Ethical Realism

November 19, 2010

Review of Nathan M. Nobis’s Truth in Ethics and Epistemology

This review is available as a free ebook (PDF file) here. (Right click/save as to download).

Terance Cuneo argued that moral realism is true (moral facts exist) based on the fact that (a) epistemic facts exist1; and (b) if moral facts don’t exist, then epistemic facts don’t exist. Around the same time Nathan Nobis wrote his doctoral thesis, Truth in Ethics and Epistemology: A Defense of Normative Realism (2004), that contained a similar argument (and it is available for free on his website). Nobis argues the following:

  1. We have good reason to accept epistemic realism.
  2. We don’t have good reason to accept epistemic irrealism.2
  3. The strongest arguments available against moral realism are weak because they could all be used just as plausibly against epistemic realism, but they don’t give us a good reason to reject epistemic realism.


Nobis defends a version of epistemic realism he calls the “deontological equivilances” interpretation of “epistemic deontology” and rejects the available versions of epistemic irrealism. However, epistemic reductionism is largely ignored and such a view could be plausible.

Additionally, Nobis makes it clear that moral realism is similar to epistemic realism in every way that people take offense to moral realism. People argue that morality is just a matter of taste or just a human invention based on the fact that moral facts seem to be mysterious, cause motivation, can’t be known by everyone equally, and so on. Nobis argues that epistemic facts equally seem to have these same objectionable features. Thus, these so-called objectionable features are not adequate reason to reject epistemic or moral realism.3

Nobis argues that we have more reason to accept epistemic realism than reject it, so the arguments against moral and epistemic realisms are implausible. The premises used to reject epistemic realism are more likely false than the possibility that epistemic irrealism is true—and the arguments against epistemic realism are analogous to the argument against moral realism. Therefore, the arguments against moral and epistemic realism are weak. Although it is possible that both moral and epistemic realisms are false, that is unlikely given our current information.

I will offer various objections and contributions to the arguments Nobis discusses including one major objection—Nobis largely ignores the irrealist position we can call “reductionism.” If we understand “epistemic reductionism” the same way that Cuneo does, then it is the view that epistemic facts aren’t overriding. We need to know why this view is implausible.4

We have good reason to accept epistemic realism.

Nobis argues that that we have good reason to accept epistemic realism insofar as it is intuitive, and at least one version of epistemic realism is plausible.

Epistemic realism is intuitive.

The reason that epistemic realism is plausible is because we have platitudes (very plausible common sense beliefs) that seem to require epistemic realism (6-7):

  1. It’s good to have evidence for one’s beliefs.
  2. All else being equal, it’s better to have consistent beliefs than inconsistent beliefs.
  3. It’s bad to be unreasonable.
  4. You shouldn’t believe something unless you have [sufficiently] good reasons to believe it.
  5. The strength of one’s belief ought to be proportional to the strength of the evidence.

These plausible beliefs seem to confirm the fact that there are epistemic facts. It is true that some people are unreasonable because there are facts (properties) about the people that make it true. They also seem to confirm that epistemic facts are potentially overriding. It’s “better” to be reasonable than not because epistemic values are “worthy” of consideration.

Epistemic realism is plausible.

Nobis argues that a plausible form of epistemic realism is the “deontological equivilances” interpretation of “epistemic deontology” (33-59). This interpretation states that if a belief is justified, then there is a positive deontic status (i.e. you ought to believe it); and if it has a positive deontic status, then it is justified (45).5 Additionally, if a belief is unjustified, then there is a negative deontic status (i.e. you ought not believe it), and if a belief has a negative deontic status, then you ought not believe it (ibid.). To say that you should believe something is to say that it is our duty to—not just because it would be better to believe it (ibid.).

Nobis defends his version of epistemic deontology from two major objections:

One, ought implies can (47-48). We can’t choose our beliefs, so we can’t have a duty to have certain beliefs. Nobis replies with the following:

  1. We can be obligated to do something, even if we can’t. For example, we can have obligations to pay our bills on time, even if we can’t (47).
  2. We can decide what we believe to some extent. By keeping an open mind and learning to think clearly, we can have better beliefs (ibid).
  3. It’s self-defeating to reject deontology based on the fact that “ought implies can.” First, if it’s true that “ought implies can” and we can’t stop believing in deontology, then we can’t say “we should reject it” because we couldn’t choose to reject it (48). Second, if deontology is false because “ought implies can,” then it’s false that we “ought not accept it” because deontology is necessary to say you ought or ought not believe something (ibid.).

Two, it’s too demanding to say that you ought to have a belief because it’s a duty (49-50). It could be better to do one thing instead of another, even if it’s not a duty. If something is a duty, then it’s wrong not to do it. For example, it’s better to cure cancer than spend our time watching television, but we don’t all have a duty to cure cancer.

