The University of Chicago Law School Faculty Blog

Ambiguity—Reply to Einer Elhauge (round 3)

Once a statute is ambiguous, how should a judge or anyone else decide how to interpret it?  Is it possible that they should consult their policy preferences?  This is the latest interesting question that Professor Elhauge raises.  A few replies to his analysis—

 

—As he notes, we are careful not to argue that policy preferences have no place in interpretation.  That is a long-standing debate, and we mean to offer results and analysis that have value to readers on either side of it.  Some people believe that judges should indeed give effect to their own preferences; but we think that even they would want to keep straight the difference between their preferences and their judgments about the text, even if they plan to give weight to their preferences in the end.  Our studies suggest that this is hard to do.  People try to put their preferences to one side, but the preferences infect the decision about meaning anyway.  In this sense it is possible to speak of "bias"—of unwanted influence—even if the respondents want to give effect to their policy preferences in the end.  They aren't trying to give effect to the preferences at that stage; so at that stage, the preferences are best considered a source of bias.

 

—Professor Elhauge speaks as though, once a statute is found ambiguous, there is nothing more the text can offer; it is time to resort to other methods of decision, like flipping a coin, estimating the preferences of others, or falling back on a default rule (such as finding for defendants, period).  We don't think our respondents were doing any of those things.  Once a statute is conceded to be ambiguous, it's still possible to have an opinion about how its text is best read.  Possible not only in principle but in practice:  the famous cases we use often produce opinions where a judge who says a statute is ambiguous goes on to argue vigorously that one reading is better than the other—not because it squares best with other evidence of legislative intent, but because it's just a preferable reading of the English.  We think this is what our respondents were doing, too.  And it might not be a bad idea normatively.  There are reasonable arguments that the ordinary meaning of a statute should be given much weight on "rule of law" grounds:  the rules that tell people what to do, and subject them to punishment if they disobey, should mean what they seem to mean. 

 

—Asking what ordinary readers would think the statute means is one way of carrying out the interpretive job just described:  "Well, I can see that the gun use statute is ambiguous; but an ordinary reader would certainly not read it as applying to trading a gun, so that's going to be my interpretation, too."  This is a coherent way to proceed, and it has the advantage of avoiding bias in the sense explained earlier—that is, of producing a judgment more likely to be distinct from the policy preferences of the person making it.  Of course it's possible to still think that the final decision should be based more on some other consideration—a statutory default rule, or whatever else.  We are concerned with the competent execution of one stage in the process of deciding a statutory problem.  Whether that stage is the last stage, or less important than others, is a question we leave for other occasions.

 

—Professor Elhauge mentions a possible inference from our forthcoming paper that, when deciding what ordinary readers would think, respondents tend to vote for defendants.  But this isn't really clear.  The pattern does appear in a few criminal cases we tested, but this might just be because the defendants in those cases happened to be taking the simpler and more intuitive view of what the statutes meant.  I consider this an open question.  And when asked what ordinary readers would say, there was movement toward the government's view by many who preferred the defendant's position as a matter of policy.  This is not consistent with the idea that respondents simply put on defendant-colored glasses when asked what ordinary readers would think.  But it is consistent with the idea that the "ordinary readers" question had a debiasing effect on them.

 

—Professor Elhauge suggests that the result might have been different if the respondents had been given more evidence of the legislature's preferences.  No doubt that is true.  I repeat that we are trying to shed light on a stage of interpretation:  the simple judgment of English meaning.  If distortions at that stage are later diluted by the addition of other information (evidence of legislative intent), fine.  But again, judges who have all that evidence still often decide the meaning of statutes in ideologically predictable ways that suggest the effects we describe aren't much diluted.  Maybe the same tendencies we find here also infect the judicial interpretation of all those rich sources of evidence that they had and that our survey-takers didn't.

 

Ambiguity--Posner's Comment

I think Frank has said it all. Whether an "ordinary" English speaker would consider a statutory or constitutional provision ambiguous is irrelevant to whether a judge or other lawyer would consider it ambiguous. The more one knows about a subject, the more--or less--ambiguous a document relating to it might appear. An ordinary English reader might think "freedom of speech" in the First Amendment unambiguous; a lawyer knows better.

