content.

Two points emerge in the light of these reflections. First, if we adopt the parlance of much ethical literature since at least Sidgwick, we may say that when one reports having the intuition that p, one normally presupposes, but does not assert, that one believes p. The second point, easily overlooked, is that when a dispositional intuition is a kind of belief, then although the intuition normally has a basis in episodic intuition, the belief need not have arisen from any intuition. What we now intuitively believe—say, having reflected on a proposition and found it compellingly intuitive—we may have previously come to believe through testimony, through simply reading it, or indeed by inference. In those cases we may form beliefs in the normal way we do when we are open to receiving information and our critical faculties are, even if ready to block the doubtful, fully open to the flow of the plausible. If a doxastic intuition must be non-inferential and has a basis in an episodic intuition, the belief constituting the former may arise beforehand and by any of a number of routes. A doxastic intuition, moreover—say, that beneficence is obligatory—need not persist even with retained belief of its propositional object. An opposing narrative may undermine one's sense of the non-inferential credibility of that proposition. But what no longer stands on its own may still be sustained by premises. Indeed, where intuition fades, philosophical reflection may cast the light of supporting argument on the formerly intuitive proposition.

Brief as it is, this phenomenological sketch suggests an explanation of the difference, at least in emphasis, in the way moral philosophers, especially ethical intuitionists, and epistemologists have tended to view intuition. Moral philosophers have regarded their main normative principles as justifiable and, if not as intuitions in the sense in which certain propositions are called intuitions, at least as fully consonant with intuitive judgments subsumable under the principles—for instance, principles about specific promises or about cases that, like the wandering toddler, call for beneficence. Moreover, even non-intuitionist moral philosophers have considered intuition to be sometimes needed to determine what one should do in difficult situations. Here, too, it is natural to take the deliverances of one's intuitive rational capacity to be doxastic, as they commonly are. Such doxastic responses are the more likely where deliverances are cautious, say to the effect that one probably should do a given thing. Here belief is more readily formed when its content carries a lesser commitment.

By contrast, at least since Descartes, epistemologists have been deeply concerned with skepticism and perennially preoccupied with the regress of justification. Consider this: beliefs admit of justification, and if intuitions are beliefs (‘convictions’, for example), then, particularly if they are fallible, as is common, they admit of justification and may need it. If, however, one takes intuitions with self-evident objects to be unprovable and, like Ross and others, considers singular intuitions to be admissible as ‘data’ in ethics, and in that way foundational, one may see no need to justify these cognitions. But if, like many epistemologists, one is concerned with stopping the regress of justification, one will seek, even for perceptual beliefs of a kind moral philosophers typically presuppose, grounds for justification that can confer it and do not admit of it, hence cannot be attacked as lacking it. Intuitional experiences can play this role. As phenomenal episodes, they are not true or false. They can be misleading, but not justified or unjustified.

Episodic intuitions, then, can serve as data at the ground level—or some ground level: some grounds are harder than others, some higher, offering wider perspective, some richer, providing better fruits. We need not take justificatory grounds to be infallible or indefeasible supports for belief, but beliefs that derive from them receive prima facie justification.Footnote ¹¹ Doxastic intuitions, moreover, as non-inferential cognitions, may rest directly on such grounds. That does not guarantee truth, but it does eliminate defeat by inferential dependence on a false premise or invalid reasoning.

Given these two patterns of discourse in ethics and epistemology, must we regiment intuitional terminology to avoid ambiguity? In some contexts this is desirable, but we need not abandon either the doxastic or the episodic conception of intuition. The two conceptions, once distinguished, can work side by side, particularly if their intimate relation is kept in mind. A doxastic intuition, as a belief, may be justified not only by appeal to the episodic intuition likely available to its possessor and commonly a basis for it, but also by argument from premises. An episodic intuition, as an experience, does not stand in need of justification and may confer it. Like sense experience, it is only a fallible indicator of truth; but much as we cannot navigate the physical world without sense experience, we cannot adequately pursue truth in ethics or elsewhere without intuition.

