intuitions & curve-fitting

From a very interesting post at Overcoming Bias:

In the ordinary practice of fitting a curve to a set of data points, the more noise one expects in the data, the simpler a curve one fits to that data.  Similarly, when fitting moral principles to the data of our moral intuitions, the more noise we expect in those intuitions, the simpler a set of principles we should use to fit those intuitions....

The fact that our moral intuitions depend greatly on how situations are framed, differ greatly across individuals within a culture, and vary greatly across cultures, suggests lots of noise in our moral intuitions.  The fact that moral philosophers don’t much trust the intuitions of non-moral-philosophers shows they agree error rates are high.  So I wonder: what moral beliefs should we hold in the limit of expecting very large errors in our moral intuitions?

The author, Robin Hanson, uses this idea to make a pretty radical argument in moral philosophy, but the question posed there is of much more general interest and relevance than that one argument, I think.

Comments

[jonathan weinberg]

(Sorry for the big font -- for some reason the typepad editor was not agreeing with me tonight.)

[Robin Hanson]

I've tried to clarify that I wasn't trying to make a radical claim as some have interpreted.

[Robin Hanson]

See http://www.overcomingbias.com/2009/05/reply-to-caplan.html

[Josh May]

Hanson provides an interesting perspective on how to approach moral theory. But I worry about thinking of moral theory as just fitting moral principles to all our moral intuitions.

Keeping with the graphing data analogy, most moral philosophers don't even plot the points of the erroneous intuitions. They plot the clear ones (e.g. kicking babies for fun is wrong) and look for principles that will help with the tough ones (e.g. abortion). And when most ordinary people seem to be confused about what should be clear cases (I think homosexuality is an example here), moral philosophers plot the correct, opposite intuition (e.g. that homosexuality is not immoral).

So I worry about the idea of plotting the noise at all. I don't think it's how moral theory tends to or should go. And if the noise isn't there in the theorizing, we won't get off the ground Hanson's idea about (extremely) simple principles. (I think there are other worries about Hanson's post even if we accept this methodology, but I put those points over the comments of Hanson's original post.)

[Robin Hanson]

Josh, you seem to think philosophers have access to "clear" intuitions they are sure have no error, even when many other people disagree. But the fact that many others disagree with your intuitions means you should be at least uncertain about them.

[Josh May]

Robin,
Thank you for the comment. In response:

It depends on what you mean by "sure" and "certain." In one sense, "certain" or "sure" may mean there's no chance that we could be wrong or that we're not 100% positive. In that sense, I agree; we probably can't be (100%) "certain" that a moral intuition is accurate. But in another sense (the more ordinary one, I think), "certain" or "sure" just means that we have very good reason to think the judgment is accurate. In that sense, I disagree. Surely we have very good reason to think some moral intuitions are correct. Even if there's a chance we're wrong, I don't think that means we shouldn't be confident in them or that we don't have a reasonable basis for distinguishing them from other intuitions that are not "clear"---i.e. not clearly correct.

You're right that there may be reason to doubt the accuracy of a judgment about a case if certain people disagree. But we shouldn't doubt it just because some people make a different judgment, no matter who those people are. You seem to think it's the sheer number of differing intuitions that matters. But that can't be right. Even if the majority of people intuit that enslaving a non-white person is okay, we in the minority needn't lack confidence in the accuracy of our intuition that it's not, so long as we have reason to think those in the majority are biased, unreasonable, misinterpreting the facts, or going wrong somewhere in some way. So the numbers don't really matter. I think it's more like only when smart, thoughtful, rational, unbiased people have different intuitions should it call for pause.