Monday, 12 May 2008

the fallacy fallacy

One of the grand worries in the business of evolutionary explanations is that they sometimes (especially when put into colloquial terms) seem to confuse is with ought. This is known formally as the naturalistic fallacy (or its converse, the moralistic fallacy, depending on which way the conflation runs). Whenever a report on scientific findings uses the word "should," I worry.

Azra Raza of 3QD quotes a piece in Scientific American:
In research that could give doctors a way to reassign sex in cases of unclear gender, scientists report this week that they have figured out why some children with genes that should make them boys are instead born as girls.
A more precise statement might go:
In research that could give doctors a way to assign sex in cases of unclear gender, scientists report this week that they have figured out why some children with genes expected to make them boys are instead born as girls.
The phrase "think deviant" comes from the notion that a goal of science is to explain why it is that things deviate from what we might otherwise expect. And so I have no problem with journalistic characterization of results as unexpected or revolutionary--science even for workaday Kuhnian normal scientists can be incredibly exciting. No new paradigm is suggested by results like this one: the Sox9 gene, which is expected to activate, sometimes doesn't. Manipulating this process can change sexual expression.
[Lovell-Badge] adds that he's very hopeful that with further analysis, scientists may determine ways to reassign gender later in life, "perhaps for cases of sex reversal or perhaps even for individuals who want to undergo sex changes," although he acknowledged that "this is getting very contentious."
And that's precisely the point: it's an exciting find and raises fascinating ethical issues by itself--issues that shouldn't be confused at the start by careless use of words like "should" when "expected" will do.


Anonymous said...

My worry with all this is the idea that someone has to change their perfectly healthy physiology (via radical surgery and drugs with massive side effects) to match society (or even to some extent their own) expectations about sexual physiology.

You may remember my newspaper clipping where someone said (apparently in earnest, but possibly ironically) something like "Transexuality is not a [psychological] disease that needs to be cured, except by drugs and surgery."

Anyway medicine has the inherently normative (but perhaps not strictly ethical) idea of "health." As opposed to biology, where you are right to bring up the naturalistic fallacy or rather Hume's Law about not deriving ought from is.

Actually as defined in the wiki article (and most places) I'm not clear that the strict naturalistic fallacy is a fallacy. If someone defines the good in terms of some natural property then she can reasonably draw ethical conclusions from natural facts in combination with that definition. However, that definition would be more than just a formal definition and actually an ethical principle. For example Searle claims (and he was not the first to do so, although nobody points this out) you can derive ought from is in the case of promise keeping and you can if you define (or rather introduce the principal of) promise keeping as a good (you would also need some extra principles to deal with when you made too contradicting promises and so on). So I think he defeats the naturalistic fallacy but not the is-ought problem. He also has to argue for the ethical principle he introduces and he never does that because he claims his definition is not actually an ethical principle (I find that galling myself).

Isaac said...

To avoid confusion, I didn't want to get into those issues in the main post. But now that we're outside of the main post...

I agree that medicine is inherently normative with respect to health. Medicine is both an advisory profession ("you should eat more vegetables") and an engineering discipline ("the function of the pancreas is such-and-thus, and so it should produce insulin").

I am not so sure doctors are well-suited to be the ones deciding whether there should be a pancreas to begin with.

Whatever we call the fallacy, I prefer 'expectation' over 'prescription' whenever possible. Both are theory-laden, but expectation seems more explicit about it, which--to me at least--makes it less worrying.