Phrases and especially sentences are meaningful only in context, and different types of reader have access to different elements of context. Contract lawyers distinguish helpfully between intrinsic and extrinsic ambiguity, where the former can be spotted with the aid of a very limited knowledge of context, while the latter is discovered only with a broader inquiry. So in the Raffles case, the contract to deliver cotton on the ship Peerless was unambiguous until one learned that there were two ships of the same name sailing between the same ports at roughly the same time (though the difference in date of arrival was what made the ambiguity costly). That is the classic example of extrinsic ambiguity.

In Chapman the major ambiguity stemmed from the weight of the cardboard relative to that of the LSD and the arbitrary consequence (or so it seemed to me) that if the cardboard was treated as part of the mixture or substance containing the LSD the sentence would be much heavier.

When judges say that the "literal" meaning of a statute should be followed unless the result is "absurd," they mean (or should I think be understood to mean) that the statute isn't clear once the context is understood.

And another way to put all this is that a statute is a communication and that in decoding a communication we draw on everything we know about the communicator and the subject matter of the communication in deciding what it means.

THE MOST IMPORTANT AMBIGUITY IN “AMBIGUITY ABOUT AMBIGUITY” By Einer Elhauge

Ok, so I think with Professor Farnsworth’s two responses, we can move past the first two ambiguities.  I do think it would be better in the follow up work to eliminate the concern by varying only the external/internal angle and making the rest of the phrasing identical.  So, given that the internal question asked, “Do you think the statute, as applied to these facts, is ambiguous?”, the external question should ask, “Would ordinary readers think the statute, as applied to these facts, is ambiguous?”  Until that is done, we cannot eliminate the possibility that other differences in phrasing caused the results rather than the external/internal difference, but I agree with Professor Farnsworth that the new study comes pretty close, so this contrary possibility is probably unlikely.

Now we come to my last – and most important – ambiguity about the paper: it’s ambiguous normative implications.  The paper is careful to avoid drawing any explicit normative conclusions, but it does describe the effect of interpreter policy preferences as a “bias”, suggesting some disapproval of it.  Whether or not that was intended, the question arises: would such disapproval be merited?  Should we encourage judges to frame ambiguity questions using an external rather than internal perspective?

If the statute had a clear meaning or we could reliably estimate enactable legislative preferences on the issue, then I think (as I lay out in my book, Statutory Default Rules), that it would be normatively undesirable for judges to use judicial policy preferences rather than statutory meaning or enactable legislative preferences.  I think such a case is usually assumed when one presumes that it is bad for judges to use their own policy preferences.

But here we do not have such a case.  Interpreters who used the external perspective found the statutory meaning unclear, and it is hard to deny the statutes were unclear given that the relevant cases made it to the Supreme Court (which focuses on cases where the law is unclear) and are classics in legislation casebooks (which focus on cases that lend themselves to evenhanded debate).  Nor were any of the interpreters given any evidence about legislative preferences.

So the question arises: if interpreters using the external perspective were not deciding case outcomes based on statutory meaning or legislative preferences or their own policy preferences, what was the basis for their decisions?  Several possibilities arise.

One is that they just decided case outcomes randomly.  But it isn’t clear why we should think that is better than having judges employ judicial policy preferences.  After all, judges are appointed through a political process, so one might think that judicial policy preferences are more likely to reflect societal policy preferences than random choice would be. 

Another is that, consistent with the fact that their own ambiguity judgments reflect their own policy preferences, their prediction about the ambiguity judgments of ordinary readers will reflect their best guess about the policy preferences of most ordinary readers, which then becomes the baseline they apply in deciding case outcomes.  Perhaps such guesses about majoritarian preferences would be better than using judicial preferences, though such guesses might also deviate from actual enactable preferences, which are not the same as simple majoritarian preferences given cloture, bicameralism, Presidential vetoes, and the fact that Senators are not allocated in proportion to population.  (Maybe it would be better to ask what ordinary legislators would think.)  If the external framing did lead interpreters to use their best guess of majoritarian preferences, then I would think that would bias them in favor of the government.  (If political campaigns reflect what politicians think most voters want, it suggests they don’t think voters favor criminal defendants.)  But a sneak peak at the second paper indicates that interpreters using the external perspective instead tend to rule for the criminal defendant.