4. The Intuitive, the Evident, and the Self-evident

The intuitive has been closely associated with the self-evident. The association is especially close in ethical intuitionism, which has prominently claimed self-evidence for certain moral principles. Our examples show, however, that (as Ross and other intuitionists realized) some intuitions have non-self-evident propositions as objects, for example, my intuition that I should protect the wandering toddler even with its apparent mother in view. Moreover, like visual experience, intuitional experience may yield false belief. The self-evident, however, like the evident, is factive. False propositions can seem evident, even self-evident, but not be so. On my view, self-evident propositions are truths meeting two conditions: (1) in virtue of adequately understanding them, one has justification for believing them (which does not entail that all who adequately understand them do believe them); and (2) believing them on the basis of adequately understanding them entails knowing them (see Audi Reference Audi2004: ch. 2). To see how something self-evident need be neither intuitive nor even believed upon comprehending consideration, focus on the proposition that a child can be borne by its grandmother. One may well have an intuition that this is false until one figures out how it can be true. Oddly, the intuition of falsity might persist even then, at least when the proposition is considered without bearing in mind the initially misleading contingencies.Footnote ¹²

Even if one agrees that (true) propositional intuitions need not have self-evident objects, one might think that, apart from oddities like our parenting case, the self-evident must be intuitive. Commonly it is, but what of such logical propositions as this: if p entails q, and q entails r, and r entails s, and s is false, then p is false? I find this intuitive, but must everyone who comprehendingly considers it find it so? It may be that once one sees how this is true, it will seem intuitive, but I doubt that this holds for the self-evident in logic generally, and it is not true for all candidates for the self-evident in ethics.

This logical example shows two other important things about the self-evident. First, contrary to most major ethical intuitionists of the twentieth century, self-evidence is compatible with provability, as the logical example illustrates. Second, adequate understanding of a self-evident proposition—the kind of understanding that guarantees being justified in believing it—may, for some people, require drawing inferences. For example, someone might need to infer from the antecedent of the theorem we began with that p entails s. Inferring this might be how one sees that if s is false, then p is also false (it is luminously true that if p entails s, and not-s, then not-p). This deduction, however, represents an internal inference: its premises—that p entails q, that q entails r, and that r entails s—are supplied (though not asserted) in the antecedent of the target proposition, the original ‘long’ self-evident proposition. The inference is not external, in the sense that the self-evident proposition justifiedly believable in the light of it is inferred from premises that this proposition does not itself (internally) supply. Self-evidence does not rule out external inferential routes to self-evident propositions, but self-evident propositions, even when provable, are not premise-dependent, in the sense that justification for them requires climbing the ladder of external inference.

In some cases, then, the kind of understanding that guarantees justification for believing a self-evident proposition can be indirect and may arise partly through drawing inferences. Understanding need not require inference, but may employ it. Indeed, in some cases, many inferential connections between p and other propositions must be grasped for p to be adequately understood in the first place. This is one reason why the self-evident need not be obvious.

Far more could be said here about understanding, but enough has been shown to indicate directions for further inquiry and to refute some stereotypes that block the appreciation of intuition as important in and beyond ethics and of self-evidence as an element in most versions of ethical intuitionism. Intuition need not be a momentary snapshot. It may be virtually instantaneous, but that is testimony to how much the mind can encompass at once, and anyway, it is not the only case. With intuition, as with perception, temporal immediacy does not entail narrow vision. Both can quickly respond to complex patterns. Moreover, the self-evident, even when intuitive, need not be unprovable. But it also need not be intuitive, and certainly not obvious, in the first place. These points explain how believing it can be withheld. We can thus reject yet another stereotype blocking the plausibility of rationalism in ethics and of intuitionism in particular. The self-evident need not be unprovable, need not be obvious, and need not be rationally beyond dispute.

5. Rational Disagreement in Ethics: The Intuitionist's Dilemma

Ethical intuitionism faces a dilemma, indeed one that can beset almost any intuitionist view. If the self-evident is by its very nature obvious and unprovable, no one who understands it should fail to believe it, and it needs no proof. If, however, it is not the case that the self-evident is by its very nature obvious and unprovable, then rational persons can disagree about its truth, and those who deny or doubt it may demand proof and may reject it if proof is not forthcoming. The first horn was wielded by Prichard and has led many philosophers to view ethical intuitionism as dogmatic.Footnote ¹³ The second yields a challenge that must be met if the new intuitionism now in development is to be sustained. Having rejected the stereotypes that underlie the first horn, I will examine the second. I will argue that neither dogmatism nor skepticism need result from intuitionism that employs the moderate conception of self-evidence and the associated notion of intuition I have described.

