This will have some big implications for my work on FantasySCOTUS and the Constitutionality of Social Cost.

What role do policy preferences play when a judge or any other reader decides what a statute or other legal text means? Most judges think of themselves as doing law, not politics. Yet the observable decisions that judges make often follow patterns that are hard to explain by anything other than policy preferences. Indeed, if one presses the implications of the data too hard, it is likely to be heard as an accusation of bad faith – a claim that the judge or other decision-maker isn’t really earnest in trying to separate preference from judgment. This does not advance the discussion, and distracts from the possibility of more interesting explanations. A promising antidote, we believe, lies in empirical study not just of large numbers of judicial decisions collected over time, as previous scholars have done, but of the immediate experience of legal interpretation.

We compile, and here present, rich evidence of what happens when lawyers in training are asked in controlled surveys to distinguish between their policy preferences on the one hand and their own interpretive judgments or predictions about courts judgments on the other. Our findings offer two lessons. First and foremost, they suggest that separating policy preferences from judgments about the meaning of statutes is very difficult. The same is true of preferences and predictions about what courts will do: respondents tend to predict that courts will do what the respondents themselves prefer. The fundamental entanglement of preferences and interpretation raises important questions about the ability of anyone – including judges – to neutrally carry out interpretive strategies meant to generate answers in close cases. Second, however, the results also show that certain ways of framing the interpretive question can reduce the influence of preference on interpretation – though perhaps not its effect on predictions. Instead of simply asking respondents how they would interpret the text of a statute or how the drafters would likely want it applied, it is better to ask respondents how ordinary readers would interpret the statute. This framing of the interpretative question can debias an individual’s interpretation of a statute.

In short, interpretative theories that elevate text alone or give the intent of drafters are both susceptible to contamination by private preferences. To immunize interpretation from these preferences, a theory that asks how ordinary readers would read a statute may be the best prescription.