Objections to Moral Realism Part 1: The Is/Ought Gap

Although I have already discussed several objections to moral realism, some of them are worth discussing in more detail. In particular, the is/ought gap has proven to be a source of confusion. The is/ought gap is ambiguous and there are at least two main interpretations: One is ontological and one is epistemological. In other words, one says that the is/ought gap is a description of reality and another says that it is a description of our evidence.

Here “is” refers to descriptive facts (nonmoral facts) and “ought” refers to prescriptive facts (moral facts). The idea of there being an ontological gap is that there is something different about description and prescription and one domain is not the same thing as the other (one domain is not reducible to another). The idea of there being an epistemological gap is that we can’t know prescriptive facts from descriptive facts. Both kinds of is/ought gaps require that we accept that something is in the “is” domain or the “ought” domain. Nothing can be in both domains.

I will discuss the following:

1. David Hume’s discussion of the is/ought gap
2. John Searle’s discussion of the is/ought gap.
3. Lawrence Becker’s discussion of the is/ought gap.
4. Intrinsic values and the is/ought gap.
5. The ontological interpretations of the is/ought gap.
6. The epistemological interpretations of the is/ought gap.
7. Two ways people have used the is/ought gap as an argument against realism.

The purpose of this paper is to consider the arguments against moral realism. In particular, I will discuss these two objections to realism:

1. The ontological argument: Moral realism requires us to accept a new irreducible kind of property, but such a property isn’t necessary. Instead of accepting irreducible moral properties, we should just admit that we are deluded about morality.
2. The epistemological argument: We can’t know about moral facts through observation, but that’s how we know about everything. Therefore, moral knowledge is impossible.

These arguments will be discussed when I discuss the different interpretations of the ontological and epistemological is/ought gap because the arguments can be understood in various ontological and epistemological ways.

1. Hume’s Discussion of the Is/Ought Gap

The is/ought gap is famously introduced by Hume, who presents us with the challenge: “How do you get an ‘ought’ from an ‘is?’”¹ Hume realized that many people argued about what is the case in order to argue what ought to be the case. This kind of argument implied that we could get what ought to be the case from what is the case, but it wasn’t yet clear how it could be done. Some moral realists have offered answers to this question by explaining how we can observe moral facts, but they can only do so with a moral theory. One way might be to merely attempt to explain moral observations we have in order to discover the moral theory that they imply. However, I have discussed a different answer to that question: We experience moral facts, similar to how we experience psychological facts. We can develop moral theory based on how we experience our final ends (benefits), and then by justifying the fact that other people will have similar experiences and final ends. Everyone’s final ends matter, not just our own.

But Hume didn’t just say that he wanted to know how to get “ought” from “is.” He also discussed that prescriptive and descriptive facts seemed quite different. Hume states that sentiments are not subject to truth or falsity, and morality seems to require sentiments. This implies that it is impossible to get “ought” from “is” because moral endorsements would then just be an emotional reaction. Consider these two quotations:

Reason is the discovery of truth or falshood. Truth or falshood consists in an agreement or disagreement either to the real relations of ideas, or to the real existence and matter of fact. Whatever, therefore, is not susceptible of this agreement or disagreement, is incapable of being true or false, and can never be an object of our reason. Now ‘tis evident our passions, volitions, and actions, are not susceptible of any such agreement or disagreement; being original facts and realities, compleat in themselves, and implying no reference to other passions, volitions, and actions. ‘Tis impossible, therefore, they can be pronounc’d either true or false, and be either contrary or conformable to reason. (Treatise, Part I, Section I)

[I]t is a requisite that there should be some sentiment, which it touches; some internal taste or feeling, or whatever you please to call it, which distinguishes moral good and evil, and which embraces the one and rejects the other.²

Here it appears that Hume is saying that emotions are neither true nor false, and moral endorsements are based on emotions. Therefore, we should conclude that moral facts can’t exist because morality is merely an expression of our emotions. This is the position of non-cognitivists. Hume never made it clear that he was a non-cognitivist, but his moral theory seems to imply that he should be.

I will now explain the various arguments and interpretations involving the is/ought gap.