It is not just Rossian moral principles and counterpart epistemic principles that we should take into account in dealing with disagreement about the self-evident. Many philosophical theses may be conceived as self-evident if true. Consider a basic epistemic principle: It is not rational to believe something having the form of ‘p and not-p’. We tend to presuppose this, and even if we can think of ways to argue for it, we do not take it to need argument. We can, of course, illustrate this principle by citing propositions of similar status—that believing that I am reading and I am not reading is not rational, that believing that this paper is white and that is not white is not rational, etc. One could call this kind of support of a self-evident proposition intuitive induction, as did Ross, citing Aristotle. But this kind of argument does not work, as supporting arguments usually do, by adducing premises knowable, or justifiably believable, wholly independently of the conclusion they are meant to support. It works mainly by enhancing the understanding of the proposition it supports. Here it may parallel the way certain self-evident propositions are learned. But the order of argumentation need not match the order of learning. Arguing for a proposition we believe by appeal to premises for it does not imply that these premises lay on our route to believing it.

To focus the question of how we should conceive rational disagreement on the self-evident, recall a basic point about self-evidence: that adequately understanding a self-evident proposition entails having justification for believing it, but not believing it. Given this account, it should be clear how one can understand a self-evident proposition and still comprehendingly consider it without believing it. A central point is that adequacy of understanding goes beyond basic semantic comprehension. A bilingual person, for instance, could understand a self-evident proposition well enough to translate a sentence expressing it into another language, yet still fail to believe it. Someone who can immediately translate ‘A child can be borne by its grandmother’ may still need reflection or even Socratic questioning to achieve understanding of the kind that justifies accepting the claim. Mere semantic comprehension of that proposition need not justify believing it. But seeing its truth through an adequate understanding of it enables one to believe it non-inferentially, presumably on the basis of apprehending the concepts figuring in it and the relations among them in virtue of which the proposition is true.

An adequate understanding of a self-evident proposition, moreover, does imply (at least in a rational person) a disposition to believe it, indeed, one strong enough so that non-belief given comprehending consideration calls for explanation. Two possible explanations are thoroughgoing skepticism and commitment to a theory clearly implying the negation of the self-evident proposition. In the case of disagreement on a complex proposition, finding such an explanation is often possible, but it may not be easy. Moreover, in some cases it is unclear whether a person lacks adequate understanding. Understanding comes in degrees. With Rossian principles in mind, let us explore how these points about the self-evident bear on rational disagreement.

If, like the major intuitionists from at least Moore on, one takes moral properties to be consequential (hence supervenient) on non-moral properties, one might think that rational persons who directly disagree on a moral proposition either do not share all the relevant non-moral information or do not each see what that information indicates. This view is highly plausible, but its applicability is limited by the impossibility, in many cases, of specifying all the information relevant: the information indicating the non-moral properties on the basis of which there is a truth of the matter in dispute. Recent epistemological work has raised a challenging problem that does not presuppose our having such comprehensive information: how to deal with possible disagreements one has with someone who seems, in the relevant matter, an epistemic peer: roughly, a person who (a) is as rational and as thoughtful as oneself (in the relevant matter, the assessment of whether p is true), (b) has considered the same relevant evidence—which need not be all the relevant evidence—and (c) has done so equally conscientiously (see Audi [Reference Audi2011] and, for further discussion, Christensen and Lackey [Reference Christensen and Lackey2013]). By contrast with most descriptions of epistemic parity, this one requires that the parties consider the relevant evidence and do so equally conscientiously. If parity requires only sharing the same relevant evidence and having the same epistemic virtues (or being equally rational in the matter, which is a similar condition), nothing follows about how fully these virtues are expressed. Hence, despite possessing the same evidence, the parties might have devoted very different amounts of time or effort to appraising the proposition. That disparity is common in philosophical controversies.

Given the complexity and idealization in the notion of epistemic parity, one would at best rarely be justified in considering a disputant an epistemic peer regarding the disputed proposition. This point alone explains how one might justifiedly adhere to a view one knows someone else rationally denies. Suppose, however, that one is justified in believing something weaker and more general: that a disputant is, by and large, an epistemic peer on the relevant subject, such as ethics. Rational disagreement still seems possible. This can be seen in relation to disagreement on the status of Rossian principles (Ross's and similar ones). Since having an adequate understanding of a self-evident proposition implies having a justification for believing it but does not entail believing it, one could rationally believe that a disputant can adequately understand a self-evident proposition, be justified in believing it, yet not believe it.

Take the strong ‘particularist’ case against Ross's intuitionism (a counterpart case may also be made for rejecting the self-evidence of certain epistemic principles). Using examples such as promises whose fulfillment would kill innocent people, some have denied that promising to do something entails any prima facie moral reason to do it. One might hold that only in particular cases viewed holistically can one tell whether promising yields any obligation. Suppose that, quite reasonably, intuitionists do not allow that one can be justified both in believing p and in believing not-p. They must then deny either that the particularist adequately understands the promissory principle (where adequate understanding implies justification for believing the principle) or that the particularist's arguments justify rejecting it. Intuitionists may deny the latter even while granting that the arguments provide some reason to accept strong particularism (this is my strategy in Audi [Reference Audi and Dreier2006] in responding to a strong version of particularism). Let me explain.

A major point here is that lack of overall justification for believing p does not imply that believing it is irrational. Being sometimes unjustified—in the sense of lacking overall justification—is common in rational persons in matters requiring rigorous reasoning; irrationality, by contrast, is a lapse they do not often suffer. In any event, failure to use reason adequately does not entail flying in its face. That failure is compatible with having some degree of justification, but the question here is overall justification of the kind that normally suffices for knowledge when the relevant belief is true and, in broad terms, appropriately grounded. This point is important in appraising disagreement. It implies that even justified confidence that a disputant is unjustified does not automatically warrant considering the disputant irrational. This implication bears on appropriate reactions to apparent peer disagreements and—the more realistic case—on disagreements that approach them: an unjustified but not irrational disputant (or one holding an unjustified but not irrational position) can give one some reason to doubt the proposition(s) at issue, even if not in general sufficient reason to doubt or withhold it.

6. Two Orders of Disagreement

So far, the depth and resolvability of disagreement on the self-evident have not been addressed. These variables are important for understanding disagreements on self-evident propositions of the kind important for ethical intuitionism and indeed for other philosophical positions. We might begin with a distinction between two kinds of disagreement in normative matters. Consider two jurors. They might agree in their verdict but disagree in their reasons for it. One takes the accused's standing to inherit a tidy sum to support guilt; the other, believing the accused to be already rich, does not. Or consider two parents who disagree on the appropriate punishment of their errant child. One believes, and the other disbelieves, that the offense calls for spanking. Disagreement in reasons concerns one or more elements that are (at least presumptive) reasons: it is difference regarding whether, in such contexts as those illustrated, the reasons succeed in the supporting role in question. Disagreement on reasons concerns reasons conceived as such, typically their epistemic or ontic status. Parties who disagree in reasons need not differ on any higher-order view of reasons, such as the thesis that reasons are constituted by evidence. If they do, their disagreement at that level is on reasons.

Disagreement on reasons is common in philosophy and is, broadly speaking, theoretical and higher-order. Disagreement in reasons, though also common in philosophy, is not theoretical but a kind of disagreement in practice, especially in providing explanations and justifications. It often concerns whether a particular consideration succeeds as a normative reason—for example, whether it justifies a lie. Disagreement on reasons often concerns the nature of a normative role—for instance, whether only facts can justify. Reasons are crucial in epistemology and other philosophical fields besides ethics, but let us focus just on ethics.

