Misreading like a lawyer

« previous post | next post »

Jill Anderson, "Misreading Like a Lawyer: Cognitive Bias in Statutory Interpretation", Harvard Law Review, Vol. 127, 2014, Forthcoming:

Statutory interpretation dilemmas arise in all areas of law, where we often script them as scenes of conflict between a statute’s literal text and its animating purpose. This article argues that, for an important class of disputes, this supposed discord between text and purpose is an illusion. In fact, lawyers are overlooking ambiguities of literal meaning that align well with statutory purpose. The form of ambiguity in question inheres not in individual words, but at the level of the sentence. What triggers a split in readings are verbs that linguists classify as "opaque," which are perfectly common in legal texts: intend, impersonate, endeavor, and regard are among them. In ordinary speech we resolve their dual readings unconsciously and without difficulty. In law, however, our failure to notice multiple readings of ambiguous language has left a trail of erroneous judicial determinations and doctrinal incoherence across a broad swath of law, from disability rights to white collar crime to identity fraud to genocide. Drawing on examples from these areas, this Article uses the tools of formal semantics to expose the ambiguity of opaque constructions and to make visible the family resemblance among the ways we misinterpret them. It then turns to the question of why lawyers misread and what we can do about it. The converging literatures of language development and the psychology of reasoning suggest an answer. When we analyze opaque sentences explicitly as statutory interpretation requires (as opposed to spontaneously in conversation), we may be particularly vulnerable to cognitive bias. Factors peculiar to law tend to amplify and propagate this bias instead of dampen and contain it, but they may also point the way toward more sophisticated legal reading.

An earlier paper by Jill Anderson on a related topic: "Just Semantics: The Lost Readings of the Americans with Disabilities Act", Yale Law Journal, Vol. 117, No. 992, 2008.  I could have sworn that I posted on that excellent earlier paper, but it seems that I just mentioned it in a comment.

Some LL background on semantic "opacity" in legal (and other) interpretation:

"Rarely better than de re", 10/23/2005
"Adverbial modification at the Supreme Court today", 2/25/2009
"Grammatical justice is served", 5/7/2009

And a few of the posts where I've indulged in the rent-seeking (but still valid) argument that more lawyers ought to get some relevant training in linguistics:

"Lawyers in need of linguistic training", 8/7/2006
"Law as applied linguistics", 7/25/2009
"The indubitable equivalent of such claims", 12/16/2009
"Corpus linguistics in a legal opinion", 7/20/2011

[Tip of the hat to Garrett Wollman]

 

Share:



33 Comments »

  1. Jonathan Mayhew said,

    September 27, 2013 @ 8:55 am

    I like this one: 'The court acquitted him, albeit reluctantly. There had been voter fraud by impersonation, certainly. But the court fixated on the object of the impersonation and concluded that because a dead person could not vote, the defendant had not impersonated a “person entitled to vote.”'

  2. Neal Goldfarb said,

    September 27, 2013 @ 9:53 am

    I've only taken a quick glance at the article (which looks good, by the way), but one thing that I noticed was its discussion of "fast" and "slow" forms of cognition:

    Cognitive science has been transformed over the past several decades by dual process theories of cognition and the role of heuristics and in reasoning. This view of human thought has recently been popularized in Thinking, Fast and Slow by Nobel laureate Daniel Kahneman, who with Amos Tversky pioneered the field and whose work is widely cited in interdisciplinary law and psychology. According to this family of theories, our thinking is negotiated between two types of cognitive mechanisms. System 1 or “fast” thinking is cognitively easy, intuitive, associative (making connections by similarity as opposed to computational reasoning) and automatic. By contrast, “slow” System 2 operations are effortful, analytical, algorithmic, and conscious.

    Not having read the article, I'm not in a position to comment on what it goes on to say about the two types of thinking and how they relate to interpreting statutes. But I think it's worth noting two points.