2. Searle’s Discussion of the Is/Ought Gap

John Searle decided that we can get a nonmoral “ought” from a promise. If I make a promise, then there is a sense that I should do what it takes to fulfill the promise. So, we can get “ought” from “is” because we can get a prescriptive statement (you should do x) from a descriptive fact (a promise to do x).³

This kind of “ought” is not a moral ought. There can be moral considerations that override my reason to fulfill my promise.

Although Searle does not answer Hume’s challenge because Hume wants to know how to get moral “oughts,” Searle still attempts to explain how we can get a kind of prescriptive fact from a descriptive one. If Searle is right, that means that prescriptive facts are not exclusively moral, and prescriptive facts might be somehow connected to descriptive facts.

3. Becker’s Discussion of the Is/Ought Gap

Lawrence Becker agrees with Searle that we can get a nonmoral “ought” from “is,” but he thought we could also get a “moral” ought from nonmoral “oughts” given that there are no overriding reason against doing so. If you should accomplish a goal all-things-considered, then you morally ought to accomplish the goal.⁴

We can get a nonmoral “ought” from a goal. (If you have a goal to eat chocolate, you should buy a chocolate bar.) If you have no overriding reason not to eat it, then (all-things-considered), you morally ought to eat it.

The problem with Becker’s account of moral prescriptive facts is that they aren’t necessarily important. If the goal is important, then the moral “ought” is important; but what’s so important about eating chocolate? Eating chocolate doesn’t sound important enough to be worthy of being a moral “ought.”

4. Intrinsic Values and the Is/Ought Gap

Intrinsic values seem important enough to help us get a moral “ought” from “is.” If intrinsic values are descriptive facts, then we can get prescriptive facts from descriptive facts. All things equal, if human life has intrinsic value, we shouldn’t kill people.

Even intrinsic values can lack importance. The pleasure from eating chocolate might have a small amount of importance, but it still doesn’t sound important enough to be worthy of being called moral. Perhaps we have conventionally required things to be relatively important in order for us to label it is “moral,” but this is just a matter of degree. There is nothing particularly different about eating chocolate for pleasure or reading philosophy for pleasure in the sense that both actions are done out of pleasure. The only difference is the quality and/or quantity of pleasure involved.

5. Ontological Gap

The ontological gap states that “is” and “ought” are different kinds of being (existence). “Goodness” and “wrong” refer to a different kind of property than “hot” or “solid.” One kind of property doesn’t depend on people, but the other does. There are at least three different ways of understanding an ontological gap:

1. Moral facts are not reducible to nonmoral facts (or vice versa). Moral facts and nonmoral facts are two different domains.
2. Moral facts are not reducible to descriptive facts (or vice versa). Moral facts and descriptive facts are two different domains.
3. Morality is not factual at all.

I will consider each of these:

Moral facts are not reducible to nonmoral facts.

Moral realists will agree that moral facts are not reducible to nonmoral facts. That’s the whole point of intrinsic value. Some things are important, but particles and energy are not constitutive of moral facts. Nothing important happens on the level of physics, so physics could be said to only entail nonmoral facts.⁵

There is an objection against realism involving a view that moral facts aren’t reducible to nonmoral facts, but this is a strange argument considering that part of my definition of realism is precisely that moral facts can’t be reduced to nonmoral facts. For example, reductionists who believe that everything reduces to physics might argue that morality is delusional because it can’t be reduced to physics. Everything that doesn’t reduce to physics must be rejected. However, reductionism is not a persuasive reason to reject moral realism because we don’t yet know how to reduce sociology, mathematics, or psychology to physics either. The fact that we can’t reduce these kinds of facts is better evidence that reductionism is false than evidence that they don’t really exist.

The real problem for realism is if we can reduce moral facts to nonmoral facts. Such a reduction would prove that moral facts are dispensable. We could just talk about psychology, for example, instead of morality. (We would also find out that moral facts don’t really matter. Importance would just be a matter of something like desires, and other people’s pain would be of no rational concern to each of us.)