A final possibility is that interpreters who are not deciding based on statutory meaning, enactable preferences, or their own preferences are using a default rule that favors criminal defendants, probably not because they know the rule of lenity yet, but because they have gotten the general sense from television crime dramas that ties go to the criminal defendant.  This seems consistent with the data in the second paper.  It also happens to be the right result from my perspective because (as my book argues), I think that judges who cannot reliably ascertain statutory meaning or estimate enactable legislative preferences should use the default rule most likely to elicit legislative preferences, which here is the rule of lenity because prosecutors have a much easier time getting legislative clarifications ex ante or overrides ex post, especially given that prospective criminal defendants may not know who they are or be reluctant to identify themselves.  But it isn’t clear to me what would happen in non-criminal areas where television is less likely to lead to implicit default rules.

Finally, it bears emphasis that the relevant questions did not give the interpreters any evidence on legislative preferences or likely effects, nor any background on statutory default rules like the rule of lenity.  It is possible that such evidence and background would change the results.  Interpreters seem more likely to apply their own policy preferences when they know nothing of the legislature’s preferences or about statutory default rules.  Further, if we tested a case where interpreters did know about legislative preferences, we might find that the external perspective would cause interpreters to use their guesses about majoritarian voter preferences rather than rely on legislative preferences, which would seem less desirable. 

So we have some things for yet another follow up paper to explore, which just confirms the value of this generative paper.

Ambiguity in Legal Interpretation: Reply to William Eskridge

Some of Professor Farnsworth's recent reply to Professor Elhauge also speaks to points raised by Professor Eskridge.  In the most recent study we did -- not of whether people find statutes ambiguous, but what of they think the ambiguous statutes mean -- we asked "which reading better fits the ordinary meaning of the statutes text," and variations on this.  We don't think this is a hard question to understand.  And the results we have obtained with these wordings are similar to the results we obtained with the wordings that have been criticized as potentially harder to understand or potentially ambiguous.

As for the use of criminal statutes:  We did these experiments with students in their first week of law school and with students at the end of their first year.  The latter group might be aware of the rule of lenity that Professor Eskridge mentions, but the first group would not be.  But there was no significant difference between the responses from those two groups.  This also suggests that law school doesn't do much to erase the cognitive tendencies we think best explain the results.  We also used a couple of questions based on civil rather than criminal statutes, with the same results shown here. 

Now maybe more years of living and practicing law do have that effect, as Professor Eskridge suggests.  Or maybe they don't.  We can't prove anything about judicial behavior by studying what non-judges do; and we don't claim to have done so.  But we think our results are suggestive, and might help explain a discrepancy familiar to all students of judicial decisionmaking.  On the one hand, there have been studies of judicial behavior from a distance that suggest judges probably give effect to their policy preferences when they interpret ambiguous statutes; and Professor Eskridge thinks this, too.  Yet when you talk to judges you often find that they don't experience themselves as doing any such thing, and some of them get mad when an academic suggests otherwise. 

Our experiments suggest -- in a stylized but plausible way, with a non-judicial population -- some mechanisms by which this discrepancy can be produced.  Respondents in our new study were told not to let their policy preferences infect their judgments about meaning.  It happened anyway.  The subjective experience of separating one's judgment about meaning from one's preferences is not a reliable indicator that the separation is actually happening.  We doubt (and the empirical studies of judicial behavior do not suggest) that this changes all that much when a lawyer gets older and becomes a judge.  It would be like imagining that judges are better than the rest of us at avoiding the endowment effect, or self-serving bias, or other stubborn sources of cognitive trouble.  Maybe they are, but we doubt it. 
 