Nobis responds to this objection by saying that there is no less demanding “ought” to describe the importance of having a justified belief rather than an unjustified one (49). If a belief is sufficiently justified, then it’s okay to have that belief; but if a belief is unjustified, then it’s wrong to have that belief. Nobis argues that if you don’t do your moral duty, then you should do something “intuitively morally neutral;” but there is no epistemic equivalent—there is nothing “epistemically neutral” we can do rather than having a justified belief (ibid.). Refusing to have a justified belief by “suspending judgment” would not be appropriate.

Imagine that you see your hand in front of your face while fully alert, sober, and sane. It is justified to believe that you have a hand, and it would be unjustified to believe that you don’t. To suspend judgment and refuse to have any belief would be absurd.

We don’t have good reason to accept epistemic irrealism.

The fact that epistemic realism is intuitive and seems plausible is already based in part on the counterintuitive and absurd results of epistemic irrealism. In particular, there is nothing you really ought to believe if irrealism is true. No one would be irrational, no belief would be justified, it would never be “right or wrong” to have a belief. Additionally, Nobis already pointed out that epistemic irrealism seems self-defeating insofar as it is false that we should believe it—and insofar as there can be no reason to believe it. Nonetheless, Nobis considers two actual forms of irrealism proposed by philosophers in particular, and he briefly discusses something like reductionism.

Allen Gibbard’s epistemic expressivism

Allen Gibbard proposes that we accept epistemic expressivim and suggests that our positive evaluations are made “to express one’s acceptance of norms (or commands or imperatives) that permit it” (235). To say that “belief in God is unjustified” is actually just an expression similar to “norms that justify the belief in God, boo!” This is true in both the moral and epistemic domains. To say that a belief is justified is actually to express our acceptance of the norms that would make it justified. For example, to say that I am justified to believe that I have a hand while I look at it right in front of my face is to express that I accept the norm, “Let’s believe what we see!”

Gibbard is an epistemic naturalist (a person who thinks that we should only believe what could be known through science), so he rejects non-naturalism (the view that we can know self-evident truths) with little argument (237). However, Gibbard also rejects that we can know evaluative truths though naturalistic methods because doing so would eliminate the “endorsement” aspect of our evaluations (236). When we say that a belief is justified we think people should do it, but Gibbard seems to think that it’s impossible to find out something should be done based on observation and experimentation.

Once naturalism and non-naturalism have both been rejected, Gibbard argues that we are only left with the possibility that there are no evaluative facts—because our evaluative judgments aren’t meant to state facts in the first place.

Nobis reminds us that Gibbard’s irrealism will be self-defeating, as usual. He has not and could not possibly give us reasons to support epistemic expressivism. After all, what we believe to be “reasons” are actually just endorsements (239). The fact that Gibbard endorses expressivism doesn’t tell me what I should believe.

Gibbard might be right that norms he endorses would license expressivism, but why should we think his norms are better than others? In fact, Nobis thinks it is plausible to think that some norms are better than others, but expressivism couldn’t possibly allow us to come to that conclusion (241). Gibbard could reject valid logical inferences, such as modus ponens, but we know that such norms will be inferior to the alternative. They will lead to falsehood and irrational thinking.

Additionally, there are going to be other counterintuitive consequences of accepting Gibbard’s expressivism that Nobis didn’t mention:

  1. It seems obvious that when I say that rejecting valid inferences is unjustified that I am not just expressing my endorsement of a norm. Gibbard seems to make it impossible to ever do more than that despite the fact that it seems quite easy to actually say that one norm is superior to another. My point is that Gibbard discounts our actual experiences to the point of absurdity.
  2. We debate about what we should believe. For example, we want to know, Should I believe in God? We can then argue about it. However, Gibbard makes such a debate impossible because we would actually just express our endorsements. No one’s expression would be better than another.
  3. We debate about what we should believe and what norms we should accept, but Gibbard’s expressionism would suggest that many of these “debates” involve two people talking past each other when each person endorses different norms. We couldn’t possibly talk about the same thing when we don’t endorse the same norms. One belief would be something I would accept if I endorse valid inferences, but it would be something I would reject if I don’t.

Hartry Field’s epistemic relativism

Hartry Field proposes a sort of epistemic relativism he calls “non-factualism.” He agrees that there are evaluative truths, but they aren’t based on facts (evaluative properties). Instead, they are based on an epistemic framework. If I have an epistemic framework that demands no evidence to have beliefs, then all of my beliefs are justified. However, if I have an epistemic framework that is very demanding, then many of my beliefs would be unjustified.

Field argues that we should endorse relativism because it is less mysterious than epistemic realism and it can explain all our epistemic experiences just as well as epistemic realism (243). (In other words Field thinks that it is simpler and conforms to Occam’s Razor. Immediately we might wonder why Occam’s Razor is a good characteristic for a theory to have.)

