Ethical Realism

July 14, 2010

A Review of Moral Realism by Torbjörn Tännsjö

One of the defenses for moral realism that makes use of common sense was given by Torbjörn Tännsjö in his book Moral Realism published in 1979 and revised in 1990. The reason that his argument makes use of common sense is because it demystifies the strangeness of morality by opening us up to the fact that moral observation is possible. Tännsjö technically mainly only defends moral realism because he argues that there is no good reason to reject moral realism of the sort he defends. If his defense succeeds and we can fully justify his beliefs, then the following argument for moral realism is implied:

  1. If we have moral knowledge, then moral realism is true.
  2. We have moral knowledge.
  3. Therefore, moral realism is true.

The argument can be rephrased as: We observe that at least some experiences of pleasure and pain have intrinsic value just by experiencing them, and “moral realism” is the hypothesis that there are moral facts beyond our mere beliefs or desires. (If something is intrinsically good, then it is intrinsically good even if we don’t desire it and even if we don’t believe it to be good.)

I agree with Tännsjö that we observe intrinsic values, but I think his argument is slightly lacking. One, he argues that we experience pleasure as good and pain as bad, but it isn’t clear from his argument that we observe pleasure or pain as intrinsically good or bad. Two, Tännsjö seems to think that only pleasure and pain have intrinsic value, but it is a fairly common assumption that human and animal existence also has intrinsic value. I will discuss his defense of moral observation and my objections to his argument.

Tännsjö’s Argument

Premise 1 is trivial. If we observe intrinsic values, then we have moral knowledge of a moral realist sense. Therefore, if we have moral knowledge, then moral realism is true. Premise 2, however, is the main premise that Tännsjö must defend. There are many moral theories of justification, such as reductionism, intuitionism, pragmatism, coherentism, and inductivism & fallibilism. First, Tännsjö attempts to show that these sources of justification are all inadequate insofar as they could attempt to justify moral beliefs without observation, and he argues that observation is the best way to justify moral beliefs. I will not discuss these alternative sources of justification and I will only concentrate on Tännsjö’s proposal that observation is the best way to justify our moral beliefs.

Second, Tännsjö discusses why he believes moral observation is the best way to justify moral beliefs and various objections to such a proposal. I will discuss (a) what Tännsjö thinks “moral observation” consists in using only common sense everyday terminology and (b) how moral observation can meet the appropriate criteria to constitute moral knowledge. He discusses the following criteria that many believe to be required to constitute knowledge:

  1. It must give us a peculiar (causal) contact with realities (55).
  2. It must be indefeasible (56).
  3. Observation must be its own best explanation (57).
  4. It must be intersubjective (56).
  5. It must be incorrigible (57).

Tännsjö argues that the first three criteria are necessary for observation to provide us with knowledge and that moral observation can meet these demands. Moral observation can provide us with peculiar contact with realities, be indefeasible, and be its own best explanation. The final two criteria (intersubjectivity and incorrigibility) are not necessary for observation to provide us with knowledge.

What is “moral observation?”

Tännsjö defines observation as the following: “To ‘observe’ is generally held to be to make an immediate statement of some sort in response to a concrete situation without any conscious reasoning having taken place” (54). Additionally,

The requirements that observational statements be of a certain form, that they describe concretely existing entities, and that they be consistently assented to whatever our sensory receptors are stimulated in certain ways all seem to be a necessary requirement of observational statements.. They are not sufficient (in the present context), however. I have taken observations to be capable of confirming or confuting theories. (55)

Observation is immediate and without thought. When I observe my hand in front of my face I know immediately that I have a hand without having to think about it just because of the situation I am in. I don’t have to think of an argument or justification to prove my observation is reliable. In this way Tännsjö argues that we can make moral observations. Observations can confirm or disconfirm theories, hypotheses, and beliefs. My assumption that I have a right hand is confirmed if it is right in front of my face and it is disconfirmed if my hand is shown to be missing.