    First, statutory interpretation is a System 2 process. It's "effortful, analytical, … and conscious"; algorithmic, not so much, but three out of four ain't bad. Of course System 1 processes play a role: the first step in interpreting a statute is simply to read it and comprehend it as a natural-language text. So the process requires the use of the kinds of System 1 processes that are among the subjects of psycholinguistics. But from that starting point, it goes on to conscious reasoning, which is part of System 2.

    Second, and here's where it gets interesting, to the extent that the goal of statutory interpretation is to determine the "ordinary meaning" of the statutory text, the System 2 process of interpretation can be thought of as an attempt to arrive at an idealized version of the outcome of the System 1 process of comprehension. In other words, when lawyers and judges talk about the ordinary meaning of statutory language—or, as one often sees, about "the most natural reading"—what they're really talking about, I think, is what fluent speakers of English would understand the language to mean, where that hypothetical understanding is the result of simply reading the text without putting any conscious effort into thinking it about what it means.

  3. Garrett Wollman said,

    September 27, 2013 @ 11:02 am

    I'd note that this article cites one of MYL's old posts (on de re vs. de dicto readings) and also work by LL'er Barbara Partee.

  4. D.O. said,

    September 27, 2013 @ 12:24 pm

    I was rereading old Prof. Liberman's post "Rarely better than de re" and thought about this example our host gives: "Kim wants to score higher than Leslie scored." It actually includes (at least) 3 distinct possibilities, not 2.

    1) Kim knows that there is some already achieved high score which he wants to beat, without the knowledge that it was in fact Leslie who achieved it. (Say, Kim wants to score better in a man-against-computer game than the current highest score, without the knowledge about who made that highest score. But we know it was Leslie)

    2) Kim wants to achieve a score, which in fact would be a higher score than Leslie has achieved, without knowing (and/or care) that this particular score was achieved by somebody, to say nothing about Leslie. (Imagine, Kim wants to win 5 jackpots in his lifetime without the knowledge that the best result so far was 3 jackpots. That result, in fact, was due to Leslie.)

    3) Kim competes directly with Leslie and wants to beat his score no matter what the letter might be.

    I am not sure how to put it in terms of the scope of existence operators, but those 3 are clearly distinct. We can call the second possibility complete de re, the first choice de re/de dicto mix, and the third one complete de dicto.

  5. D.O. said,

    September 27, 2013 @ 12:30 pm

    Man, those philosophies are difficult. In my example 1 it should read: Kim wants to score better in a man-against-computer game than the current highest score, exclusively because it is the highest score, no matter what that score in fact is and maybe without even the knowledge of what this score is and without the knowledge about who made that highest score. But we know it was Leslie.

  6. J. W. Brewer said,

    September 27, 2013 @ 12:39 pm

    Very very few close observers of the U.S. legal system (either inside or outside the academy) think it suffers at present from the problem of not sending a sufficiently large number of people to prison, so a scholarly perspective that the prisons are insufficiently full because of an alleged persistent failure by all participants in the legal system to read statutes as draconianly as the legislators supposedly intended is bracingly contrarian.

    The other thing is the importance of what she perhaps too quickly hand-waves away as "path dependence." Statutes are often harder to read then the judicial decisions authoritatively (whether correctly or incorrectly in the abstract) glossing them, so once a statute has been interpreted, you counsel your clients to try to avoid violating the statute-as-interpreted. There are substantial problems of fairness and due process in criminally prosecuting someone for conduct that the courts have already said isn't illegal on the basis of a clever argument that those courts all got it wrong because they muddled up de re w/ de dicto (or even suing them under a complicated federal statute like the ADA – you want employers to develop policies and procedures ensuring compliance with what the law has been interpreted to mean; keeping yourself out of trouble by also trying to comply with speculative academic theories about what the law would be interpreted to mean if the courts did a better job, which will inevitably fail to harmonize with each other, is not an achievable goal).

    That said, it's an interesting piece which I hope to read with greater care.