Field admits that he is personally interested in knowing the truth and he accepts intellectual norms that “lead to more truth and less falsehood” (245). However, it isn’t clear that everyone has a strong interest in finding the truth or that epistemic relativism is true.

Field also admits that some epistemic frameworks are better than others, but it’s not clear why he thinks that. In particular, we would expect him to find a “better” framework to be one that is only better given a certain epistemic frameworks. I suppose some epistemic frameworks could be self-rejecting in the sense that they enable us to realize that there are better alternatives out there. However, we might find Field’s irrealism itself to be a self-rejecting inferior framework.

Nobis argues that an epistemic realist has accepted a particular epistemic framework and will find no reason to change frameworks to Field’s irrealist framework because the realist does not accept Field’s norms (246-247).

However, I think that the moral realist could accept the norm of simplicity and could be tempted to be a relativist based on Field’s argument. That’s not to say that the norm of simplicity really does favor Field’s irrealism. In particular, I don’t think Field’s irrealism can account for our epistemic experiences as well as realism can:

  1. Realism makes sense of the fact that we should accept certain beliefs irrespective of our framework. For example, we should accept valid logical inferences.
  2. Realism makes sense out of the plausible belief that some frameworks are better than others, but relativism doesn’t.
  3. Realism makes sense out of the plausible belief that we genuinely argue with people who have differing norms about what we should believe. We aren’t talking past each other. When I say that we should believe that it’s irrational to believe 1+1=3, I’m not saying that it’s only justified for people like me who share my norms. I’m not talking past someone who argues that 1+1=3 is a justified belief given their personal norms.


Reductionism is the idea that what we ought to believe reduces to something else, such as what would help us accomplish our goals. In particular, the goal of knowing the truth seems relevant. If a justified belief is more likely true than an unjustified belief, then we might not need epistemic realism. Instead, we just say “you should believe 1+1=2” based on the assumption that the other person wants to know the truth. Reductionism seems to lack overriding duties or reasons, but there is a sense that it can be overriding for a reductionist. A reductionist could argue that justifications are overriding reasons to believe something insofar as the only relevant information for justified beliefs is whether or not they are probably true. If a belief makes you feel good (such as the belief, “Unicorns exist”), that is irrelevant to whether or not a belief is justified insofar as it doesn’t tell me whether unicorns probably exist or not. When exactly a belief is justified is controversial, but we could start with the following suggestion—A belief for a reductionist can be sufficiently justified when the belief has a good chance of being true (as far as we know). If we have reason to think our belief doesn’t have a good chance at being true (or is probably false), then it would be unjustified.

Nobis doesn’t talk about reductionism as such, but he does talk about it indirectly when he discusses that an epistemic irrealist can point to a belief being “probably true” and say we “ought to believe it” based on that fact (96-97). We know that we sometimes say we “ought to do something” based on our goals. This is known as an “instrumental ought.” For example, you ought to drink water if you want to stay alive. Epistemic oughts can be based on what we ought to believe if we want to know the truth.

Nobis argues that an irrealist can point to the logical relations of various beliefs or statements, and they can show when some propositions are inconsistent with others (ibid.).6 He says that we might then read epistemic arguments and ask ourselves, “’Why is [so and so] telling me this, even if what [he or she] says is true? What is supposed to follow, in terms of what I believe, from my recognition of this natural feature?’ They would suspect that [so and so] is giving premises to an argument” (ibid.). Nobis then claims that there is no clear answer to this question from the reductionist because the logical relations can be dealt with in different ways. Given an argument that shows two beliefs to be inconsistent, we could keep one belief or the other, or reject that there is inconsistency (ibid.)

A reductionist could reply to Nobis that what follows is based on what a person ought to believe who wants to know the truth. That is a common understanding between people who read epistemic arguments. Someone who doesn’t want to know the truth would probably not read such arguments or debate with others about epistemic evaluations. Nobis doesn’t seem to take seriously the idea that epistemic judgments could be instrumental based on the assumed goal of attaining knowledge.

Nobis’s argument continues, but he doesn’t make it entirely clear why reductionism should be rejected. This is a major weakness of his thesis. (Cuneo’s argument suffers from the same weakness.) I am personally not sure that reductionism should be rejected, but I have the following suggestions that could help us find a reason to reject it:

  1. There are counterintuitive implications to reductionism. In particular, we might think that we ought to believe something because it’s probably true whether you care about attaining knowledge or not. This might be based on the intrinsic value of knowledge itself.
  2. A reductionist wants to tell us which beliefs are “probably true,” but there is something mysterious about knowing what is “probably true.” It’s not entirely clear how to know how strong an argument is or how likely it is for a conclusion to be true. It’s also not entirely clear why a belief that conforms to the value of certain theoretical virtues, such as Occam’s razor, are more likely to be true. Perhaps it is easier to explain why a belief is “probably true” given epistemic realism.