Professor Eskridge questions whether our “external” question would be as helpful to judges as it is to our respondents.  But he describes the external question as just a way of wording the inquiry, which makes it natural to suggest (as he does) that judges are too “savvy” to be affected by this.  But it isn’t just a matter of wording; the external question really is a different inquiry than simply asking people how clear the statute seems to them.  Sometimes judicial arguments about ambiguity are arguments partly about whether the external question (as we style it) is the right question ¬ as in the Smith case Anup mentioned (the one where the defendant was charged with “using” a gun because he tried to trade the gun for a bag of cocaine). 
 
Another thing to remember is that judges aren’t the only ones who interpret statutes.

Professor Eskridge cites a fascinating article by Brudney & Ditslear (89 Judicature 220 (2005-2006)) that examines the relationship between use of legislative history and the ideological tilt of opinions in the Supreme Court.  That article does not, however, demonstrate that use of legislative history mitigates the effect of policy preferences on judgments.  What it reports is that legislative history is used at roughly the same rate in opinions with liberal outcomes and opinions with conservative outcomes.  That result is also consistent with the prediction that liberals can use legislative history to reach liberal outcomes and conservatives can use it to reach conservative outcomes.  Fortunately the data set that Brudney and Ditslear have gathered may be able to test Professor Eskridge's claim.  We recommend a regression of whether an opinion is liberal on the interaction between use of legislative history and whether the author is a liberal.  If the interaction has the opposite as the main effect on the author being liberal, it means it mitigates the effect of policy preference on judgments.

Finally, in comparing different methodologies, both Professors Eskridge and Elhauge note that our findings suggest that textualism appears not to mitigate the effect of policy preferences.  This is correct.  But to provide a complete picture, it should be admitted that our data do not suggest that textualism is any worse than purposivisim or intentionalism, which are captured by the our drafters' intent question in the second set of surveys.  Both the textual cue and the drafters' intent cue allow policy preferences to affect judgments at similar rates.

Ambiguity in Legal Interpretation: Frank Easterbrook's reaction

I enjoyed the paper. The distinction between internal and external standards of ambiguity is fruitful.

Although the authors ask about "ordinary readers," it may help to figure out who the addressees of a statute are and then ask how these people would understand the text. This is the approach of using an interpretive community to identify meaning—the "community" part being important because linguistic conventions differ (think of the difference between how lawyers and physicians use language), while no one really knows who the "ordinary reader" may be. The more precise we are in identifying the interpretive community, the more likely we are to agree on how they understand a text, and thus what it means.

Unfortunately, a survey of first-year law students (or any other group) needs to hold down the length of questions and therefore does not include all of the information needed either to identify the right interpretive community or understand what tools that community would employ. For current purposes, let's suppose that the right community for criminal statutes is all federal judges serving at the time of enactment. Because the three statutes used as examples in the paper all are recent, we needn't worry about the difference between the interpretive community at the time of enactment and the interpretive community at the time of application.

Judges read statutes in context, not as isolated clauses. So for Chapman v. United States, 500 U.S. 453 (1991), the LSD case, the judges asked not whether the LSD's carrier medium (blotter paper) is part of the same "mixture or substance" as LSD in the abstract (or as a matter of chemistry), but whether LSD-in-blotter-paper is the same sort of mixture as cocaine-in-mannitol or other common dilutants. And the judges knew (as the students answering the survey did not) that the statute provided alternative punishment levels for pure PCP and diluted PCP but treated the weight of all other drugs as including the entire "mixture or substance". This implied that the weight of LSD's carrier counts.

Even so, every judge of the seventh circuit (which sat en banc, see 908 F.2d 1312 (1991)) and every Justice of the Supreme Court deemed the statute ambiguous. It's a surprise to me that so many of the first-year students found clarity where, despite the aid of context, the members of the interpretive community found ambiguity. What divided the judges was not disagreement about the existence of ambiguity but disagreement about how to respond to that ambiguity. Some thought it appropriate to use linguistic context to make a best estimate of meaning; others wanted to put the linguistic context aside and consider practical effects—which when coupled with the rule of lenity or a desire to avoid constitutional questions led to a vote for the defendant.

In a judge's daily work, there is substantially more disagreement about the consequences of ambiguity than about the fact of ambiguity. Textualists see ambiguity in statutes and regulations as often as pragmatists or intentionalists do (well, almost as often). The paper under discussion does not tell us much about how that ambiguity should be resolved. But it does caution against allowing one's own ideas of wise policy (the "internal judgment") to influence when a text will be declared "clear." For Ward Farnsworth reports (in his response to Einer Elhauge) that the internal judgment not only leads respondents to report less ambiguity but also leads them to treat texts as "clearly" meaning what the respondents think is the better policy. Judges can't be voted out of office, so they need an interpretive methodology that reduces the role of their own views of wise policy. No methodology can eliminate it, but the paper does tell us that using an external judgment will reduce the risk that judges will smuggle their own preferences into the law by a confident declaration of "plain meaning".

Ambiguity in Legal Interpretation: Reply to Einer Elhauge (round 2)

Professor Elhauge is perceptive again in pointing out another ambiguity in our paper about ambiguity.  We asked some respondents to our surveys this question:

"Do you think ordinary readers of English would disagree about which side's reading of the statute is better?"

Might some respondents have thought this question was asking them to predict the policy preferences of ordinary readers?  We doubt it; we think the first part of the question (referring to "ordinary readers of English") makes clear enough that by "better," we mean better as a matter of English.  But in any event, we did administer a set of surveys that instead asked “Do you think ordinary readers would disagree about which reading better fits the statute's text?”  This question removed the possible ambiguity.  It produced the same results as the question that concerns Professor Elhauge.

Note this as well:  In our surveys that focused on what the respondents thought the statutes meant, we cured the ambiguity that concerns Professor Elhauge.  This time we compare answers to the follow questions:

1.  Setting your policy preference aside, which reading better fits the ordinary meaning of the statute’s text?

2.  Setting your policy preference aside, which reading of the statute is a better fit to what the drafters of the statute intended?

3.  Setting your policy preference aside, which reading of the statute would ordinary readers of English think is a better fit to the ordinary meaning of the statutes text?

Here again the ambiguity is gone, but the effect turns out to be the same:  the answers to the first two questions tracked the respondents' policy preferences much more closely than the answers to the third. 

On Professor Elhauge's other points:  We don't ask our respondents whether ordinary readers of English would say the statute is ambiguous.  We ask whether they think ordinary readers would agree about which reading of it is better -- not quite the same thing.  Also, Professor Elhauge suggests that the internal question (is the text ambiguous?) is simply a logical question about whether the text will admit of more than one reading. We aren't so sure.  We don't define "ambiguity" for the respondents.  They define it for themselves.  And it's pretty obvious from our fact patterns that the texts are always ambiguous in that logical sense -- that two different readings are possible in principle; yet many respondents nevertheless say that the statutes aren't ambiguous in fact.  Evidently they aren't defining ambiguity in the logical sense that Professor Elhauge has in mind.  We speculate that if we gave our respondents the "draw the drapes" hypothetical, they would rate it as pretty unambiguous in reply to any of our questions about it. 

This discussion shows that ambiguity is itself ambiguous, and that subtleties in the idea are often unappreciated -- which is one of our claims.  Professor Elhauge has helped fortify it by coming at the issue from a slightly different angle than we did, and (we think) from a different angle than courts use.  Showing that a statute is ambiguous just in the logical sense that "draw the drapes" is ambiguous is not generally enough to get a court to declare the statute ambiguous (in a sense that would trigger the familiar consequences of such an announcement). 

Ambiguity in Legal Interpretation: William Eskridge's Response

1. I have argued in both historical work (101 Colum. L. Rev. 999) and empirical work (96 Geo. L.J. 1083) that judges applying textual plain meanings or finding ambiguities are influenced by their personal or political preferences and (relatedly) by normative concerns, at least in the "hard cases" SCOTUS hears. Does the authors' study confirm some evrsion of this idea? I am not sure.