  7. Joe said,

    September 27, 2013 @ 1:40 pm

    @D.O.: The "Just Semantics" paper agrees with you that there are 3 ways to parse the statement (it even has pictures!). However, the paper would categorize your 1) and 2) both as "de re" constructions (ie, having a particular high score as the "res"). According to the paper, 1) is a "wide scope de re" and 2) is a "narrow scope de re".

  8. Neal Goldfarb said,

    September 27, 2013 @ 1:49 pm

    I've just begun to read the article, so I'm still not ready to comment on it, but I can certainly talk about the background assumptions that are going to inform my reading of it. And one of those assumptions is that there is has not been (to use J.W. Brewer's formulation) a "persistent failure by all participants in the legal system" to recognize certain kinds of possible interpretations.

    My guess is that in most if not all cases in which the statute contained a de dicto / de re ambiguity, the interpretation that the court didn't adopt was advocated by one of the parties. So it's not as if lawyers are blind to the ambiguity.

    The real problem, IMO, is that lawyers in general do a good job of arguing the issue, because they don't have the analytical toolkit or vocabulary that's needed to do so. I know from prior communications with Jill Anderson (author of the article) that she basically agrees with this, and the point is discussed in the article: Two of the sections are headed, "Tools lawyers lack" and "Tools lawyers use that don't work."

  9. D.O. said,

    September 27, 2013 @ 2:10 pm

    @Joe. Thank you. I know I have to read "primary sources" for the blogposts, but…

  10. Steve said,

    September 27, 2013 @ 2:14 pm

    I certainly found it very interesting, and instructive in many ways. However, I find myself feeling like a bit of sleight of hand occurred in the "genocide" discussion, not that I in any way disapprove construing the statute in question as forbidding the systematic slaughter of a people. However, I have a problem with the general statement that, in a de dicto reading of a sentence with several categories, each separated by an "or", a statute can be literally interpreted as meaning that a thing which shares some of the qualities of the specified items thereby fits within the definition, even if it does not fit within any one of them. She facilitates this reading by inserting "a people, to whit" in front of race, ethnicity, nationality, and religion, in her de dicto paraphrasing of the statute.

    However, the statute, as originally codified, does not mention a "people", And it references "group" only in conjunction with the specified criteria. And, importantly, I think, it does not include catch all language such as "or another group with similar characteristics" or "or any other group defined by an immutable characteristic." It is certainly reasonable to infer, based on its history, that the drafters DID intend such a result, but I think one can get there only by importing the legislative history into the analysis, and not through what purports to be a "literal" reading of its language.

    It should be noted that legislatures very frequently DO include catch all provisions of this type, precisely because it is sometimes difficult to conceive of every thing that qualifies as the "x" that one wants to regulate. Where such langauge is present, it is entirely proper (indeed, necessary) for a court to determine what "other" things are similar enough to, or share enough characteristics with, the specified items. Where such language is not present, it is problematic to insert it, absent either vagueness, ambiguity, or the need to avoid an absurd result.

  11. Anthony said,

    September 27, 2013 @ 2:37 pm

    " It then turns to the question of why lawyers misread …"

    There's a very simple answer to this: Lawyers misread when it is in the interest of their clients to do so.

    Of course, it's only in the interest of their client if they can convince the Court that a particular misreading generates sufficient ambiguity to benefit their client, rather than convincing the Court that the lawyer is an idiot and the client obviously in the wrong if the only avenue they have is to deliberately misinterpret the words of the statute or previous decisions. Thus, it's partially also due to the judicial culture.

  12. J. W. Brewer said,

    September 27, 2013 @ 3:35 pm

    I had never heard of Pierce v. United States. I just read it. (It's short.) I don't expect U.S. law professors to be able to coherently speak about linguistics, but I do expect them to be able to accurately read and summarize U.S. Supreme Court decisions. Professor Anderson's criticism of Pierce does not appear to me to reflect an accurate reading or understanding of Pierce. This frankly makes me question whether I should bother to read the rest of the article, if I'm not willing to invest the time to fact-check all fact-checkable assertions made therein.