The strongest arguments available against moral realism are weak

Nobis considers arguments against moral realism given by A. J. Ayer, C. L. Stevenson, R. M. Hare, J. L. Mackie, and Gilbert Harman and finds that their arguments threaten epistemic realism in the same way. Additionally, the epistemic irrealist “alternatives” seems to be less plausible than epistemic realism. He then concludes that moral and epistemic realisms are more plausible than irrealism even when considering the strongest arguments against them that are currently available. Nobis argues that all the forms of irrealism are implausible for similar reasons and I have already discussed them above. I will grant that epistemic realism is plausible for the sake of argument and discuss the various arguments against moral realism discussed by Nobis:

  1. Positivism requires us to reject moral realism. (Moral facts aren’t observable.)
  2. Moral judgments provide us with motivation, but that is unacceptable for moral facts.
  3. Self-evidence is unacceptable, but that’s the only way we could know moral facts.
  4. It’s hard to see what moral facts consist in.
  5. Moral naturalism and intuitionism are both unacceptable, but one of these theories must be true if moral realism is true.
  6. People disagree about moral facts even when they agree on all the nonmoral facts, and that suggests that there are no moral facts.
  7. Moral facts would be radically different from all other parts of reality, but we can explain all our experiences without believing in moral realism.
  8. Morality is supervenient on nonmoral facts, but that is too mysterious to accept.
  9. Moral demands must be acceptable to those they apply to, which suggests that different moral demands are acceptable to different people.
  10. Moral facts don’t explain or cause anything, so we should reject their existence.

1. Positivism requires us to reject moral realism.

Logical positivism stated that a belief must be analytic (true by definition) or empirically verifiable (or falsifiable), or it is meaningless. If a belief isn’t true by definition and it has no implications to our observations, then it doesn’t mean anything. A. J. Ayer argues that moral realism must be rejected because it fails to fulfill the demands set out by logical positivism:

  1. If morality is compatible with logical positivism, then moral facts must be known through observation or some other way. These two possibilities have been explored by moral realists describing their realism as a form of “naturalism” or “non-naturalism.”
  2. Naturalism claims that moral facts are identical with nonmoral facts, but that isn’t something you can know through observation (63-64).7 For example, “pain” doesn’t seem to mean the same thing as “bad.” It is possible that it does, but it isn’t something I can prove through observation or experimentation. Therefore, naturalism violates the demands of logical positivism.
  3. Non-naturalism claims that moral facts are irreducible to nonmoral facts, but Ayer thinks that such facts could only be known though intuition, which would be a non-empirical form of knowledge—and it would therefore violate the demands of logical positivism (64).8
  4. Therefore, moral facts are incompatible with logical positivism.
  5. If moral facts are incompatible with logical positivism, then we should reject them.
  6. So, we should reject moral facts.

Nobis only challenges premise 5—he doesn’t find logical positivism to give us plausible epistemic demands. We seem to know things that are incompatible with logical positivism. In particular, epistemic realism seems to be incompatible with logical positivism for the exact same reasons as moral realism (70-71). To say a belief is rational or justified doesn’t seem to be something we can observe. In fact, such facts seem to be known intuitively (without observation). Logical positivism isn’t a good reason to reject epistemic realism, so it’s not a good reason to reject moral realism either.

Additionally, Nobis mentions that logical positivism is self-defeating. We can’t observe that beliefs are meaningless unless they are analytic or empirical. Therefore, logical positivism violates its own demand because it is neither analytic nor empirical.

2. Moral judgments provide us with motivation.

C. L. Stevenson argues that moral facts can’t exist because moral judgments provide us with motivation. He thought that only desires rather than beliefs could provide us with motivation, so he rejects that we even have moral beliefs in the first place (100-101). Imagine that I think it is bad to feel pain. I certainly would seem insincere if I said pain is bad, then caused myself a great deal of pain for no good reason. It would seem that my judgment that “pain is bad” involves my desire to avoid pain.

Nobis finds it unlikely that moral judgments are necessarily motivating and thinks it is possible to feel no motivation to do what we believe is right. However, he agrees that there is at least some plausibility to think that there is a connection between moral judgments and motivation. Unfortunately he thinks it is just as plausible to think that epistemic judgments are motivational. Imagine that I say that it is completely irrational and unjustified to believe in bigfoot. It would seem like I am insincere if I said this and believed that bigfoot exists. Therefore, if motivation gives us good reason to believe that moral judgments are desires, it also gives us good reason to believe that epistemic judgments are desires. However, this is absurd given the plausibility of epistemic realism. The motivational character of epistemic judgments isn’t a good reason to reject epistemic realism, so it’s not a good reason to reject moral realism either.

