An Argument for Contextualism about Ethics

The point of this post is to float an argument for contextualism about ethics. I don’t want to _endorse_ the argument. Indeed my main interest is in seeing how it compares to arguments for contextualism about other domains, especially epistemology. The argument takes as its starting point some experiments performed by Daniel Kahneman and Amos Tversky summarised “here”:http://www.workingpsychology.com/lossaver.html.

Kahneman and Tversky asked separate groups of doctors two questions. Here is the first.

bq. Imagine that the U.S. is preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs to combat the disease have been proposed. Assume that the exact scientific estimates of the consequences of the programs are as follows: If program A is adopted, 200 people will be saved. If program B is adopted, there is a one-third probability that 600 people will be saved and a two-thirds probability that no people will be saved. Which of the two programs would you favor?

The doctors generally preferred A over B by an overwhelming margin. A second group was asked this question.

bq. Imagine that the U.S. is preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs to combat the disease have been proposed. Assume that the exact scientific estimates of the consequences of the programs are as follows: If program C is adopted, 400 people will die. If program D is adopted, there is a one-third probability that nobody will die and a two-thirds probability that 600 people will die. Which of the two programs would you favor?

This group favoured D over C by an equally overwhelming margin. Now the doctors were asked a question that is about their personal preferences (which would you favour) but it’s hard to imagine that their answers would have changed if they were asked an explicitly moral question (e.g. which is the right thing to do). If that’s right, their answers to moral questions would seem to depend crucially on features of the context.

So now we are faced with three options.

# Option A/C is really better than option B/D, and the doctors in the second group have a systematic incompetency
# Option B/D is really better than option A/C, and the doctors in the first group have a systematic incompetency
# What the right thing to do is context-dependent, and since the doctors were in different contexts, in particular the contextually salient baseline was different in each case, the majority is correct on both occasions

Now there are several reasons to favour the third alternative. It best follows the principle of charity. It makes the best predictions about intuitions about new cases. And, given the extensive work by Kahneman and Tversky, we can see how to do both of these things in a systematic way. (The general idea here is that people treat losses and gains differently. But whether something is a loss or a gain is dependent on the conversational frame. The contextualist says that hence whether we can properly call an action ‘right’ depends on whether it is producing a gain or alleviating a loss, which is a context-dependent feature.) To the extent that these are the motives driving contextualism elsewhere, we have a motivation for contextualism in ethics as well.

Are there any weaknesses in this contextualist argument that don’t carry across to other contextualist arguments as well?

* _Just Plain Nuts_
It might be thought obvious that A is better than B iff C is better than D. But the contextualist can say this is true. In _our_ context we can say this. What is at issue is whether the doctors who (correctly) picked A over B can also (correctly) pick C over D. Once this objection is carefully framed, it looks just as weak/strong as similar objections against other contextualisms.

* _Not Empirically Supported_
To be sure I’ve done a bit of armchair speculating here about what would (or more precisely wouldn’t) happen if we tinkered with the example a little. But this is surely nowhere near as armchairish as what normally happens in contextualism.

* _Not Stable_
Presumably the doctors will not hold onto their opinions when presented with the AB and CD choices side-by-side. But no one will hold on to their sceptical _and_ anti-sceptical opinions when the cases are presented in an amalgm of the two kinds of cases that ground epistemic contextualism either. Asking both of the questions amounts (from the ethical contextualist’s perspective) to changing the context, so intuitions in the new context don’t count.

* _One of the Answers is Obviously Correct_
If one could make a good argument that one of A/C is better than B/D is correct, then we could argue that one particular group is incompetent. But as such this is actually a hard case, and intuitions don’t tell either way. Note that this is a respect in which the contextualist argument is _stronger_ in this case than in the epistemic case. When an agent moves into a ‘high-standards’ context she normally regards her earlier knowledge attributions as wrong. This is an asymmetry in the cases. (There’s a reason contextualists always elict pro-knowledge judgements _first_ and anti-knowledge judgements _second_, it’s because it is hard to view earlier scepticism as _wrong_ when moving to a lower context, but that’s exactly what the theory predicts you should do.)

* _Semantic Blindness Arguments_
As Cappelen, Lepore, Hawthorne, Stanley and others have argued, contextualisms of any kind make systematically mistaken predictions about the behaviour of the allegedly context-sensitive terms in many complex constructions, especially speech reports. The ethical contextualism floated here is no different. But it is not in this respect _worse_ than other contextualisms.

* _Not Enough Options_
In the cases I’ve given, there are only two contexts. A good contextualism should allow for many contexts. Fortunately the kind of contextualism I’m floating does that. What is set by context is the ‘baseline’ relative to which actions are assessed. This can vary continuously depending on how the conversation goes. (And the variations might matter – the difference between a huge gain and a very huge gain might matter differently than the difference between a small gain and a moderate gain.)

* _Explainable_
The usual way of taking Kahneman and Tversky’s prospect theory is as a psychological explanation of how and why people go wrong in certain kinds of reasoning. Since we have such a good psychological theory, why not use it? I’m sympathetic to this response, so I don’t want to really argue that the contextualist has a way out. But I do want to argue that the ethical contextualist isn’t worse off than any other philosophical contextualist here. Contextualists face a dilemma. Either they provide a systematic story about how intuitions about cases vary with contextual features or they do not. If they do, then we are a long way towards having a psychological theory of the kinds of errors people make, and once we have a theory of how and when errors get made we can discount the intuitions. If they do not, then we don’t have a contextualist theory at all, just a ragtag collection of data. In the ethical case it’s clear that the contextualist will have to deal with the first horn, but I don’t see how *any* contextualist can avoid having to deal with this horn sooner or later.

I’m sure I’m missing something here, but it looks to me like this kind of contextualism – one that takes Kahneman/Tversky experiments to reveal how to change the context in relevant ways rather than ways in which subjects make systematic mistakes – is just as strong as other kinds of contextualism. This suggests that the stakes in contextualism debates are somewhat higher than we might have thought.