Intrinsic values can’t be observed in the sense of seeing a hand because intrinsic values are properties (of mental states) rather than solid objects. Pleasure has the property of being intrinsically good and pain has the property of being intrinsically bad. Tännsjö believes that we observe intrinsic values in a similar way that we observe heat (58). To feel intense pleasure and observe, “this pleasure is intrinsically good” is much like to feel a flame and observe, “this flame is hot.”

An objection: Tännsjö considers the objection that the flame really does have a property of hotness, but pleasure does not have a property of being intrinsically good because such a property can’t exist. He argues that to say that any property exists is a strange thing to say, but there are at least two ways in which intrinsic values can exist.

One, “if we allow ourselves to talk of possible objects, it is sufficient that there is some possible object that is good in order for there to exist a property of goodness. In this sense there certainly exist moral properties” (ibid).

Two, to say that there are no moral properties could imply that they can be reduced to something else, but we have no reason to think that is true at this time. For example, what we think of as “intrinsically good” could be identical with pleasure just like hotness might be identical with high kinetic energy (59). However, the fact that high kinetic energy always coincides with hotness does not show that they are identical any more than the possibility that pleasure always coincides with intrinsic goodness would show that pleasure is identical with intrinsic goodness. “These reductions, if they are possible, don’t show, then, that there is no property of hotness. On the contrary, they presuppose that there is one. In the same vein, a possible intertheoretical reduction of moral laws to psychological ones would not show that there are no moral properties. On the contrary, it would presuppose that there are such properties” (60). Tännsjö is not sure how we could reduce a property to show it to not exist. It is true that H2O is water, but that is merely the discovery that one object is nothing more than the sum of its parts. Water behaves in exactly the same way we would expect H2O to behave.

Does moral observation provide us with moral knowledge?

If moral observation can meet all the requirements for providing us with knowledge, then we have good reason to think it can provide us with knowledge. He considers five reasons that moral observation might be rejected:

  1. If it fails to give us a peculiar contact with realities.
  2. If it fails to provide indefeasible knowledge.
  3. If it fails to be its own best explanation.
  4. If it fails to be intersubjective.
  5. If it fails to provide us with incorrigible knowledge.

It must give us a peculiar contact with realities.

If seeing my hand in front of my face has nothing to do with there really being a hand in front of my face, then it would be strange to say that I am really observing my hand. If we have moral observations despite there being no moral properties, then we have a good reason to say that moral observation is impossible. Observation is a satisfying source for knowledge precisely because it is supposed to give us access to part of reality. If we have no access to moral reality, then it would be strange to say that we could have moral knowledge. Of course, it is possible to be tricked into thinking we have observations. I might think I see my hand while looking at someone else’s hand, but I could discover that I didn’t observe my hand after all. There are therefore legitimate and illegitimate sorts of observation depending on whether or not an appropriate contact with reality has been made.

Objection 1: From a the perspective of eliminative materialism there is nothing in existence except atoms. From that perspective, mental and moral properties (and entities) will be equally objectionable (67-68). However, we don’t currently know how to explain away our mental experiences, so eliminative materialism is far from established as true. It is quite plausible that psychological facts are real and such facts could even be physical. In the same way moral facts seem to be real and could be physical.

Objection 2: It has been argued that we would need a special sense perception to observe moral facts, but this seems just as strange as to think we need a special sort of sense perception to observe psychological facts. “[I]ntrospection can… be conceived of without assumptions being made about any faculties besides our ordinary reasoning and perceiving” (68). Simply put, we know about our psychological facts and moral facts directly from first hand experience. I’ve seen green, so I know what green looks like to me. I’ve felt pain, so I know what pain feels like to me.

Objection 3: J.L. Mackie argued that moral properties must be strange in the sense of being intrinsically motivating (69). However, Tännsjö disagrees. “Perhaps it is true that most people do as a matter of fact to feel somewhat inclined to produce a certain event, if it is perceived by them to be good… But I find it obvious that there can be people who pay little attention or none at all to the value of the consequences of their actions. The notion of value as I see it does not ‘logically’ (whatever that could mean in this context) provide us with motives” (69-70).

