Blog Coordinator

« Doxastic Self-Control | Main | The Transparency of Belief »



Feed You can follow this conversation by subscribing to the comment feed for this post.

Nice finds! Here's a link for Rose, Buckwalter & Nichols:

Here are four survey items from Rose, Buckwalter & Nichols, about which I was moved to make predictions, before looking at their results. I quote the Neuro-Prediction version only, because that was what my predictions were about.

(Aware Change) When Jill became aware that she was going to vote for Green as Governor, could she have voted for a different candidate for Governor instead of Green? [Yes/No]

(Change Mind) When Jill became aware that she was going to vote for Green as Governor, could she have changed her mind and voted for a different candidate for Governor instead of Green? [Yes/No]

(Activity Change) After the pattern of brain activity occurred, could Jill have voted for a different candidate for Governor instead of Green? [Yes/No]

(Different Pattern) If the pattern of brain activity which lead Jill to vote for Green as Governor had not occurred but a pattern of brain activity which leads to voting for a different candidate for Governor did occur, Jill would have definitely voted for a different candidate for Governor. [Yes/No]

I predicted that % Yes responses would be high (and about equal) for Aware Change and Change Mind, and with somewhat less confidence I predicted Yes% would be lower than that for Activity Change. I separately predicted, without comparison of %s to the other three, that the Yes% would be high for Different Pattern, *despite* the fact that in the authors' minds this apparently "conflicts" with saying Yes to Aware Change and Change Mind.

I was right on all counts. Aware Change: 81% Yes, Change Mind: 83%, Activity Change: 69%, Different Pattern: 75%. How did I know?

Hint: the phrase "could have" appears in Aware Change, Change Mind, and Activity Change, but not in Different Pattern.

Paul, can you say more about whether you think your predictions indicate some flaw in the Rose et al. studies. Jason Shepard and I responded to their paper at the SPP and raised several worries, but I am curious what you think (and whether any critiques would also apply to our studies).

Yes, I am making a criticism. My criticism of Rose et al is a partial defense of Nahmias et al. As I pointed out above, the authors evidently feel that, given the vignette, a Yes to Different Pattern contradicts a Yes to any of Aware Change, Change Mind, or Activity Change. Note that over half of the respondents commit the first two "inconsistencies".

I have a more charitable interpretation of their responses. Respondents take "Jill could have" to mean "Jill had the ability", where this need not contradict "Jill definitely will not exercise the ability." The "could" is not read as the epistemic could of conditional probability, for example P(Jill votes not-Green | Jill becomes aware that she will vote Green) > 0. Nor do they read, say, Activity Change as "it is nomologically compossible that Jill has the pattern of brain activity yet votes not-Green," and not just because respondents don't go around thinking in terms of nomological compossibility.

"Could" is systematically ambiguous. Sometimes it's about ability, sometimes probability, sometimes laws of nature. This, by the way, is how I correctly foresaw that respondents would agree more with Change Mind and Aware Change than with Activity Change. The former two encourage an "ability" reading of "could" via their psychological emphasis. Whereas Activity Change is easily and naturally read either as about ability or in one of the other ways.

Different Pattern doesn't use "could" at all, which is why so many respondents, including Free Will affirmers, get it right. Of course from my perspective, the majority response on *all* the items is right, given a natural reading.

I think Rose et. al. do give good evidence for intrusion. But the number of answers affected by intrusion need not be very large, and the conclusion of Nahmias et. al. may very well stand: It's OK if my brain made me do it.

Dear Paul Torek,

Thank you for your critical comments. I think you make some nice points about possible interpretations that people “could” (here we ago again) have about particular experimental items, that with some effort, could end up being consistent with the stated text. However I’d also encourage you to evaluate the evidence across all the studies presented, and on the basis of that evidence as a whole, consider what it is most responsible for us infer about people’s representations of these stories when making free will judgments. As you mention in your comment, we also find evidence that people think “it’s OK if my brain made me do it", we just doubt that these responses can be taken to support certain philosophical theories with too much confidence given the manner in which they seem to come about. Moreover, I agree with you that "could" is sometimes ambiguous. The suspicion that it is sometimes ambiguous in ordinary language is part of what makes me skeptical of past inferences made in the free will literature on this topic. Hopefully future studies claiming to find compatibilist judgments will measure such judgments more carefully in the context of collecting free will ascriptions, that way we'll be able to actually evaluate whether, given the processing of stories, people's judgments end up being consistent with one philosophical theory over another.



Thanks for your reply. The fact that English uses one word, "could", for these (arguably) different concepts, makes the researcher's job complex. But certainly not impossible! I look forward to future research which will disambiguate some of these questions.

I wonder if there are other languages in which ability and epistemic-possibility have only separate words, and no word which can mean either one? And if so, what would those subjects say, about such scenarios?

Paul and Wesley--

If I may, consider not just "could-have" with its associated ability connotations as against "would have" with associated time-implications derived from "will" that require an alternative time-track for assessment, but also "might have" with plausibly associated opportunity connotations that do not attribute ability in making an appropriate assessment of its use. The modals of assessment for these locutions are complex, straying across questions of ability, time-frames, and opportunity--as well as whether the questions tap metaphysical or epistemic assessments in producing answers. Frankly it seems to me a huge challenge to frame questions that demonstrably disambiguate these locutions in clear and useful ways.

The comments to this entry are closed.