What about Becker? Lawrence Becker’s answer to the is/ought problem seems to imply an answer to the gap between the moral and nonmoral. His answer is that we can get the moral from the nonmoral because he reduces moral facts to all-things-considered judgments involving goal satisfaction. This answer appears to be a challenge to moral realism. If we can get moral judgments from nonmoral judgments, then what good are intrinsic values?

One problem with Becker’s account of moral judgments lacks the importance required for moral judgments. Goals alone are not enough to give us moral judgments because they can lack importance. Although I agree that all things equal, eating chocolate might be good, it is only superficially so. Consider the following:

Becker must admit that those who want to spend hours counting blades of grass could be morally justified doing so because there are no overriding reasons not to. The fact that counting blades of grass is unimportant isn’t in and of itself an overriding reason not to do it. Although most people might have overriding reasons not to spend hours counting blades of grass considering that such a goal will conflict with their other goals, it is logically, metaphysically, and physically possible that a person would have no such conflicting goals. If Becker is correct, this person morally “ought” to spend hours counting blades of grass.

What about Intrinsic values? Notice that I earlier claimed we can get “ought” from “is” using intrinsic values. It might be true that intrinsic values are descriptive facts, but they are moral facts either way. Therefore, I admit that intrinsic values do not let us get moral facts from nonmoral facts.

In conclusion, we can’t get “ought” from “is” In the sense that moral and nonmoral facts are two separate domains. Arguments concerning what “is” the case that somehow give us what “ought” to be the case must have make use of a hidden premise, which would tell us how “is” is relevant to the ethical issue at hand. For example, capital punishment might be wrong if it kills innocent people given the hidden premise that people have intrinsic value. The hidden premise is itself a moral fact.

Moral facts are not reducible to descriptive facts.

Descriptive facts could include moral descriptions (e.g. torture is wrong), but that seems to miss the point. (Descriptive facts might include both moral descriptions and nonmoral descriptions.) The point seems to be that material facts and moral facts seem to be different kinds of things. (Material facts can include any fact within the materialist’s metaphysics: Particles, energy, minds, and anthropological facts can all exist for a materialist.) So, descriptive facts should be taken to be facts of the material world. Facts of the material world might include moral facts, so not all moral realists will agree that prescriptive facts aren’t descriptive. (The is/ought gap could be rejected by arguing that something can be both prescriptive and descriptive: Both a materialistic fact and a moral fact.) Instead, a materialist can agree that psychological and moral facts are caused by particles and energy. (We could agree to materialism as long as all material entities are causally connected.) Materialism itself doesn’t require that we accept that everything is ontologically reducible to physics (particles and energy), so it is possible for a materialist to agree that an irreducible moral domain exists. I discuss why some people agree to a materialistic is/ought gap and what it would mean to deny a materialistic is/ought gap.

If moral facts are materialistic, we can get “ought” from “is” in the sense that materialistic facts can include moral facts.⁶ In that case the premise that “pain is bad” would be a materialistic fact because such an experience of pain is materialistic, and it can give us reason to avoid pain, and it gives us reason to help other people avoid pain. However, not everyone rejects the materialistic is/ought gap. Consider these alternatives to moral facts being materialistic:

• A dualist might argue that moral facts are part of the psychological realm, but psychological facts are quite different than materialistic facts.
• A pluralist or idealist might argue that moral facts are a different domain than the psychological realm and the materialistic realm. For example, someone could argue that moral facts are based on Platonic forms.
• Some might argue that moral facts are supernatural. For example, moral facts might depend on a supernatural deity’s existence. This position is especially mysterious and requires something like divine revelation.

All three of these positions have difficulty in explaining why psychological facts and moral facts are causally linked to (or dependent on) the material world. The dualist, idealist, and pluralist might still have some access to moral facts through introspection, but tying moral facts to the supernatural make it very unclear how we could know moral facts. If the existence of moral facts depends on something we can’t experience, and the supernatural tends to be something that people can’t experience, then we can’t experience moral facts.

I do not wish to argue that we have to be materialists to understand moral facts. It might be that a dualist, pluralist, or idealist can accept the existence of a materialistic world, and it is even possible for them to accept that moral facts are materialistic (or at least tied to psychological facts).

Those who believe that moral facts are materialistic are left with a question: How can we be sure that moral facts are materialistic? This is what I will discuss next.