It must be indefeasible.

If an observation is indefeasible, then it is not “defeated” by possible objections. If an observation is not caused by the appropriate situation, then it is defeated. For example, I might believe that a cow is in the distance when it is really a cardboard cutout of a cow. This means that I didn’t really observe the cow and my observation was defeasible. The question is, can moral observations be indefeasible (undefeated and caused by the appropriate moral properties)?

Why would anyone think that moral observations are always defeasible?

One, it could be argued that moral observations can’t be shown to be defeated, and an observation can’t be indefeasible if there is no possibility of being proven to be defeasible. However, Tännsjö argues that moral observations are defeasible. If I think I see an elk but I find out it was a deer, then my observation has been defeated, and in a similar fashion I might think someone’s behavior indicates that she is feeling something intrinsically bad when I later find out she was actually feeling ecstasy (72).

Two, we are almost always biased about our moral beliefs (74). For example, the person who is profiting by harming others will want to rationalize their behavior so that they can continue profiting from it. Tännsjö admits that moral observations might be rarely impartial and indefeasible, but that doesn’t mean they always are. If it is in someone’s advantage to be moral, then they would probably hold a true moral belief. For example, suppose that torture is always wrong. Someone about to be tortured would be likely to correctly think that the act of torture about to be undertaken is wrong. This belief would not be undefeated by an impartial perspective.

Observation must be its own best explanation.

According to Gilbert Harman, an observation must be its own best explanation in order to confirm or disconfirm a theory, and Tännsjö agrees that this is necessary for observation to give us knowledge (78). Harman also argues that moral observations have no connection to moral reality and we would have the same moral observations, even if they are false. For example, we observe that children torturing a cat is doing something wrong, but the physical actions in no way confirm or disconfirm the hypothesis “that torturing cats for fun is wrong.” The moral observation reflects nothing other than our moral assumptions.

On the other hand, Harman argues that scientists can observe a proton from a scientific experiment, and they would not make that observation if there was no proton. The observation of a proton best explains itself—a proton really exists and the theory of protons is true. Consider a more familiar example. Observing people grabbing apples with their hands best explains itself—there really are people grabbing apples with hands, and my assumptions about the solidity of apples and hands is confirmed.

Tännsjö objects to Harman’s argument in two ways:

One, he agrees that we can’t observe normative facts (that an act is morally irrational or “wrong”) (80). The wrongness of an action depends on the consequences of an action compared to alternative actions, but we can’t observe all of the good and bad consequences of an action compared to alternative actions that could be taken instead. We don’t observe moral relations of this kind any more than we can observe that someone jumps higher than anyone has or will ever jump (ibid). Instead, we observe intrinsic values. My observation that my intense pleasure has intrinsic value confirms rather than disconfirms my assumption that all pleasure has intrinsic value. (An observation that some pleasure lacks intrinsic value would disconfirm that assumption.)

Two, it isn’t clear what it means for an observation to best explain itself. Harman wants to say that the best explanation of an observation of a proton existing is that it actually exists, but there are alternative explanations that are also very plausible:

  1. There was a proton there, making very obvious signs of its presence…
  2. The sensory receptors of the physicist were triggered in a way typical of his observing a proton.
  3. The physicist was confronted with the kind of signs… that make physicists normally say there really was a proton there. (82)

Although Tännsjö agrees with Harman that the observation best explained itself, he says it is far from clear why that is so. If moral observations don’t best explain themselves, we need to know in what important respect they differ from other sorts of legitimate observations.

It must be intersubjective.