3. Self-evidence is unacceptable.

It seems possible to think that moral facts are self-evident, which is to say that we could know them by understanding the concepts involved. This possibility is disregarded by C. L. Stevenson because self-evident truths “seem upon examination to be deliverances of [one’s] respective reasons only” rather than those of others (101). I suspect that Stevenson thought that self-evidence was little more than one’s prejudice or dogmatic conviction.

Nobis doesn’t say very much about whether it is plausible to think moral or epistemic facts are known through self-evidence, but what he says about intuition below might be applicable.

4. It’s hard to see what moral facts consist in.

Stevenson argues that moral facts seem too mysterious. He “find[s] nothing ‘out there’ for his attitudes to represent” (101).9 I suppose moral facts are taken by him to be so mysterious that we might suspect that they are mythological similar to faeries and ghosts. Nonetheless, Nobis makes it clear that epistemic facts are similarly mysterious. It’s not at all clear what “out there” could make a belief justified (102). After all, a justified belief could be false. Again, this argument is too ambitious. It’s implausible to think we should reject epistemic realism just because it’s mysterious.

5. Moral naturalism and intuitionism are both unacceptable.

R. M. Hare argues that both moral naturalism and intuitionism are both unacceptable, but they are the only two sorts of moral realism to choose from. If they are both unacceptable, then moral realism should be rejected. Naturalism is unacceptable because it would make moral disagreement too difficult and intuitionism is unacceptable because it would make it impossible to settle moral disagreement. Both of these views were rejected by A. J. Ayer because he believed that they are both incompatible with logical positivism.

Moral naturalism

Hare rejects moral naturalism for four major reasons: (1) if it’s true, then many moral debates would be too difficult; (2) the open question argument; (3) the motivational pull of moral judgments is ignored by naturalism; and (4) naturalism collapses into relativism.

1. If it’s true, then many moral debates would be too difficult (111-115). Moral naturalism claims that moral facts are identical to certain nonmoral facts, but not everyone agrees about what nonmoral facts they are identical to. Imagine that people from two cultures, culture A and B, meet for the first time and debate about whether lying is always wrong or not. People from culture A argue that it’s not always wrong to lie, but people from culture B argue that it is always wrong to lie. However, people from culture A believe that “wrong” means “causes more pain than the alternatives” and people from culture B believe that “wrong” means “tends to deceive others.” The people from the two cultures are actually talking past each other. However, if this debate happened in real life, we would know that they aren’t talking past each other. The word “wrong” means the same thing for both cultures. Therefore, moral naturalism leads to a counterintuitive result that contradicts our knowledge concerning moral debates. Moral naturalism makes genuine moral debate much more difficult than it really is. Hare seems to assume that no one could legitimately say of others, “You are mistaken in your moral evaluation because you do not accept my definition of the moral terms” (114).

Nobis argues that there are epistemic naturalists that seem to have the same problem. Someone might think “justified” means “produced by a reliable belief-forming process” and another might think it means “highly probable relative to the believer’s observation” (115). This could lead to the same problem of talking past each other that was discussed above. However, this in itself just seems to lead him to reject epistemic naturalism.

I am not convinced that Hare’s argument against naturalism is plausible. First, perhaps we can know the word “wrong” prior to learning about the natural equivalence. It might be that when a person says that something is wrong “because it causes more pain than the alternatives” that equivalence doesn’t reduce the word “wrong” to “it causes more pain than the alternatives” anymore than saying “water is H2O” reduces the word “water” to “H2O.” Both words have a different meaning and can be understood separately, but they are also the same thing. We might then disagree with someone about whether a liquid is water who doesn’t yet know that water is H2O but we wouldn’t be “talking past each other.” The other person simply lacks a deep understanding about water.

Second, naturalists have traditionally assumed that moral facts are only “natural” if they are “reducible to nonmoral facts,” but I find it implausible to think that all natural facts reduce to other facts. My thoughts don’t seem to reduce to brain states. They are caused by brain states, but they are still natural facts. Many natural facts could be irreducible in this way including moral facts. I know that I have thoughts because I experience them and I know that pain is “bad” from experiencing it. There is no need to reduce such potentially natural facts to non-psychological or non-moral facts.

2. The open question argument (115-116). The open question argument is the suggestion that any natural equivalence of moral facts could be legitimately questioned. If I suggest that “wrongness is whatever causes more pain than the alternatives,” then someone might ask, “But ought I to refrain from causing pain?” Nobis suggests that epistemic naturalistic equivalences suffer from the same objection. To say that “justification” means “caused by a reliable belief forming process” can be questioned by someone who asks, “But ought I use a reliable belief forming process?”

The open question argument seems to suggest that it is difficult to prove that an equivalence exists. Whatever equivalence we suggest can be questioned and we might feel uncertain about whether or not the equivalence has been found. However, I don’t find the open question argument to be persuasive because it might be possible to prove that an equivalence exists. We think that water is H2O. Someone might ask, “This is water, but is it H2O?” That question in no way seems to prove that water isn’t likely to be H2O. How exactly we know that water is H2O is an interesting question, but I assume that it can be answered. I suspect that a “careful examination” would be involved.