How could a materialist reject the materialistic is/ought gap? Minds, for example, might be caused by our brain; but minds are not entirely explained and understood in terms of our brains (or in terms of particles and energy). For example, my experience of the color green doesn’t seem like it’s the same thing as neurons firing in my brain. We can describe neurons firing in my brain, but new information is introduced when I describe my experience of the color green. A materialist can then say that minds are not the same thing as brains, but minds are caused by brains. A materialist would then say that minds are part of the physical world, minds are caused by brains, but mental facts are irreducible to nonmental facts. For example, John Searle argues that mental facts are emergent system feature of the brain. If he is right, mental facts require irreducible emergent properties to be caused by the brain.

A materialist might then reject the is/ought gap in the sense that moral facts are also part of the material world. Although moral facts might require irreducible emergent properties, those properties are caused by particles and energy, like everything else.

It can be important for a realist to reject the materialistic is/ought gap because all the relevant facts appear to be materialistic (or psychological). Torturing to others for fun involves physical actions and psychological facts, but torturing others is something we believe to be wrong based on our belief that physical and psychological facts determine moral facts. So, assuming that all relevant facts are materialistic, we must admit that we can know moral facts even if we only know materialistic facts. Moral facts are also materialistic facts.⁷

On the other hand, some philosophers will also reject a materialistic is/ought gap by denying that moral facts require irreducible moral properties. We might find out that mental facts are nothing more than various configurations of particles and energy, and moral facts might be reducible in a similar way. (This would be a form of anti-realism.)

I have not actually argued that we know for sure that moral facts are materialistic. Instead, I simply pointed out the fact that psychology can be taken to be a materialistic fact despite the fact that we don’t experience it as being part of physics. This could be seen as speculative: We can theorize about psychology being materialistic, but it hasn’t been fully justified yet.

However, there is some independent evidence that psychology is materialistic in the sense that it is causally tied to solid objects. The mind of each creature seems to depend on the complexity and configuration of its brain; brain damage can alter someone’s psychology; and our desires and beliefs can influence our body’s movements.

Now we are left with the question: Do we have evidence that moral facts are also dependent on the material world? My answer is, “Yes.” We know moral facts from a combination of psychological and biological facts. The motivations, the ability to cause pain, the ability to damage someone’s biology are all essential facts to determine if an action is beneficial, harmful, justified, right, or wrong. The badness of pain influences our psychology to avoid pain and to help other people avoid pain, but pain is a psychological phenomenon.

In conclusion, just like we have evidence that psychology is dependent on the material world, we also have evidence that moral facts are dependent on the material world. In particular, our experience of moral facts influence our psychology.

The argument for a materialistic is/ought gap: Some materialists also reject moral facts on the ground that such facts imply moral realism and require emergent properties. Such materialists accept that moral facts imply irreducible emergent properties, but they reject that there could be such properties. John Mackie introduced this position and what he called the “argument from queerness.”⁸ The main idea is that we shouldn’t accept queer entities or properties (new kinds of existence) unless we have sufficient reason to do so. He even admits that we often behave as though such irreducible moral facts exist (such as intrinsic values). Such irreducible moral facts might even be required to explain our moral experiences. However, all the worse for our moral experience. We would do better to admit that our moral experiences are delusional than to admit that a new kind of entity exists.

I have already argued that anti-realists that try to make sense out of our moral experiences will fail to do so (because our moral experiences require us to accept altruistic actions as justified, but such actions are not justified for an anti-realist.) Mackie would agree with my argument, but he would reject my belief that the burden of proof is on the anti-realist. Most philosophers will accept that our moral experiences can give us evidence of moral facts. If this is right, then our moral experiences are evidence of moral realism because anti-realists will not be able to make as much sense out of our moral experiences.

However, Mackie would argue against the belief that moral experiences are evidence of realism. Moral experiences merely prove that we are delusional. In order for us to side with Mackie, we will need to accept one of the objections I mentioned in my argument for moral realism. (Such objections were meant to argue that we can’t accept that “pain is bad no matter who experiences it.”) In particular, these two are relevant:

1. Our thoughts and feelings can’t be philosophically analyzed.
2. Pain’s subjective ontology causes it to be less real than required for it to have intrinsic disvalue. Pain is something like an illusion.