In ethical discourse, disagreement in reasons is concrete and most often occurs in one or more of three cases. It may concern (1) whether a particular factor, such as a specific promise to a friend, is a reason or an explanation, say for a specific action, (2) whether a certain normative principle, viewed in the context of someone's giving a reason in accordance with it, is true (a disagreement closely related to the first case) or (3) whether a relevant inference—for instance to the conclusion that someone did wrong in breaking a promise—is, given the relevant facts, appropriate (above all, valid or adequately probable). Disagreement on reasons is abstract and does not entail disagreement in reasons even where the reasons in question are, like promises, the very factors that figure in the abstract formulations—such as Rossian principles—regarding which there is direct disagreement, the kind manifested where one party affirms the principle and the other denies it. This explains why skeptics about morality who deny that there are justificatory reasons may, like non-skeptics, in everyday life criticize someone as having no good reason to break a promise.

The practice of giving counterexamples also illustrates the disparity in question: between uses of reasons (or indeed other notions) and views about them. Beyond this, the practice illustrates a certain fluidity in normative discourse: agreements and disagreements in reasons about concrete cases may lead to agreements and disagreements on reasons in general and vice versa. This fluidity can make the distinction difficult to draw for certain disputes but does not reduce its importance. Reasons are somewhat like tools: knowing how to use them does not imply ability to describe them, but such know-how can assist description, as good description can sometimes enhance know-how.

Disagreement in ethics need not derive from adherence to any particular theory. Consider another intuitive moral principle (also affirmed by Ross and others): Injuring other people is (prima facie) wrong. Suppose I am angry at a small boy who jostles me on running into a crowded bus. Imagine that I sharply slap his face and bloody his nose. You might be shocked by this violence and say that I overreacted and should apologize. If I replied, ‘Why? I was angry at him and had no reason not to strike him’, would you assent? Isn't one's first thought to the effect that there was of course a reason not to slap him? This thought would indicate a disagreement in reasons. Would any normal person doubt that there is some reason not to slap him? Much as promising to A is virtually always accepted as a reason to A, the fact that doing something would injure a person is virtually always accepted as a reason not to do that. This pattern of response does not depend on any theory, perhaps not even on any implicitly accepted moral principle. The term ‘reason’ is also not crucial, nor is any higher-order view of reasons. It is no more likely that I would respond to my critic with ‘I like slapping boys who annoy me, and I don't see why I shouldn’t’, where this is not to give an excuse or appeal to any theory but to reject avoiding injury as any reason to abstain.

Disagreement on reasons tends to be less deep than disagreement in them, especially in controlling thought and action. Suppose I reject the promissory principle because it conflicts with my desire-satisfaction theory of reasons for action. An agent's desires need not be optimally satisfied by maintaining promissory fidelity. Still, in spontaneous self-expression, as where parents chide their children for selfishness, I might criticize someone for breaking a promise and might unhesitatingly reject the claim that, because a change of desires made the promise-keeping aversive, the agent had no reason to do the promised deed. Spontaneous critical practice often belies a theory that should explain it.

Reason-giving is a natural part of communicative discourse. It is a common element in most adult human lives, but knowing how to give reasons does not imply knowing how to describe how we do it. Even after considerable reflection, one may easily formulate a normative principle (or a description of a reason) inadequate to one's own practice or reject a principle or description devised by someone else that is adequate to it. Much the same holds for principles that account for our linguistic usage. There, too, intuition is an essential guide to normal usage but, even as governing the concrete cases to which linguistic principles are responsible, may well not yield an ability to formulate such principles or prevent rejecting correct principles formulated by others.

Distinguishing between disagreements on reasons and disagreements in reasons does not depend on any particular terminology, especially ‘reason’ or any equivalent; much explanatory and justificatory discourse appeals to the relevant notions in other terms, and I intend what I say about disagreement in and on reasons to apply to various other normative notions, including that of justification. A crucial element here is norm-guidedness: if, for instance, ‘rational’, ‘sensible’, or even ‘ok’ are used in the right critical manner, then noting their applications can help us to determine whether people agree or disagree in reasons. Moreover, much agreement is implicit, a matter of such variables as what classifications one is disposed to make and of what inferences one would draw or reject. An important instance is moral reasons. We appeal to such considerations in numerous ways. Just saying, for example, ‘I can’t—it would hurt her feelings’ or ‘I won't say that; it would be a lie’, implicitly recognizes the status of lying or hurting feelings as providing reasons.

Agreement with Rossians on the relevance and, especially, the force of promissory reasons and other reasons that intuitionists emphasize may be, in various ways, implicit. Even if I reject the Rossian promissory principle, I may still virtually always (1) assume that I am morally obligated to do a thing given that I have promised to; (2) consider wrong—or at least wrong if not excusable—broken promises I learn of; and (3) infer (or presuppose), when people tell me that they have promised a third party to do something, that they ought to do it. Here, (1) indicates a classificatory response to a reason for action (a reason grounded in promising); (2) illustrates both that response and a judgmental response to acting contrary to what a reason calls for; and (3) indicates an inferential response to promising as providing a reason for action. These responses, in a common, spontaneous pattern, weigh heavily against rejecting the promissory principle. Now I submit that such cognitive disparities between our—in my view—basic cognitive behavior and our high-level general beliefs and dispositions to believe are not uncommon and that part of the task of philosophy is to eliminate such disparities where possible. Here intuitions about the classification of cases and about the validity or plausibility of inferences are important elements in achieving reflective equilibrium. As philosophers know well, it takes great skill to formulate principles, or to specify what types of elements constitute reasons, in ways that do justice to one's own careful usage and capture one's own (non-formal) inferential practices. This is one reason why disagreement on reasons may be resolvable by clarification, often through presentation of cases or through the search for reflective equilibrium.

Disagreements in ethics, then, are easily taken to indicate greater cognitive differences than they in fact show. Vague disagreements, whether on or in reasons, may be mitigated or eliminated by clarification, and when deep disagreements are on reasons, they may still be accompanied by a high degree of concurrence in using and responding to reasons. Kantians and certain utilitarians, for example, often largely agree in reasons important for political decisions and often endorse liberal policies. It is quite possible, moreover, to be reasonable in retaining a plausibly contested conviction when one is respectfully resolved to explore the objections. The reasonableness of adhering to a disputed view may in fact depend on its higher-order context in the mind of the adherent.

It should now be clear how a rational person can understand a self-evident proposition adequately, acquire justification for believing it, yet find plausible arguments against it and thereby excusably and, in some at least minimally rational way, deny it. I leave open the question of conditions under which one can justifiably believe that a dissenting interlocutor is in the epistemically undesirable position just described. My concern here is to indicate how rational disagreements can extend even to the self-evident, to facilitate their resolution by distinguishing disagreement on reasons from disagreement in reasons, and to show how the former is compatible with considerable agreement, between the disputants, in reasons. Agreement in reasons is often a basis for resolving disagreements on them, much as coincidence in our use of tools can lead to agreement on descriptions of them. In dealing with rational disagreement, then, I cannot see that intuitionist ethics is significantly worse off than any other major view in moral philosophy or, for that matter, worse off than philosophical views that do not purport to defend substantive self-evident propositions.

7. Intuitions without Intuitionism, Intuitionism without Realism

I have refuted some common misconceptions of intuition, intuitionism, and the self-evident. I have also shown how natural it is to appeal to intuitions in seeking evidence for philosophical views and even in supporting some everyday moral propositions. If I am right, we should expect to find appeal to intuitions, sometimes without the name, in many fields. Here are four historically significant cases drawn from philosophers not generally considered intuitionists.

Aristotle might be seen as providing intuitive anchors for virtue theory when he said that ‘not every action or feeling admits of a mean. For some have names immediately connected with depravity, such as spite, shamelessness, envy, and, among actions, adultery, theft, homicide’ (Nicomachean Ethics 1107a [2000: 31]). Aquinas, in one place (Summa theologiae Iallae Q 91, Art 2), referred to ‘natural reason, whereby we discern what is good and what is evil’ as if he took natural reason to be what I have called the faculty of intuition. He affirmed, as the first precept of natural law, ‘good is to be done and promoted, and evil is to be avoided . . . so that whatever the practical reason naturally apprehends as man's good (or evil) belongs to the precepts of the natural law as something to be done or avoided’. These precepts he said (earlier in the same article) are ‘self-evident principles’ (Summa Iallae Q 94, Art 2). Third, might we not see Kant's ringing affirmation, ‘Now I say that man, and in general every rational being, exists as an end in himself, not merely as a means for arbitrary use by this or that will’ as an expression of an intuition? (Kant [1785]Reference Kant and Paton1948: 95). Even Mill, who disparaged what he called ‘the intuitive school,’ apparently regarded two of his important claims as intuitive: ‘The only proof . . . that an object is visible is that people actually see it. . . . In like manner, I apprehend, the sole evidence it is possible to produce that something is desirable is that people do actually desire it’ (Mill [1861]Reference Mill1979: 34; italics added). Nothing said here implies that these philosophers lacked or could not find premises for the propositions in question. My point is that they took these propositions to be plausible, perhaps indeed knowable, in their own right, as is common with propositions that, on considering their content, one finds intuitive.

