Desires vs. values: A new experiment

Ever since  Watson's classic paper, philosophers have been concerned with the distinction between desires and values. The usual view is that this distinction is a purely psychological one -- the sort of thing that might figure in a scientific theory of the human mind.

Erica Roedder recently suggested to me that this view might be leaving out an important aspect of our ordinary folk concept of values. She pointed out that certain moral norms might be playing an irreducible role in the concept.

Together, we designed an experiment to test this hypothesis:

Subjects in one condition were given a story about a character loosely modeled on Huck Finn. On a conscious level, he believes that the racist practises of his society are morally right, but he sometimes feels a pull in the opposite direction. He finds himself feeling guilty when he performs racist behaviors and sometimes ends up doing things that promote racial equality as a result. Subjects in this condition were then asked whether, despite his conscious beliefs, the character actually values racial eqaulity.

Subjects in the other condition were given a story about a kind of 'inverted Huck Finn.' On a conscious level, he believes that he ought to treat all races equally, but he sometimes feels a pull in the opposite direction. He finds himself feeling guilty when he performs behaviors that advance other races at the expense of his own, and he sometimes ends up doing things that promote racial discrimination as a result. These subjects were then asked whether, despite his conscious beliefs, the character actually values racial discrimination.

The results showed a marked asymmetry. People who had been given the first story tended to say that the character's attitude toward racial equality was one of his values, whereas people who had been given the second story tended to say that the character's attitutde toward racial discrimination was not one of his values. This result is puzzling, since the two stories seem exactly parallel on a psychological level, differing only in the moral worth of the attitudes themselves.

Any interpretations?

Comments

[Kip Werking]

Perhaps this lends some support for asymmetry theories such as Susan Wolf's?

Also: maybe the selection of race discrimination prejudiced the subjects. Regardless of how the fact-pattern is setup, it might be more difficult for people to say that anyone "values" racial discrimination, in the same way that a person "values" racial equality. I also tend to think of racists as people who feel comfortable with their prejudice, and feel uncomfortable with the prospects of racial equality, and not as people who positively value racial discrimination (although such people certainly exist).

One way to filter out this (speculative) bias would be to consider unusual people who have conflicted values/desires for things ordinary people feel indifferent about. Suppose that a person strongly values eating apples over oranges, but sometimes feels a pull towards oranges. That is just an example off the top of my head, one should be able to contrive less silly ones.

The results are, nevertheless, very interesting.

[jon kvanvig]

I wonder if you could run the experiment and leave out evaluative terms like equality and discrimination. So, in place of valuing racial discrimination, you could put it in terms of valuing whites over blacks, for example; and in place of valuing equality, you could put the case in terms of not valuing either race over the other. I'd bet you'd get something closer to symmetry with such questions, but I've been wrong before...

[Jussi Suikkanen]

Hi Joshua,

here is one suggestion (which perhaps is even empirically testable is one believes in such things...). What if the respondants actually used something like Davidsonian principle of charity in their interpretation of the actions of the persons in the described cases? This would mean that as the respondants themselves value racial equality they project this value to explain the actions of others. Here the fact of whether this value produces actions on conscious or unconscious level would not be seen as relevant. Value of discrimination would thus be a value which the respondants would feel uncomfortable in ascribing to the persons in the stories because this would make them in part to be irrational - having false values would be seen here as a kin to having false beliefs. The discriminatory actions would then be explained more likely with other psychological mechanisms than 'acting from values'.

Just a suggestion. What one could test this with would be to look at the respondants' reactions to the question of whether the persons in the described cases have the explicit values they say the have, and to run the test with persons with 'discriminatory values'.

[Jason D'Cruz]

I think that the results say more about the ordinary connotation of the word "values" than any substantive philosophical thesis. Generally, when non-philosophers use the word "values" they mean "values that are in accord with morality" (as in "family values", "traditional values", "American values" etc). In the first case the character's attitude was indicative of his "values" because his deepest belief (non-racism) is in accord with morality; in the second case the character's attitude was not indicative of his "values" because his deepest belief (racism) is immoral.

[Rob]

Joshua,

Could you divulge the text of the stories your subjects were given?

[Joshua Knobe]

Thanks for all the comments! The suggestions so far have been really helpful.

Jon and Kip, These are definitely important alternative hypotheses, and the only way to rule them out would be to do some additonal experiments. Do you think it might be helpful to try an experiment in which we just use one vignette and then give it out to people with different moral views? The hypothesis would be that people would attribute different 'values' to the agent depending on what they themselves regarded as morally good.

Jussi, That looks like a promising explanation for the effect. But would you expect that people would also be reluctant to attribute desires for outcomes that they do not regard as truly desirable?

Rob, I'm sorry that I didn't post the vignettes before. The first one goes as follows:

--
George lives in a culture in which most people are extremely racist. He thinks that the basic viewpoint of people in this culture is more or less correct. That is, he believes that he ought to be advancing the interests of people of his own race at the expense of people of other races.

Nonetheless, George sometimes feels a certain pull in the opposite direction. He often finds himself feeling guilty when he harms people of other races. And sometimes he ends up acting on these feelings and doing things that end up fostering racial equality.

George wishes he could change this aspect of himself. He wishes that he could stop feeling the pull of racial equality and just act to advance the interests of his own race.
--

And the second one goes like this:

--
George lives in a culture in which most people believe in racial equality. He thinks that the basic viewpoint of people in this culture is more or less correct. That is, he believes that he ought to be advancing the interests of all people equally, regardless of their race.

Nonetheless, George sometimes feels a certain pull in the opposite direction. He often finds himself feeling guilty when he helps people of other races at the expense of his own. And sometimes he ends up acting on these feelings and doing things that end up fostering racial discrimination.

George wishes he could change this aspect of himself. He wishes that he could stop feeling the pull of racial discrimination and just act to advance the interests of all people equally, regardless of their race.
--

[Jussi Suikkanen]

Joshua,

no, I would not expect them to be reluctant to do that. As far as I see this, desires and beliefs about desirability are states of mind on quite a different levels. I take it that beliefs about desirability of things and values are pretty much the very same thing. Desires, on the other hand, are the direct pro-attitudes towards the outcomes of the actions that make us act. They are, in the causal chain of things, inbetween the values and the actions. Only naive emotivists would say that these two states of minds are basically the one and the same state of mind. This of course doesn't leave room for weakness of will, which some take to be an absurd implication of naive emotivism.

So, in the case of someone doing something I think is highly undesirable, it seems obvious and non-problematic to attribute a desire for the person to do that thing she is doing. That seems to be constitutive for the event to be an action in the first place. The question, of course, is about how to explain this desire. The suggestion I made earlier was that in this situation we might first attempt to attribute the same 'true' beliefs about desirability and values to the agent we have and try to explain the action as some kind of a local failure in rational agency. This would take the agent to be more rational on the whole. The desire to do the undesirable would rather be seen as the agent's failure to do what she really thinks is desirable or what she values. The desire would be explained by some other, deviant causal mechanism.

Of course, when such attempts for explanations fail, we have no option but to attribute beliefs about desirability to the agent we think are false to explain her desires. We might still try to be charitable and attribute as little false beliefs about desirability as we can in order to be able to understand her actions. But attributing a set of beliefs about desirability we do not think make sense from the beginning does little for us in our attempt to understand why the agent does what she does.

Never thought of this before myself. It starts to make some sense though. Maybe there is something to this after all...

[Rob]

"He often finds himself feeling guilty when he helps people of other races AT THE EXPENSE OF HIS OWN. And sometimes he ends up acting on these feelings and doing things that end up fostering racial discrimination." (my emphasis)

Joshua,

I wonder whether the asymmetrical results aren't owing to the zero-sum-like suggestiveness of the above portion of the second vignette. Perhaps the results would be more symmetrical if the corresponding sentence of the first vignette read: "He often finds himself feeling guilty when he harms people of other races TO THE BENEFIT OF HIS OWN RACE" (or some such way of paralleling the second vignette on this score); or, perhaps, simply remove "at the expense of own" from the second vignette. In other words, the first vignette doesn't invoke the idea that the promotion of equality will cost those (like George) who hold the basic racist viewpoint in the culture at issue, whereas the second vignette does invoke the idea that there is a cost involved in the promotion of equality; and perhaps your subjects' response to the second vignette reflects their sympathetic sensitivity to this matter of cost.

Just a wild speculation.

[Keith]

Well, speaking as a novelist, not a philosopher...

The two passages aren't symmetrical. "Extremely racist" and "believe in racial equality" are not equivalent in their two contexts. People usually respond more to connotation than denotation, and "racist" has an emotional charge and connotation that "believes in racial equality" can't begin to approach.

My suggestion would be to delete the first two lines of each passage. Start with an emotionally neutral statement of what George believes, e.g., "George believes that he ought to be advancing the interests of all people equally, regardless of their race."

In my experience, denotation and strict definition are about the last things many (most?) people perceive when encountering certain words. "Racist" is certainly one of those.

[Richard Double]

I agree with Kip Werking and Jason D'Cruz. I add a wrinkle. Whether we have read Twain’s novel or not, most of us have been exposed to the general type of Huck Finn persona/script through TV, movies, other reading, etc. In that story, the hero growing up in a racist environment is really [deeply] anti-racist, that is, VALUES non-racism. On the other side, the reverse-Huck story strikes me as not as common. Subjects, thus, instead of giving their answers by reading them off the script have to decide on other grounds. Now, calling racism “a value” is difficult for most subjects, because they fear that it might sound as if the subjects THEMSELVES were expressing tolerance for racism. One does not need to be a wild-eyed emotivist to believe that calling things “values” is more positively emotive than calling them “desires.” I suspect that subjects’ answers are often driven by their aim to say what they think the experimenters want to hear, although I don’t know of work that speaks to my suspicion.

[Bertram Malle]

I want to support two themes featured in several posts:

(1) In our research on values (http://hebb.uoregon.edu/04-01tech.pdf) we found pretty strong evidence that people _conceptualize_ values as being directed at desirable things, primarily prosocial content. I wouldn't rule it out that people might ascribe a negative value to despicable characters (or, more likely, groups), but we haven't found any. All the negative stuff seems to be the content of goals (e.g., exploitation, fame, power, etc.)

(2) Wishful thinking (projecting): The participants' own (probably egalitarian) values make it more likely that they like the first character better (he is more like them) and gladly infer that he has the desirable value. The second character is both a little strange and may make participants uncomfortable; they would rather not make any strong inferences, especially not ones that make him look awful.
A classic reference: Ross, L., Greene, D., & House, P. (1977). The "false-consensus effect:" An egocentric bias in social perception and attribution processes. Journal of Experimental Social Psychology, 13, 279-301.
A nice critique: Dawes, R. M. (1990). The potential nonfalsity of the false consensus effect. In R. M. Hogarth (Ed.), Insights in decision making: A tribute to Hillel J. Einhorn (pp. 179-199). Chicago: University of Chicago Press.

[Joshua Knobe]

Thanks again for all of the comments! They are really helping us see the important issues here.

We had originally been thinking that the concept of values was a prototype concept. That is, we thought the concept had a number of different features, such that none of them were strictly necessary but all of them had a certain weight. The basic idea was that the actual moral significance of the attitude's content was one of the features people used to determine whether or not a given attitude was a value. It would not be a strictly necessary feature, but it would have at least some weight in the process of classification.

A number of people seem to be suggesting that the moral significance of the content is not really an aspect of the fundamental criteria underlying the concept of values. The suggestion is that people are simply using the moral significance as a way of getting at some other feature that is truly the criterial one. (This idea comes through especially clearly in Bertram's very interesting suggestion that the results might be due to the false consensus effect.)

But now a new question arises. If the moral significance is not actually part of the criteria, then what sort of feature in the actual criteria might it be affecting?