I have already discussed why these are not good reasons to reject that “pain is bad for everyone,” and my arguments will also be equally relevant concerning evidence that point to morality being irreducible.

I have discussed why introspective evidence can be a reliable source of knowledge (that we have observation, for example), and now I will argue that introspective evidence can be a reliable source of ontological justification. Introspective evidence is very relevant to ontological knowledge. In particular, we have reason to believe that the mind might not be reducible to the brain in the sense that our experience of green doesn’t appear to be the same thing as neurons firing in a certain way. The fact that an experience of green is multiply realizable in the brain (different brain states can cause a specific experience of green) coupled with our knowledge of experiencing green gives us a strong reason to reject that certain brain states are “exactly the same thing” as our experience of green. It might make more sense to say that brain states can cause our experience of green (than t say that the experience of green is nothing other than brain states).

Given my example, we have pretty strong evidence that our introspective evidence can give us a justification for ontological beliefs. In particular, the fact that an experience of the color green is not “exactly the same thing” as a certain brain state.

If introspective evidence concerning our moral experiences can be used as evidence of moral ontological properties (just like it can give evidence concerning psychological ontological properties), then we also have reason to accept that pain is bad, and to accept that pain is bad for others; and therefore, that pain is intrinsically bad. The argument I gave for moral realism could then be considered to be based on reliable evidence.

One could object here that I haven’t yet given introspective evidence that morality is irreducible. Sure, we can’t understand the experience of the color green using non-psychological facts, but maybe we can understand moral facts using nomoral facts. My reply to this objection is that we can’t understand the badness of pain through a nonmoral description. We can experience the badness of pain, but no amount of nonmoral facts will ever be able to fully describe the experience of the badness of pain.

At this point the anti-realist would need to give us a reason to believe that introspection involving mental ontology and moral ontology are disanalogous. They must be different in some important sense, or the reliability of psychological introspection should indicate the reliability of moral introspection. Introspection involving moral ontology could give us reason to believe that there are irreducible moral facts, just like there appear to be irreducible psychological facts.

In conclusion, we do have reason to believe our moral experiences are reliable just like our psychological experiences are reliable. We therefore have some reason to accept that moral facts are materialistic. If intrinsic values are descriptive (materialistic) facts, then we can get “ought” from “is” in the materialistic sense using intrinsic values. The fact that pain is bad is enough to give someone an aspirin, and that fact might be part of materialistic metaphysics.

Morality is not factual at all.

If morality isn’t factual at all, then there can’t be moral statements. No moral sentence could be true or false. “Hitler is viscous” wouldn’t be true or false, and “charity is good” wouldn’t be true or false. This is the commitment held by non-cognitivists (people who deny that moral sentences can be true or false). Non-cognitivists are anti-realists, so their arguments are relevant. I will treat the arguments against non-cognitivism given by other philosophers to be sufficient. In particular, non-cognitivism is against our moral experience.⁹ Additionally, a non-cognitivist will have to reject that our moral experience is reliable, but I already argued above why we have some reason to believe that our moral experience (introspection) can be reliable. (If our moral experiences are reliable, then we have a good reason to accept moral realism.)

If a non-cognitivist rejects our moral experiences, then it isn’t clear why they don’t side with Mackie, who also rejects our moral experiences. I suppose they want to preserve more of our moral experiences than Mackie, but then they appear to want things both ways: They want to agree with Mackie that intrinsic values are queer and unjustified, but also accept that our moral experiences are worthy of consideration. Then the problem is that our moral experiences will provide us with our evidence for intrinsic values rather than non-cognitivism.

In conclusion, noncognitivism will be rejected and moral facts are possible. We can discuss which moral statements are true or false, even if all moral statements are false.

6. Epistemological Gap

Most philosophers seem to refer to the epistemological is/ought gap and believe that we can’t know prescriptive facts from descriptive facts. Observation, for example, seems to give us evidence of descriptive facts rather than prescriptive facts. I will first discuss the different kinds of epistemological is/ought gaps, and then I will discuss the argument against moral observation. The argument states that we can’t know prescriptive facts because all facts are justified by observation, but we can’t observe prescriptive facts.