Clearly, we need not regard the philosophers just quoted as intuitionists. Philosophers in any field can make essential use of intuitions without being committed to intuitionism. May we go so far as to say that, as Ross implied, intuitions are data not only of ethics but of philosophy, much as sense perceptions are data of science?Footnote ¹⁴ Consider the analogy to perception. My seeing a copper beech tree about 100 feet from where I am seated is an orienting perception. I have a perceptual belief that it is about 100 feet away and must take it into account in certain activities. When I go inside, I cease to have this perceptual belief but remember the proposition that was rendered a datum by perception. I can normally trust this memory even if I cannot perceptually confirm it. Without trusting memory in such cases, we cannot function normally.

Now compare the case of intuition. I have an intuition that one should act protectively toward the wandering toddler I described. The proposition is before my mind, and I find it highly credible in its own right, independently of thinking of any premises for it. Granted, the sense of looming danger is a factor in my finding it so, but this no more implies an inference than does my analogous sense of goodly distance as a factor in coming to believe that the copper beech is about 100 feet away. There is also a difference from the perceptual case: if I look at the tree and consider whether it is about 100 feet away, I normally must form the belief that it is, whereas I can often resist belief formation when an intuitive proposition comes before my mind.Footnote ¹⁵ This does not by itself imply that intuitional impressions are less reliable than perceptual ones: it is not inconceivable that the world be such that the probability of truth given the former is higher than that of truth given the latter. This is apparently not our world, however, and that point is one reason why perception tends to be more trusted than intuition. Human knowledge as we know it depends on both perception and intuition. Philosophy does not depend on perception in the way or to the extent that it depends on intuition. For philosophy, intuition has at least this advantage: we can at will call propositions before the mind to subject them to appraisal, whereas nature must cooperate in ways it often does not if we are to produce or reproduce perceptions. Reflection has wings and can go where it will. Perception is imprisoned in our bodies and depends on our surroundings.

The important point here, too rarely appreciated by philosophers, is that episodic intuitions evidence propositions in a non-inferential way that bears important similarities to the way in which perceptions evidence propositions essential for scientific inquiry and indeed for ordinary navigation of the world. In both cases there is conferral of a degree of justification and there is reliance on memory for retention of the propositions so evidenced. To be sure, there are differences. When beliefs arise from intuitional experiences and are retained in memory, we tend to call them intuitions, but when beliefs arise from sense perception and are memorially retained, we do not call them perceptual beliefs. That is probably because, whereas the very thought of the propositional object of an intuition tends to evoke the (or a) grounding intuitional experience, the parallel point plainly does not hold for propositions believed on the basis of perception and then recalled from memory without the same occasion for sensory confirmation. Granted, where a belief is occurrently based on perception, as is my visual belief about the copper beech when I view it, we can speak of seeing (hence perceiving) that it is about 100 feet away. This is a propositional perception based on seeing the tree. But for beliefs arising from perception and retained in memory, we cannot revisit their grounds just by calling their propositional objects to mind; for beliefs arising from episodic intuition, we commonly can do this. This is epistemologically significant. In the crucible of reflection, we can review, reappraise, and, often, confirm.

Much more could be said about the strength of the analogies and disanalogies here. But it should at least be clear that only a thoroughgoing skeptic would deny that some degree of rational support is implied by clear and steadfast sensory appearances, and I see no reason to consider intuitive experience different in kind on this point. Descartes, preoccupied with defeating skepticism, used the metaphor of the light of nature for the faculty of intuition operating at its best: delivering clear and distinct ‘ideas’. The metaphor is good if we include the deliverances of sense experience among the beams of light. Here I am content to hold that without the light of intuition there are both objects and propositions that we either cannot see at all or cannot see clearly enough for justification and knowledge. The same holds, of course, for the light of perception.

If, in philosophical and other thinking, intuition has the pervasive role I have sketched, and if intuition is similar to perception in the ways I have described, then it should be evident that recognition of this role does not foreclose the possibility of naturalism in ethics, even intuitionist ethics. Given that, historically, intuitionists have opposed naturalism, one might wonder how this can be so. A short answer is that neither the role of intuition regarding singular moral propositions nor even its role in providing justification for self-evident generalizations implies that the properties in question are non-natural. A priori relations can hold among natural as well as among non-natural properties. What is round, for instance, is not square.Footnote ¹⁶

As to anti-realism, Berkeley provides a historically important example of anti-realism about the objects of perception together with no less reliance on intuition than one would expect in philosophy. Might noncognitivist anti-realists in ethics do as well? They cannot take assertive moral sentences to ascribe properties and thereby be truth-apt, but they may still take moral judgments and the moral attitudes they express to be non-inferential when their basis is of the kind intuitionists have posited, such as the perception of an apparently endangered toddler. Such attitudes may also be understood as norm-guided in much the way linguistic behavior is norm-guided: in neither case is one compelled to posit non-natural properties as essential to explaining the critical practices the norms prescribe. Moreover, acceptable claims of self-evidence can be conceived in terms of reasonable assertibility on the basis of understanding. Philosophical ingenuity can salvage all the accoutrements of truth value even while it jettisons the properties usually considered essential to its basis.

Ethical intuitionism explicitly gives intuitions a major evidential role regarding both singular moral judgments and moral principles. But this role is not one that intuitions cannot play in other ethical theories, other branches of philosophy, and indeed in aspects of scientific inquiry. Moreover, although ethical intuitionism is committed to countenancing non-inferential justification, it may be adapted to an empiricist epistemology and to ontological naturalism. It may even be adapted to a noncognitivist, anti-realist moral ontology by taking non-inferential justification to apply to judgments viewed as expressions not of propositions but of attitudes, stances, plans, or other non-truth-valued elements. Such noncognitivist views can incorporate the normative standards Ross and other intuitionists have articulated, and they can account for the way in which, both in content and through evoking motivation, those standards can guide conduct. My own ethical intuitionism is moderately rationalist, realist, and uncommitted to ontological naturalism. My primary aim has not been to defend my own ethical theory but to clarify the notion of intuition and to show its overall importance in moral inquiry. If I have been right, intuitionism must be counted among the major kinds of ethical theory adequate to guiding theoretical reflection in ethics and relevant to guiding moral conduct in everyday life.

May we conclude, as Ross did, that intuitions are the data of ethics? Whether in their doxastic or even their episodic forms, they are not the only data of ethics, but in both forms they are often durable strands in the fabric of moral sensibility and indispensable supports for moral judgment. They are essential for ethical reflection. They are crucial for serious reflection of any kind. They are starting points for philosophical theories even if the structures we initially build from them ultimately need other foundations. What we build philosophically goes beyond the deliverances of intuition, but intuitions are unavoidable in the appraisal and reappraisal of our results. This conception of philosophy does not in the least imply that scientific findings and theories are irrelevant to its work. Far from it: accounting, in a certain way, for scientific knowledge and indeed empirical knowledge in general is a major task of philosophy. The larger philosophical question is the power of reason: the question of the kind and extent of justification and knowledge possible through reflection. Here data must come from objects of intuition as well as from sense perception. Intuitions, even when their objects are self-evident, do not put those objects beyond rational dispute. Intuitions provide ideas for exploration, evidence for judgment, and guidance for action. If intuitions never preclude disagreement on substantive matters, they also do not undermine our routes to rational resolution. Philosophical reflection, whether or not it is intuitive, is not isolated from common experience or scientific progress, not even on its normative side, but the role of intuition and a priori reflection portrayed here remains a central element in both the pursuit of knowledge and the autonomy of philosophy.