3. The motivational pull of moral judgments is ignored by naturalism (116). I already discussed this objection to moral realism above, so there is no reason to discuss it again.

4. Naturalism collapses into relativism (117). Hare suggests that naturalism collapses into relativism because everyone will merely use the word “wrong” in different ways. There will be no single correct meaning to the word “wrong,” but how a group uses the word will influence who is thrown into prison and so on. Nobis correctly replies that this objection is unpersuasive because “Hare seems to assume… that on any naturalism the majority’s view on the truth conditions of a term determines the truth conditions of the term, or somehow makes the analysis true by definition and thereby unquestionable” (ibid.). However, it is quite possible that “wrongess” really is equivalent to certain nonmoral facts. People who use the word the wrong way are wrong about what “wrongness” is. (Not to mention that I already discussed that the meaning of “wrong” can be understood to some extent before learning what nonmoral equivalence is involved.)


Hare rejects intuitionism because intuitionism can’t resolve disagreement (119-122). Hare argues that intuitionism requires us to have an ability to “just see” that something is right or wrong, but we have differing “intuitions.” Some people “just see” that lying is always wrong and others “just see” that lying is sometimes the right thing to do. Additionally, intuitionism requires that our intuitions are infallible. Therefore, people disagree with each other based on what they “just see” to be true and there is no way to resolve their disagreement because each person’s beliefs are infallible.

Nobis notes that intuitionists clearly don’t have to think that their intuitions are infallible (ibid.). They can admit that someone’s intuition has to be false when people come up with conflicting intuitions. It could be that people have differing intuitions because intuition often requires a certain degree of maturity and education.10

Nobis also argues that if we grant that intuitionism can’t resolve disagreements based on Hare’s objection, then we will have just as much reason to grant that epistemic intuitionism should be rejected based on an analogous argument (121-122). It’s not entirely clear how we can know that a belief is unjustified without intuition, but it’s also not entirely clear how intuitionists can resolve disagreements about when a belief is justified. For example, there is some evidence that bigfoot (or some other unknown great ape) has been leaving footprints around, but people disagree about whether the evidence is “sufficient” to have a justified belief in bigfoot.

6. People disagree about moral facts even when they agree on all the nonmoral facts.

J. L. Mackie (and Gilbert Harman) argue that the amount of moral disagreement that exists suggests that moral realism is false (149-176, 196). This is known as the “argument from disagreement” or “argument from relativity.” Mackie thinks that there is “radical disagreement” concerning moral truth and there wouldn’t be such disagreement if moral properties existed (149-150). It’s not entirely clear how much disagreement there is and whether or not moral disagreements are based on disagreements concerning “nonmoral facts.” Mackie seems to think that there would be disagreement about moral facts even if we agreed on all the nonmoral facts.

The assumption that Mackie is making is that moral properties would be “equally salient to everyone” (165). If moral properties exist, then they would be equally known or perceived by everyone. However, if they exist, then they are not equally accessible to everyone. Mackie decides that it is more likely that moral properties don’t exist than the possibility that moral properties are unequally accessible to people.

Nobis suggests that it’s not clear that moral properties, if they exist, are probably equally accessible to everyone. Perhaps moral properties can be seen clearly only by those who have undergone decades of philosophical or moral training” (165) or “[p]erhaps moral properties can be seen more clearly only when one has had traumatic personal experiences of (what many would call) evil” (166). “Some people aren’t well placed to see true answers to moral questions… Some people just can’t ‘see’ mathematical or logical truths, and perhaps there are people who just can’t see moral truths (or epistemic truths) either” (167). There are reasons that some properties in the world can be seen by some people better ta others.

I would also want to suggest that we often find moral properties to be inaccessible. When that happens it isn’t entirely clear what one ought to do. People might often disagree about moral facts because they have moral beliefs despite having insufficient information to have certainty. Moral properties seem to be immeasurable and incommensurable. Suppose you have a choice to save the life of an elephant or a dolphin. Which life is more important? It’s not clear what the answer is, but some people might have an opinion one way or the other.

Nobis also argues that the the breadth and depth of moral disagreement is unclear (170). We might expect that moral properties, if we have any reason to believe that they exist, would be at least somewhat accessible to people—but they might be. People do tend to agree that various moral beliefs are “almost certainly true,” such as the belief that torturing babies is wrong.

Finally, Nobis argues that if we have any reason to think that moral properties must be equally accessible to people for moral realism to be true, and there is too much moral disagreement for moral realism to be true; then we seem to have equal reason to reject epistemic realism for the same reason (173). People seem to disagree about whether certain beliefs are justified even when they agree on all the nonepistemic facts, and it is unclear how (or whether) such disagreements can be settled. However, this is absurd. The fact that people disagree about epistemic facts doesn’t seem to be a good reason to reject he existence of epistemic facts entirely.

I would add that moral and epistemic disagreement is a poor reason to reject moral or epistemic realism because if we know a single moral or epistemic fact to exist, then realism is proven to be true. I think it is plausible that at least one such fact can be known with confidence.

7. Moral facts would be radically different from all other parts of reality.

This is what J. L. Mackie calls the “argument from queerness” (177-187). Moral facts require there to be a completely different part of reality than the other parts, and we can only know moral facts with a completely unique sort of perception (a “special faculty”) different from all other ways of knowing things.11 If all our experiences can be explained without positing a new sort of reality or perception, then we should prefer to do so. Therefore, we should reject moral realism because it’s too ambitious. It requires metaphysical and epistemological commitments that we should prefer not to make.

Nobis argues that the argument from queerness seems to apply equally well to epistemic facts. If epistemic facts exist, they are totally different from other facts, and (just as plausibly) we could only know them with a “special faculty.”12

Additionally, if moral facts and epistemic facts both face the argument for queerness in sufficiently similar way, then moral facts wouldn’t be so unique or “queer” because it might be the same sort of fact as epistemic facts. If both moral and epistemic facts require a totally different sort of reality (the normative or evaluative realm), then normative facts require the same sort of reality—and if moral facts can only be known through intuition (a special faculty), then it seems plausible to think that epistemic facts can also only be known the same way.

8. Morality is supervenient on nonmoral facts.

J. L. Mackie argues that morality is supervenient on nonmoral facts—that the nonmoral facts are relevant to morality (such as experiencing pain); and if the moral facts are different, then the nonmoral facts must be different. If it’s wrong for me to hurt someone one day because I have no reason to hurt that person, but it’s right for me to hurt someone the next day; then there must be something different in my situation that makes one action right and the other wrong. Mackie argues that supervenience suggests that moral realism is false because it’s too mysterious (179-182). There must be a reason that moral facts are supervenient on nonmoral facts, but Mackie says that he doesn’t have any idea what “in the world could be signified by this ‘because?’” (179).

Nobis notes that mental and epistemic facts both seem to be supervenient on other sorts of facts as well. Our minds seem to depend on our brain states, and epistemic facts depend “on psychological facts about [a] believer’s mind (e.g., her perceptions, memories, thoughts, and so on) or on the processes that her beliefs resulted from or some other non-epistemic features of the situation” (180). At the same time it would be absurd to think there are no epistemic or psychological facts based on the fact that their supervenience is “mysterious.”

9. Moral demands must be acceptable to those they apply to.

Gilbert Harman argues that moral realism should be rejected because it requires us to reject the principle that “ought implies can.” In particular, he thinks that (1) not everyone can rationally accept the same moral principles, so they ought not; and (2) “it can be rational for two different people to accept different moral demands ‘all the way down’” (190). Two people can have totally different moral principles.

1. Not everyone can rationally accept the same moral principles, so they ought not.

Should we agree with Harman about this? Nobis thinks not. He argues that we often think people should do what they have no personal reason to do (or ought not do what they have no personal reason to not do) (191). For example, we would judge a child molester as doing evil acts, even if that person had no reason to accept our moral principles. Some people seem to have wrong moral beliefs, and there can still be objective moral facts that they don’t know about.

2. It can be rational for two different people to accept different moral demands ‘all the way down.’

Why would anyone accept this principle? Harman finds it intuitive because he thinks there are successful criminals, cynical politicians, business executives etc. that may have no reason to care about strangers (190-191). Although Harman seems to think these sorts of people support his conclusion, Nobis argues that most people will think the opposite—they are examples of evil people who have no reason within their inferior “moral framework” (set of moral assumptions) to do what is objectively right.

Nobis then argues that if we give Harman the benefit of the doubt, epistemic realism should be rejected for the same reason (192-193). Different people have different “epistemic frameworks” (set of epistemic assumptions) that cause them to have differing epistemic beliefs, and it is theoretically possible that to people could accept totally different epistemic frameworks “all the way down.” At the same time this is implausible. Some beliefs really are better than others, and some epistemic frameworks are inferior to others. It seems plausible that there are epistemic facts and some people simply know less about these facts than others. In fact, we seem to know that a “foolish person’s beliefs don’t become justified by his thinking that ‘anything goes’ in terms of belief” (195).

Harman suggests that relativism is plausible for a domain when the following criteria are met (209):

  1. There are differences of opinion about when to apply an absolute notion.
  2. We have no way to resolve these differences.
  3. We can see how these differences could arise from differing frameworks (“different salience of certain relations.”)

However, Nobis argues that all of this criteria can equally be applied to epistemic realism, but it would be absurd to reject epistemic realism using such criteria (209-210).