2. The methodology for the questions seems to me to invite normativity from respondents: The authors repeatedly ask which reading of the statute is "better". I have no idea how a first-year law student would understand this kind of inquiry; I'd understand it as asking me to see "ambiguity" (or whatever) as a normative rather than a purely descriptive endeavor. It is hard for me to tell how serious a concern this is.

3. Using criminal statutes may muddy the experiment in several ways, as the authors discuss and maybe in one additional way. When I was a first-year law student, I was aware that criminal statutes were supposed to be construed in favor of the defendant; some students may know this, and that kind of knowledge may make questions about "ambiguity" ambiguous. Probably a minor concern, but it bugs me.

4. Can we generalize from first-year law students (taking the survey) to judges (who decide cases)?  At this point, I do not see why we should make that leap. Judges know a lot more than law students, and they have been socialized into legal thinking; additionally, their decisions are accountable to critical scrutiny from other judges, from the press, from academics, and (for many state judges) from the public/voters. Are judges more "disciplined" in fencing out norms? I do not think so, and my empirical work suggests not, but I do not see how surveys of law students offer conclusive (or even strong) evidence about judges.

5. I am intrigued by the authors' finding that law students were less likely to bend text to reflect their preferences when the question concerned external meaning rather than their own understanding of meaning. For the reasons noted above, I am not sure the authors have proven this, but it is a neat hypothesis. It does strike me that older, legally savvy judges might not be so much influenced by the wording of the inquiry, however.

6. How best to prevent judges (not just law students!) from importing their preferences into judgments about plain meaning/ambiguity?  This may be an impossible task, but if some headway can be made against this tendency, there is empirical work suggesting one source as potentially constraining.  Jim Brudney and Corey Ditslear have done important empirical work arguing that considering legislative history, and not just text and textual canons, actually counterbalances political preference. They have an article in Judicature making this point, and perhaps more recent articles.

Anyway, this is an area for fruitful inquiry.  Thanks to the convenors and the authors.

SOME AMBIGUITIES IN “AMBIGUITY ABOUT AMBIGUITY” – Part II By Einer Elhauge

You know you are dealing with great productive scholars when you point out there is an ambiguity in their study that merits a follow up article, and they are have already  written it.  I gather from Professor Farnsworth’s post that their follow up article will resolve my first ambiguity, finding that interpreters who see less ambiguity are more likely to make decisions that reflect  their own political views.  This brings me back to my point that this study does seem to undercut the argument that textualist methodologies that are designed to avoid conclusions of ambiguity will reduce the influence of the judge’s own politics.

Now for the second ambiguity I saw in the original paper.  The authors write as if the key distinction is between the internal versus external perspective: that is, whether  the question is whether the interpreter finds the meaning ambiguous or whether the interpreter believes that ordinary readers would find the meaning ambiguous.  But that is only one of two dimensions varied by the questions.  The “internal” questions ask a linguistic question about the logical parsing of text: that is, they ask whether the statute is ambiguous or whether another reading is plausible.  The “external” questions ask a predictive question of which reading most readers would find better.  The linguistic/logical question is not really the same as the predictive question.

Suppose, for example, my wife asks me to “draw the drapes.”  If you ask me whether the meaning is ambiguous, I would take it as a linguistic/logical question and have to admit the answer is yes: it could mean either that I should pull the drapes to the side to let the light in or that I should draw a nice picture of the drapes.  If you ask me to predict which interpretation most people would find better, then I would say the pretty clear answer is to pull the drapes to the side.  So I would answer the two questions differently.  But is it because of how they vary the internal/external perspective?  I think not, because if you asked me to predict which interpretation I would find better, I would say the pretty clear answer is to pull the drapes aside.  On the other hand, if you asked me whether most readers would conclude the phrase was ambiguous (in a case where both possible interpretations were pointed out to them), I would say the answer is yes.  So at least in this example, the linguistic/logical nature of the question is what does the work rather than the internal/external element.  To sort these possibilities out, it would be better to have questions that varied only one element at a time.