Just like the ontological gap, there is more than one way of interpreting the epistemological gap:

1. We can’t know moral facts from nonmoral facts.
2. We can’t know moral facts from materialistic facts.
3. We can’t know moral facts because all facts are descriptive.

The epistemological gaps are tied to the ontological gaps, as I explain below:

We can’t know moral facts from nonmoral facts.

If moral facts are not reducible to nonmoral facts, then we can’t know moral facts given nonmoral facts. Although moral realists will agree with this statement, they point out that a moral theory can be sufficient to derive moral facts from nonmoral facts. Seeing children torture a cat is enough to judge the children as doing something wrong.

Of course, the moral theory might have to be justified on moral grounds rather than nonmoral grounds. If we only know nonmoral facts, then we can never know moral facts. So, how do we ever get to know any moral facts? Because we experience them. I already explained this position in my post, A Moral Realist Perspective.

My position is not one necessarily endorsed by all realist philosophers. Some seem to merely believe that we non-reflectively start off with moral intuitions or moral beliefs, and we are then able to observe moral facts. We could then theorize about which nonmoral facts determine moral facts based on our actual moral observations, intuitions, and assumptions.

In conclusion, we can know moral facts from nonmoral facts given bridging premises, such as “human life has intrinsic value.” This premise would help us conclude that killing people is a bad idea given biological facts involving death and how essential living bodies are for our own existence.

We can’t know moral facts from facts of materialism.

If moral facts are not reducible to materialistic facts, then we can’t know moral facts given facts of materialism. I already mentioned how this gap can be rejected by moral realists. We have some reason to believe that moral facts are facts of materialism. Moral facts might be an irreducible sort of material fact. Once realists reject the materialist is/ought gap, they can escape the objection that all we know are materialistic facts, so we can’t know moral facts. If I am right that moral facts are materialistic, then I am right that we can know moral facts from materialistic facts.

We can’t know moral facts because all facts are descriptive.

One might admit that moral statements are “descriptive” in the sense that they describe part of the material world, but that is not what is meant by the assertion above. What is being asserted is the position that moral sentences are noncognitive (neither true nor false), so we can’t know any nonmoral facts. This is just a trivial conclusion based on noncognitivism. Moral realists will reject noncognitivism, so they will not agree to this epistemological problem.

In conclusion, we will not agree that we can’t know moral facts because we reject that all facts are (by definition) descriptive. Some facts might be prescriptive. It isn’t necessary at this point to admit that there are true moral facts because moral statements might all be false. Noncognitivism can be rejected, even if Mackie is right that all moral statements are false.

The argument against moral observation

Although there are different interpretations of the epistemological is/ought gap, any of them could lead to a single problem: It doesn’t seem possible to know moral facts. This is the conclusion of the argument against moral observation:

The argument against moral observation is the following:

1. We can’t know moral facts from observation.
2. We know everything from observation.
3. Therefore, we can’t know moral facts.

This argument can be based on any of the three epistemological is/ought gap interpretations, so it is relevant no matter which interpretation of the gap we are considering.

The problem with this argument is that neither premise has been proven or sufficiently justified. The argument is mainly just a challenge to realists to explain how we can know about moral facts. I will discuss how each premise can be questioned:

Premise 1: Realists have often accepted premise 2 (that we know everything from observation), but rejected premise 1: I have already given the arguments given by Geoffrey Sayre-McCord, Nicholas L Sturgeon, and Richard N Boyd, who argue that we can observe moral facts. They basically argue that observation is theory-relative, and we need a moral theory to help explain our moral observations. No observation is reducible to our actual experiences. Observations require assumptions and theory in order to make sense. I can see my hand, but that observation requires me to have assumptions about solidity, biology, and selfhood.