Harman suggests that epistemic relativism should be rejected because (1) epistemic frameworks could be based on human nature and (2) only adequately good epistemic frameworks would survive (210-211). We know a belief is justified based on our instincts, and people with faulty epistemic instincts die off. Of course, Nobis is unpersuaded and argues that people who have faulty epistemic instincts (or inferior epistemic frameworks) haven’t died out (211).

10. Moral facts don’t explain or cause anything.

Gilbert Harman argues that we should reject moral realism because moral facts don’t explain or cause nonmoral facts (212-213). Some people have argued that Harman is wrong and such facts can cause and explain nonmoral facts, but Nobis grants that it is at least initially plausible to think that moral facts don’t cause anything to happen. At the same time Nobis thinks it is equally plausible to think that epistemic facts don’t cause or explain anything either, but it would be absurd to reject epistemic realism just because epistemic facts don’t cause anything to happen (213-219). Therefore, the fact that something doesn’t cause anything to happen (or explain anything) isn’t a good reason to reject its existence.

Additionally, Nobis argues that although some people would suggest that our beliefs are only justified if they explain something, this “test” is self-defeating because the test itself doesn’t explain anything (220).


Nobis succeeds in showing how epistemic realism can be plausible, he argues well that all of the arguments against moral realism seem to apply equally well against epistemic realism, and he shows that a great deal of irrealist literature is unpersuasive. Nonetheless, we still have to wonder if the epistemic irrealist position known as “reductionism” is a worthy alternative to epistemic realism. If we should reject epistemic realism in favor of epistemic reductionism, then we might also have a good reason to reject moral realism.


1 Epistemic facts are facts about justification, rationality, reasonableness, and knowledge. If there are no epistemic facts, then people can’t be rational or reasonable, or have knowledge or justified beliefs. “Epistemic realism” is the view that epistemic facts exist.

2 Many people call the view that moral facts don’t exist “moral anti-realism” and the view that epistemic facts don’t exist “epistemic anti-realism.” However, there is a growing trend of now calling such things “irrealism” rather than “anti-realism,” and Nobis calls it “irrealism.”

3 Cuneo argued that both moral and epistemic facts can be overriding, but I didn’t find his argument to be entirely convincing. More needs to be said about the issue.

4 Cuneo discussed epistemic reductionism at length, but I didn’t find his argument against it persuasive.

5 This is not to say that a belief is justified because you ought to believe it. Why a belief is justified is another matter entirely.

6 This argument was actually against a sort of emotivist epistemic irrealism that argues that epistemically evaluative terminology are merely expressions of our feelings.

7 I don’t understand why we could only observe moral facts if they are identical to non-moral facts. This position seems to require the assumption that we can’t observe moral facts as such. We seem to observe psychological facts (thoughts, feelings, etc.) even though we don’t necessarily think they are reducible to non-psychological facts.

8 To even suggest that irreducible facts are “non-natural” is to use loaded words. Certainly we don’t want to commit ourselves to thinking morality or thoughts are “supernatural.” I prefer to call these views “non-reducitionistic” or “emergentist” rather than “non-natural.” I suspect that we experience lots of irreducible facts. Chemistry doesn’t necessarily reduce to physics, and certainly psychology doesn’t seem to reduce to physics. At the same time we experience psychological facts (thoughts, desires, etc.) I don’t think we can dismiss the possibility that moral facts can also be known in this way, but Nobis remains silent concerning this issue.

9 The same objection was made by J. L. Mackie who says he “cannot frame for himself any clear, detailed, picture of what it would be like for [moral] values to be part of the fabric of the world” (182).

10 See my review of Robert Audi’s The Good in the Right for more information.

11 Part of the argument from queerness asserts that moral facts are somehow motivational, but it seems just as likely that epistemic facts are as well. I already discussed this argument above.

12 Robert Audi doesn’t think intuitive or self-evident truths must be known with a special faculty. Instead, they can be known through reflection.

but they don’t give us a good reason to reject epistemic realism.


  1. Thanks for this. I hope to read and comment soon. I can point to a 2000 word (!!) version of the main argument from my dissertation here:

    “Moral Nihilism, Intellectual Nihilism & Practical Ethics”

    Nathan Nobis, Ph.D.
    Philosophy, Morehouse College, 830 Westview Dr. SW, Atlanta, GA 30030
    Community Health and Preventative Medicine, Morehouse School of Medicine, Atlanta, GA
    Animal Studies, Humane Society University, Washington, DC

    Comment by Nathan — November 24, 2010 @ 1:47 pm | Reply

  2. I was hoping that you could do a similar review of David Benatar’s “why it is better never to come into existence”. Most of the points given against this particular paper are always given from a moral irrealist point of view. I was hoping to hear your thoughts on it, and some possible objections moral realism has towards his view.

    Comment by Ilias — December 17, 2010 @ 5:37 pm | Reply

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Create a free website or blog at

%d bloggers like this: