Last semester I taught a seminar at Harvard Law School on “law and cognition.” Readings consisted of about 50 or so papers, most of which featured empirical studies of legal decisionmaking. I will now & again describe some of them.
One of the most interesting was “The Plasticity of Harm in the Service of Punishment Goals: Legal Implications of Outcome-Driven Reasoning, ” 100 Calif. L. Rev. (forthcoming 2012), by Avani Sood—whom I convinced to attend the seminar session in which we discussed it—and John Darley (a legendary social psychologist who now does a lot of empirical legal studies).
The paper contains a set of experiments in which subjects are shown to impute “harm” more readily to behavior when it offends their moral values than when it doesn’t. This dynamic, which reflects a form of motivated reasoning, subverts legal doctrines rooted in the liberal “harm principle”—which prohibits punishment of behavior that people find offensive but that doesn’t cause harm.
I liked this paper a lot the first time I read it—as an early draft presented at the 2010 Conference on Empirical Legal Studies—but was all the more impressed this time by a new study S&D had added. In that study, S&D examined whether subjects’ perceptions of harm were sensitive to the message of a political protestor who was alleged to have “harmed” bystanders by demonstrating in the nude.
S&D first conducted a “between subjects” version of the design in which one half the subjects were told that the protestor was expressing an “anti-abortion” message and the other half that the protestor was expressing a “pro-abortion” one. S&D found that subjects more readily perceived harm, and favored a more severe sanction, when the protestor’s message defied the subjects’ own positions on abortion.
That was in itself a nice result (it extended other studies in the paper by showing that diverse moral or ideological attitudes could generate systematic disagreements in perceptions of harm) but the best part was a follow-up, within-subject version of the same design, in which all subjects assessed both pro- and anti-abortion protestors. Subjects now rated the behavior of both protestores—the one whose message matched their own positoin and the one whose message didn’t—equally harmful, and deserving of equally severe punishments.
The result was valuable for S&D because it addressed a potential objection to the paper: that subjects in their various studies didn’t understand that offense to their (or others’) moral sensibilities doesn’t count as a “harm” for purposes of the law. If that had been so, then the results in the within-subject design presumably would have reflected the same correspondence between protestor message and subject ideology as the results in the between-subjects design. The difference suggested that the subjects who had evaluated only one protestor at a time had been unconsciously influenced by their own ideology to see harm conditional on their opposition to the protestor’s message.
This result in fact made me feel better about some of the cultural cognition studies that I and my collaborators have done. In a number of papers, we have been exploring the phenomenon of “cognitive illiberalism,” which for us refers exactly to the vulnerability of citizens to a form of motivated reasoning that subverts their commitment to liberal principles of neutrality in the law.
One of the possible objections to our studies was that we were assuming such a commitment—when in fact our subjects could have been consciously indulging partisan sensibilities in assessing “facts” like whether a fleeing driver had exposed pedestrians’ to a “substantial risk of death” or a political demonstrator had “shoved” or “blocked” onlookers. I think we had reason to discount this possibility before. But based on S&D’s result, we now have a lot more!
I also really like the S&D result because of what it suggests about the prospects & even the mechanics of “debiasing” in this setting. The disparity between their between- and with-subject designs demonstrated not only that their subjects’ conscious commitment to liberal principles were being betrayed by the sensitivity of their perceptions to their ideologies. It suggested, too, that making their subjects conscious of the risk of to this sort of defeat could equip them to overcome it.
One might be tempted to think that all one has to do is tell citizens to “consider the opposite” if one wants to counteract culturally or ideologically motivated reasoning. Sadly, I don’t think things are that simple, at least outside the lab. But that’s a story for another time.