Double Effect & the Knobe Effect

Here's a new paper, arguing that the intuitions to which the doctrine of double effect appeals are normatively loaded. As I see it, this is a problem for proponents of the doctrine, because the normative element should be an output of application of the doctrine, not an input into it. If I am right, it is only because people judge that a certain action is permissible, say, that they judge it to be brought about unintentionally, rather than the other way round.

I would like to build on these results in the following way. My view commits me to the following prediction: people’s prior normative views will influence whether they see actions as intended or not, or as goals or side-effects. This is testable: a prior screening stage could sort subjects into those who, for instance, are pro-life or pro-choice, and then we could see whether these prior beliefs influence their judgements regarding intentions or goals given appropriate vignettes. I don’t have the facilities to run this test (if you read the paper, this will be very obvious: the data here is somewhat flaky). Anyone who wants to volunteer to collaborate with me on getting better data please let me know! (I should add that I can fund a study; it’s access to subjects that is the problem for me).

Comments welcome, of course.

Download Double Effect and the Knobe Effect (pdf).

Comments

[Jonathan]

Prof. Levy,

You note in the document that your "claim is that the distinctions are too normatively loaded to serve as a criterion to sort difficult and controversial cases."

However, is it not that if what you claim is true, then any attempt at the use of DDE is simply rationalization of something else?

If so, then it seems to me that you should begin with an attack on the least subtle of cases (those of ectopic pregnancy, for instance), for if your argument cannot shake those lose, then more difficult cases might just be instances of careless arguments, etc.

--Jonathan

[Jonathan]

Make that "loose."

-J.

[Neil]

Hi Jonathan,

Suppose that normative judgments are an input into the functionally modular subsystem whereby we judge whether an action was intended or merely foreseen. Then on easy cases, everyone will form much the same judgments. That's why I thought the hard cases were the ones to test: it is when there is normative disagreement that we will see disagreement about intentions (mutatis mutandis for goals/side-effects).

Neil

[Jonathan]

Neil (if I may),

Your argument concerning DDE, if I understand it correctly, is that the reason the "side-effect" of the death of the attacker (for example) while fending off the attack is acceptable is that it is already acceptable in the normative beliefs of the individual (considering the scenario or experiencing it) to kill such an attacker.

I agree that you have a potential criticism of DDE here.

--Jonathan

Interesting results!

I was hoping that you could address/clarify a particular point:

It seems that what was manipulated in your scenarios was a moral-relevant feature of the actor (and not an action or effect).

So, for example, in the ill pregnant woman example, both variations contained the same act (taking medicine) and the same effect (feeling better and death of fetus), but what differed was the woman's attitude toward children (like or didn't like children).

It seems plausible to think that subjects are more willing to attribute "goal" to fetus death in the "don't like children" manipulation because the actor's *motives* have become questionable.

And it seems that motives would be important to a defender of DDE. So, it seems that if the increase attributions of "goal" turned on some consideration of potential motives, the DDE defender may have a potential objection to your line of argument.

Perhaps I am missing something obvious?

[Neil]

I agree that one can reasonably infer from an agent's attitudes what their motives were. But the second interpretation of double effect was supposed to be an alternative to the first. If we attribute goals to agents on the basis of their intentions, then we have returned to the first interpretation. And I take the Knobe effect to be bad news for the first interpretation.

[Robin Aldridge-Sutton]

Hi Neil

I enjoyed your paper and found it quite convincing.

I was wondering if there is experimental evidence that the Knobe effect works via permissability judgements about harm rather than directly from the causing of harm, or is this just the consensus view about how to interpret Knobe's results? If there were experimental evidence that would seem to me to make your main argument more solid, since people sometimes think harm is permissible.

I think it's really plausible that direct/indirect effects will turn out to work just like intentional/unntentional effects, as your data suggests. They sound like the same thing to me...

Cheers

[Andrew Papanikitas]

The 'experimental artefact' or 'normative loading' (depending on which you believe) is illustrated in British Medical Law in the contrast between R v Adams 1957 (high doses of painkillers did not equate to an intention shorten life though this was forseeable) and R v Woolin 1998(throwing one's baby onto a hard surface in the foresight of harm equated to an intention to harm). Lawyers are still struggling to reconcile the two legal directions, and even argue that 'intention' works differently for doctors!(see Huxtable R and Williams G under subject of double effect).