The issue is related to the fact that the questions ask respondents which reading is “better.”  The authors acknowledge some ambiguity in whether respondents would take this word to mean better as a matter of pure interpretation or as a matter of policy, but they argue this issue cannot explain their findings because all the questions had the same “better” phrasing.  However, there is a crucial difference.  The ambiguity/plausibility questions began by asking whether the statute was ambiguous or had another plausible interpretation and thus in context clearly indicated that “better” meant better along the dimensions of ambiguity or plausibility rather than of policy.  The ordinary readers question had no such antecedent, it just asked about the extent to which ordinary readers would disagree about which reading is “better.”  One might well answer this question assuming that “better” means better as a matter of policy, and thus that the question asks about the extent to which most readers would agree about which reading is better as a matter of policy.

But perhaps this will also be addressed in the follow-up paper that I eagerly anticipate.

Ambiguity in Legal Interpretation: Some thoughts in reply to Judge Williams

Judges often have to decide whether statutes are ambiguous, because that determines, e.g., how the Chevron doctrine applies, or whether some canon of construction applies, or (depending on the judge) whether certain sorts of evidence of the statute's meaning should be considered.  Of course the lawyers' briefs will point out alternative possible meanings, and the good judge will take them seriously.  But that doesn't mean the judge inevitably will find ambiguity.  Sometimes the judge just decides that despite what the lawyers say (and maybe despite what other judges say), one interpretation is so much better than the other that the text really isn't ambiguous.  And that decision can be influenced to a greater or lesser degree by the preferences of whoever makes it.

That's where our study comes in.  It shows that for at least some populations, asking "is the statute ambiguous?" isn't the same as asking "would ordinary speakers of English agree probably about the meaning of the statute?"  Judicial opinions often argue about ambiguity without making clear which of those two questions they mean to be answering.  Perhaps judges assume that they really are the same question.  But they needn't be.  To say a text is "unambiguous" could be a statement that the judge (or whoever is talking) is very sure how best to read the text.  Or it could be sort-of empirical claim that most people would agree about what the text means.  Sometimes judges seem to argue with each because some of them are coming at the question the first way and others are coming at it the second way.  (This is one of the "ambiguities about ambiguity" we refer to in the title of our paper.)

Our study suggests that there is an important difference between these two ways of thinking about ambiguity.  The point isn't quite that people sometimes think a text is unambiguous when in fact there is lots of disagreement about its meaning—though of course that's true.  The point, rather, is that thinking about those two questions produces different likelihoods that the answers will be entwined with (probably "influenced by") the policy preferences of the reader.  Those who consider whether ordinary readers would agree about the meaning of a text—which is a question some judges fuss about more than others—are more likely to be able to give answers that are independent of their own policy judgments.  This may not achieve true "neutrality" (which for these purposes might just mean an answer uninfected by policy preferences), but it does better than other questions we have researched.

Of course some might say that they want policy judgments to play a part in interpretation.  That's fine, but we still think that most judges like the idea of being able to think separately about questions of ambiguity and meaning on the one hand and about their own policy preferences on the other, even if they go on to weigh those considerations differently when they decide who wins the case.  (Another objection is that we didn't experiment with judges; we experimented with law students.  Fair enough.)

Ambiguity in Legal Interpretation: Reply to Einer Elhauge

Einer is right to point out an ambiguity in our paper:  okay, so people with strong policy preferences are more likely to find statutes unambiguous – but do they then read the statutes to mean what they want them to mean? The answer, as it turns out, is yes.  Perhaps we didn’t make that as clear as we should have in the paper we recently published, because we wanted to save those second findings for the next article (which will be released soon).  But in short, and as Anup said in his first post, people with strong policy preferences tend to read statutes to conform with those policy preferences – or at least they do that when they are asked which reading is most consistent with the ordinary meaning of the text.  But when they are asked what reading they would expected ordinary readers to think most consistent with the ordinary meaning of the text, they are able to let go of their preferences and give answers independent of them.