Tännsjö admits that our beliefs could be justified to a greater extent when other people share our observations. (Sharing observations make them “intersubjective” rather than “subjective.”) It is not clear that other people can observe the intrinsic value found in our pleasure, so it might not be possible for moral observations to be shared (75). Nonetheless, Tännsjö does not believe that intersubjectivity is a necessary condition of knowledge. I can know that I am experiencing a hallucination even if no one else can observe my hallucination, and I can observe the intrinsic value of my pleasure even if no one else can observes it.

It must be incorrigible.

Some people suggest that observations can only provide us with knowledge if they are incorrigible—completely justified (57). The problem is that we prefer falsifiable evidence. We want our observations to be disprovable by other observations. It might be that my observation of my own existence is incorrigible, but most observations can be disconfirmed (ibid). If I see my hand in front of my face, but then I look back and notice that I was actually looking at someone else’s hand, then my original observation is disconfirmed.

My Objections

Pain is bad, but is it intrinsically bad?

Tännsjö doesn’t sufficiently differentiate intrinsic values from other sorts of value, and it isn’t clear that we observe intrinsic value rather than other sorts of value. Sure, we experience our own pain as being bad, but Tännsjö doesn’t make it entirely clear what makes pain intrinsically bad or what intrinsic value means. I take it that he thinks some things can be good or bad irrespective of beliefs and desires.

Still, I agree with him that intrinsic values can be observed and I have differentiated our experience of intrinsic values from other sorts of values in my discussion, An Argument For Moral Realism. For example, Tännsjö doesn’t fully explain why intrinsic value seems to be irreducible to nonmoral facts. There are good reasons for having this belief beyond the mere fact that we can’t currently do it.

Can animal existence have intrinsic value?

Tännsjö argues that “concrete experiences” can be observed to have intrinsic value in various situations (54). We don’t observe that animals have intrinsic value from concrete experiences given various situations. However, we might experience our own existence as having intrinsic value by knowing what it is like to exist. We might be able to contrast the value of our existence based on various situations involving our happiness, intelligence, and state of mind. I suggest that even when depressed or in pain we still believe that life is worth living which suggests that our existence still has intrinsic value. Our pleasure and happiness are not the only things we experience as having intrinsic value. Our own existence is also experienced as having value. Pain is not generally bad enough to invalidate the value of our existence.

What about others an animals? We an generalize that others also experience their existence as having intrinsic value. Other people respond in similar ways as we do, and other animals seem to have some characteristics that seem to give our existence intrinsic value as well.


Tännsjö’s defense of moral realism is very similar to my argument for moral realism. I agree that we can observe intrinsic values and we can show that our experience of pain is not only bad, but it’s intrinsically bad.

Tännsjö also believes that moral reason and normativity in general is based on human interests rather than an independent reality from human interests. We should say that an action is wrong if it leads to much more intrinsic disvalue than the alternatives, and we should say that an action is morally irrational if we expect that it is wrong based on our current information (95-98). This is a sharp contrast to Terance Cuneo‘s view that moral realism is irreducibly normative. Tännsjö would probably argue that moral reasons are categorical insofar as intrinsic values are the only relevant considerations to moral assessment, but Cuneo might find a way to argue that moral reasoning means more than that. For example, it might be wrong to violate human rights even if we expect our action to promote intrinsic values better than all the alternatives.

You can buy Torbjörn Tännsjö’s Moral Realism at



  1. […] James Gray critically reviews Torbjörn Tännsjö’s book Moral Realism at some length, summing up his response to the book briefly by writing: I agree with Tännsjö that we observe intrinsic values, but I think his argument is slightly lacking. One, he argues that we experience pleasure as good pain as bad, but it isn’t clear from his argument that we observe pleasure or pain as intrinsically good or bad. Two, Tännsjö seems to think that only pleasure and pain have intrinsic value, but it is a fairly common and uncontroversial assumption that human and animal existence also has intrinsic value. […]

    Pingback by A Philosophers’ Blog Carnival! – Camels With Hammers — August 31, 2010 @ 5:17 am | 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 )

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s

Blog at

%d bloggers like this: