In her excellent article, Misreading Like a Lawyer: Cognitive Bias in Statutory Interpretation,1 Professor Jill Anderson explains an intricacy of sentence meaning that is well known by linguists, generally handled adequately by most of us in normal conversation, but apparently misunderstood and badly handled by lawyers and judges. The misreading in question is based on what linguists call “opaque constructions”: texts whose structural ambiguity creates alternative readings. These alternative readings are generally known (as Anderson points out) as “de re” and “de dicto” interpretations. As the article clearly explains, opaque constructions differ from other sentences by exhibiting “existence neutrality, the availability of nonspecific readings, and substitution resistance.”2 Opaque verbs tend to include those that speak of hypothetical states of affairs or mental states (for example, “intending,” “promising,” and “believing”).3
Anderson’s contribution is excellent and important, and the comments that follow are meant only as quibbles on marginal matters. One concern is primarily one of emphasis: while Anderson asserts that the problem of misreading legal texts that contain opaque verbs is unrelated to conventional disputes between those who favor textualism in statutory interpretation and those who favor purposive interpretations,4 I believe that a focus on statutory purpose can be helpful in these sort of cases (and that Anderson’s own analyses support this conclusion). I will return to this point below.
There is also a second and lesser worry I want to mention briefly: the article presents the problem of reading opaque statutes as pervasive in law — we are told that “in law . . . such off-base responses to statutory language abound” and that “[i]ts costs are enormous.”5 However, I am not sure that the examples the article offers make out that claim. The article only offers a handful of examples, and even among these few there are cases that are many decades old or from foreign and international jurisdictions. To justify a claim of present epidemic or crisis, one might have preferred some showing of numerous examples just in the last few years in the federal appellate courts or over some similar limited period in some other small sample of courts.
The selection of cases might nonetheless be justified if they were all exemplary (and notorious) examples of misreading legal texts with opaque constructions. This does not seem to be the case. Consider one of Anderson’s main examples, the misreading of the statute in the Arthur Andersen obstruction of justice prosecution.6 Upon closer inspection, this appears to have been a problem caused more by an error in understanding precedent than a direct misreading of an opaque construction. The relevant statute made it a crime to obstruct or impede “the due administration of justice.” An 1893 Supreme Court decision7 properly overturned a conviction under the statute for miners whose action had violated a court’s restraining order, but where there had been no evidence that the miners were aware of the litigation or the order. However, that 1893 case was misread by later cases to require (a) that there be an existing judicial proceeding, and (b) the obstruction be directed at that proceeding. To be sure, that misreading was possibly encouraged by a de re rather than a de dicto reading of the statutory language, but the fault in the reading of precedent in the line of cases seems to overshadow the issue of de re versus de dicto.8
To return to the role of purpose, while I agree with Anderson that the problem of misreading norms with opaque verbs is not, strictly speaking, a question of textualist as against purposivist approaches to statutory interpretation,9 the general idea of purposivism — that one should focus on the purpose of the enactment, or, more precisely, the mischief which it was meant to prevent or cure10 — can help direct judges and lawyers to the correct de dicto reading of applicable statutes. For example, a focus on what the statute was meant to prevent or accomplish clearly grounds and supports the reading Anderson (rightly) supports in the genocide case and the voter-impersonation case,11 and other cases as well.
In the Rwanda genocide example, where Hutus accused of the mass murder of Tutsis were prosecuted by an international tribunal under the Genocide Convention, the charge was acting “with intent to destroy, in whole or in part, a national, ethnical, racial or religious group, as such.”12 Arguably this example13 is only secondarily about de re/de dicto; it is primarily about when a list (“national, ethnical, racial or religious group”) should be understood as illustrative rather than exhaustive. Anderson’s suggested reading, that the four categories in the Genocide Convention should be read as illustrative of the larger category (“as a group”) and should focus on the perpetrators’ subjective intentions, is certainly tenable and perhaps even the best reading, but it is hardly an obvious or undeniable interpretation. As the article points out, the conclusion that the list should be read as illustrative is easier to come by in opaque/de dicto readings, but even when a de dicto approach has been chosen, there is still argumentative work to be done before the illustrative conclusion is reached. As Anderson indicates in her own analysis, this is work to be done by a focus on the purpose of the Genocide Convention, in particular, on the evil the Convention was meant to deter (or, when deterrence failed, punish).14 When the Hutus killed Tutsis simply because they were Tutsis, this was clearly the sort of crime the Convention was aimed to deter and punish.
Similarly, the case involving the man who pretended to be a Tennessee Valley Authority worker, who was then prosecuted for “pretend[ing] to be an officer or employee . . . of the United States” but had his conviction reversed because TVA employees were not technically “officers or employees of the government,”15 is comparable. In that case, Anderson’s proposed reading of the statute turns on reading “officer or employee . . . of the United States” broadly.16 Anderson’s preferred reading may presuppose a de dicto reading, but it requires more than that, for one could still tenably argue for a narrower reading. Once more, though, a focus on the purpose of the statute, the mischief it was meant to prevent, could give a solid ground for the broad reading Anderson suggests.
One final note: it is worth observing that the sort of problem discussed in Anderson’s article comes up (as she indicates) in cases involving questions of knowledge and intention (for example, general intent versus specific intent).17 Thus, there are connections with the notorious problem of impossible attempts: if Sarah thinks she is selling cocaine, but the substance is actually baking soda; or if Tom tries to kill Mary by shooting her but Mary is already dead; has either or both committed a crime?18 As Anderson points out, our uncertainty about such cases comes from the use of texts that include terms (for example, “attempt,” “conspire,” or “impersonate”) that relate to mental states and hypothetical or imagined worlds.19 What did the defendant want to do (even if she did not or could not), or what did she want other people to think? Arguably, a focus on “purpose” or “mischief” could once again helpfully guide those applying such rules: we do not want people shooting to kill even if those they shoot turn out already to be dead; and we do not want people trying to sell certain sorts of drugs, even if it turns out that the powder they possess is not actually the drug in question.
Anderson suggests that the problems of misreading on which she focused might be solved by some combination of better training in law school and technology that might warn lawyers and judges when legal texts contain opaque constructions.20 Both recommendations are worth taking seriously. I would add only that responses to opaque constructions should also include an understanding of the importance of focusing on the purpose legal norms serve, the mischief they are meant to prevent.
* Frederick W. Thomas Professor of Law and Philosophy, University of Minnesota.
1. Jill C. Anderson, Misreading Like a Lawyer: Cognitive Bias in Statutory Interpretation, 127 Harv. L. Rev. 1521 (2014).
2. Id. at 1531.
3. Id. at 1531–32.
4. E.g., id. at 1525–27, 1536–37. Anderson writes: “One claim of this Article, then, is that many clashes among theories of statutory construction are simply irrelevant to a significant class of problems that legal actors regularly confront.” Id. at 1526. Anderson then immediately refers to the “clash” between textualists and purposivists as an example of a disagreement she will show to be irrelevant. Id. This Response tries to show that purposive approaches are far from irrelevant to the proper understanding of legal texts with opaque constructions.
5. Id. at 1536, 1592 (emphases added); see also id. at 1525 (“Our mishandling of opaque constructions creates considerable and costly havoc . . . .”).
6. See id. at 1523, 1544–1551.
7. Pettibone v. United States, 148 U.S. 197, 205 (1893), discussed in Anderson, supra note 1, at 1548–50.
8. Anderson considers but rejects the suggestion that this category of cases is primarily a problem of “path dependence,” with a single misreading then locked in by precedent. Anderson, supra note 1, at 1567. I agree that path dependence does not explain all, or even most of these cases, but it may explain some of them.
9. E.g., id. at 1525–27, 1536–37.
10. In English statutory interpretation, this is known as “The Mischief Rule” or “The Rule in Heydon’s Case.” See Heydon’s Case, (1584) 76 Eng. Rep. 637; 3 Co. Rep. 7 a.
11. See Anderson, supra note 1, at 1523–24, 1538–42, 1553–63.
12. See id. at 1523, 1554 (quoting Convention on the Prevention and Punishment of the Crime of Genocide art. II, Dec. 9, 1948, 78 U.N.T.S. 277) (internal quotation marks omitted).
13. Where the outcome, by the way, was the one Anderson would have wanted — conviction under the Convention. She is not content, though, with the way the Convention was interpreted and the reaction of other scholars to the court’s reasoning. Id. at 1554–56.
14. See id. at 1557.
15. See id. at 1542 (quoting Pierce v. United States, 314 U.S. 306, 306 (1941)).
16. The impression that the defendant should have been convicted, and the conviction upheld (even under a different reading of the statute), is supported by another fact Anderson reports: that “the defendant had told at least some plaintiffs that he was employed by the federal government.” Id. at 1542.
17. See id. at 1581.
18. E.g., J. Temkin, Impossible Attempts—Another View, 39 Mod. L. Rev. 55 (1976).
19. See Anderson, supra note 1, at 1531–32.
20. See id. at 1584–91.