It was Geoffrey Sayre-McCord who related the problem of observation to the is/ought gap.¹⁰ He agrees that there might be an is/ought gap; but if there is, then it is no more a problem for morality than it is for psychology. Just like the is/ought gap, there appears to be something like an is/thought gap. (We might not be able to reduce psychology to non-psychological facts, or morality to nonmoral facts.) Observation by itself isn’t sufficient to give us moral facts, and observation by itself isn’t able to give us psychological facts. Of course, we do observe both moral and psychological facts once we realize that certain assumptions or theory is involved. Certain observed behavior and biology indicates certain psychological facts, and certain observed behavior and biology indicates certain moral facts.

Premise 2: I have already discussed the fact that premise 2 seems false, which asserts that we know everything through observation. I argued the opposite: We know about moral facts through personal experience rather than purely from observation. This is equally true about mental phenomena: We directly experience mental phenomena. In that case “we don’t know everything from observation,” so the second premise would be false. In that case we might suspect that it is true that we “don’t know moral facts through observation” just like it might be true that we “don’t know psychological facts through observation.” Instead, we can experience psychological and moral facts, and we can know about them through introspection.

In conclusion, the argument against moral observation is unconvincing because observation is not the only way we know about the world. We also know some things through introspection. (Of course, some people might define observation in a way that includes introspection. In that case the argument is false because we can know about moral facts through personal experience.)

Conclusion

How do you get “ought” from “is?” It depends on what you mean by the is/ought gap:

• If you mean, “How do you get moral facts from nonmoral facts?” then you can’t get “ought” from “is.” You can only get “ought” from “is” by making use of a moral premise. Moral facts can’t be known if we are only given nonmoral facts.
• If you mean, “How do you get moral facts from materialistic facts?” then you can get “oughts” simply from the fact that some materialistic facts are already moral facts. (Although a dualist might argue that moral facts are mental facts rather than material. In that case we can still know about moral facts in the same way through introspection.)
• If you mean, “How do you get moral attitudes considering they aren’t factual?” then we will have to reject the assertion that moral attitudes are noncognitive.
• If you mean, “How do we know moral facts from nonmoral facts?” then we can only know moral facts through introspection or given other moral facts. “Pain is bad” will imply that we shouldn’t torture cats, and that people who are torturing cats are doing something wrong because of the cat’s biology and psychology.
• If you mean, “How do you observe moral facts?” then the answer might be that we don’t. Instead, we can experience moral facts through introspection.

1 Hume, David. A Treatise of Human Nature, Book III, Part I, Section I. 18 Jan. 2010. <http://books.google.com/books?id=5zGpC6mL-MUC&dq=hume+treatise+human+nature&as_brr=1&client=firefox-a>.

2 Hume, David. Enquiry Concerning The Principles of Morals, Appendix I.V. 18. Jan. 2010. <http://books.google.com/books?id=fV0AAAAAMAAJ&dq=Enquiry+Concerning+The+Principles+of+Morals&as_brr=1&client=firefox-a>.

3 Searle, John. “How to Derive ‘Ought’ From ‘Is,'” Philosophical Review 73, 1964, 43-58.

4 Becker, Lawrence. A New Kind of Stoicism. Princeton, NJ: Princeton University Press, 199.

5 To say that moral facts are not reducible to nonmoral facts means that we can’t get moral facts from nonmoral facts of anthropology, psychology, or physics.

6 If moral facts are materialistic, then we can have something a lot like a science of morality. A study of the material world can give us moral facts. Theism and mysticism will be unnecessary for attaining moral truth. If I am wrong, then we might lack a reliable method to learn about moral facts.

7 Of course, a substance dualist might argue that psychological facts aren’t materialistic, but dualism is not currently considered to be a viable option by philosophers.

8 Mackie, J. L. Ethics: Inventing Right and Wrong. 38.

9 I experience myself talking about true and false moral statements, such as “Pain is bad.” Non-cognitivists seem to deny that I can do this. Instead, they want to say that I am just expressing my emotions.

10 Sayre-McCord, Geoffrey. “Moral Theory and Explanatory Impotence.” Midwest Studies in Philosophy, Vol. XII (University of Minnesota Press, 1988), pp. 433-457. (Chapel Hill Philosophy. 18 Jan 2010. <http://philosophy.unc.edu/people/faculty/geoffrey-sayre-mccord/on-line-papers/Explanatory_Impotence.pdf/view&gt;.)