  13. J. W. Brewer said,

    September 27, 2013 @ 4:05 pm

    Separately, arguments about the Genocide Convention are made complicated by the fact that the rather ad hoc court dealing with alleged offenses in Rwanda was composed of judges from a number of different nations, a majority of whom were probably not L1 Anglophones and were not professionally trained in Anglo-American-style legal systems. I'm not sure what level of sophistication in decision-making you can possibly expect under such circumstances, but it would presumably be politically unpalatable to have an international tribunal composed only of judges from such similar linguistic and cultural backgrounds such that you might reasonably hope for high-quality application of a single (necessarily contestable because not universal) judicial methodology.

    The Genocide Convention recites on its face (if the English text I found on the internet is accurate . . .) that it is supposed to be pentalingual, and that the English, French, Spanish, Russian, and Chinese texts are all equally authoritative. (This is perhaps happy-clappy U.N. propaganda that was not intended to be taken seriously, but let's pretend to take it seriously as a thought experiment.) Are all five texts equally susceptible of the same alleged de re / de dicto ambiguity (where I'm frankly having a bit of difficulty following the argument even just looking at the English text)? Are the arguments for the relative plausibility of one reading over another of equal strength in all five languages? Is there a single lawyer or judge in the world with sufficient fluency in that particular set of five languages to even have an informed opinion on those questions?

  14. J. W. Brewer said,

    September 27, 2013 @ 4:13 pm

    Sorry, just to make sure I'm not read as supporting Anglophone hegemony or the universalization of American legal culture, I think a court composed solely of Francophone judges from various countries whose legal/judicial culture (including standard methodologies of statutory interpretation in a criminal context) was derived from that of the Napoleonic Code etc., all working off the French text of the Genocide Convention, would be just as coherent an approach as a court composed solely of Anglophone judges from, let's say, British Columbia, New Zealand, Barbados, and Bangladesh, all working off the English text. Would those two hypothetical tribunals converge on the same interpretation? I have no idea.

  15. Garrett Wollman said,

    September 27, 2013 @ 5:22 pm

    @J.W. Brewer: The French text reads:

    Dans la présente Convention le génocide s'entend de l'un quelconque des actes ci-après, commis dans l'intention de détruire, en tout ou en partie, un groupe national, ethnique, racial ou religieux, comme tel: [...]

    The Engilsh text:

    In the present Convention, genocide means any of the following acts committed with intent to destroy, in whole or in part, a national, ethnical, racial or religious group, as such: [...]

    So the translation is nearly word-for-word. I am not sure if this weakens or bolster's Anderson's claims. Of all the examples she gives in her article, this is the one whee I find it most difficult to credit her alternative reading.

  16. Dan Hemmens said,

    September 27, 2013 @ 5:45 pm

    Professor Anderson's criticism of Pierce does not appear to me to reflect an accurate reading or understanding of Pierce.

    I'm not a lawyer, but what do you think was wrong with Anderson's reading of Pierce?

    I'm not sure I'm convinced by the linguistics (then again, I'm not a linguist either), but the law seems reasonable (it does seem to have been overturned because an employee of the TVA is not an employee of the federal government, and that therefore to impersonate the former is not to impersonate the latter).

  17. J. W. Brewer said,

    September 27, 2013 @ 6:20 pm

    I will admit I'm having trouble following (albeit perhaps due to my ignorance of the fine details of Rwandan sociology) why people found apparently found it so difficult to categorize the Tutsi as an "ethnical . . . group" (assuming "ethnical" is just a weird or archaic synonym for "ethnic"), since ethnicity is already a nicely vague category,* but if there's some fuzziness around the outer boundaries of what can be accurately called an "ethnical group" in English, I wouldn't assume that the fuzziness and/or definitive outer boundaries are located at exactly the same distance from the paradigm case when you're talking about what can be accurately called "un groupe ethnique" in French. The English NP could easily be either broader or narrower in semantic scope than the seemingly-equivalent French NP, and if for whatever reason the Tutsi are a borderline case, that difference in scope could matter.

    French wikipedia says in one place (no guarantee there aren't contradictory statements elsewhere): "Tutsis : Eleveurs du Rwanda et du Burundi, assimilés à un groupe « ethnique »." Does that mean *not* "un groupe ethnique," strictly speaking?

    *There's I suppose the path-dependence possibility that some prior legal authorities tried to deal with the potentially worrisome vagueness of "ethnicity" (or "ethnicality"?) by trying to gloss it with a more precise-sounding definition that, when looked at carefully as applied to Rwandan circumstances, is too narrow and doesn't cover the full range of what it ought to cover, with the judges dealing with the Rwandan cases not feeling that they had the authority to discard that imprudent prior gloss and start again fresh with the language of the treaty itself. But that's a different problem.

  18. J. W. Brewer said,

    September 27, 2013 @ 6:54 pm

    Dan Hemmens: Assume three possibilities: a) the defendant falsely claims to work for the federal government, without saying anything specifically about working for the TVA; b) the defendant falsely claims to work for the TVA, and falsely (but believably, given popular confusion at the time about the technicalities of the TVA's connection with the government) claims that the TVA is a federal agency; or c) the defendant falsely claims to work for the TVA, but makes no claim that the TVA is a federal agency (but either the victims of the fraud or, perhaps more problematically, members of the jury might have a preexisting erroneous impression that the TVA was a federal agency). Assume further that the defendant might have said all three versions at different times to different victims, since conmen have been known to vary the details of their script, and thus alleged instances of all three are floating around in the trial testimony (although the jury is obviously entitled to pick and choose which testimony it believes and which it doesn't). Anderson as I read her seems to think that the Court said B is not a federal crime (analogously to the case of the guy who voted while claiming to be his dead neighbor). As I read the decision, the Court said no such thing, but merely held that the defendant had under the circumstances a procedural right to have the jury affirmatively told that C, by itself, was not a federal crime, and that he shouldn't be convicted if the jury thought that C was all he had done (or for specific incidents where C was all that he had done), and that the conviction therefore ought to be reversed (presumably without prejudice to a retrial with proper jury instructions) because his contemporaneous request to have the jury given that instruction had been wrongfully denied by the trial judge. It was, to use the jargon, an instructional-error issue, not a sufficiency-of-the-evidence issue.

  19. Dan Hemmens said,

    September 28, 2013 @ 5:24 am

    Ah, I see what you mean.

    I think I agree that Anderson is misreading the situation (although, as I say, I am neither a lawyer nor a linguist), it's just that I think she's misreading the language rather than the law. That is, I read her argument as being that impersonating a member of the TVA *constituted* impersonating an employee of the federal government (so that C should be prohibited by the de dicto reading of the statute, even though the TVA was not a branch of the federal government).

    I've not read the whole paper yet, but I think Anderson's argument hinges on a very slightly tenuous interpretation of the existence-independence of opaque verbs. I think her argument about Pierce vs United States maps closely to her argument about Rwanda. She seems to be arguing that it is possible to impersonate an employee of the federal government, even if you are not impersonating an employee of an organisation which is not actually a branch of the federal government (and, crucially, whether you say that the organisation you represent is part of the federal government or not). Similarly, she seems to be arguing that it is possible to act "with intent to destroy an ethnic group" even if the group of people you are attempting to destroy do not, in fact, constitute an ethnic group.

    This strikes me as a little bit tenuous (but not completely unsupportable). It feels like arguing that (to use some of her earlier examples) that a person can be said to "want to eat a cupcake" even if they want to eat a salad, or to "want to buy a house on Green Street" even if they actually want to buy a house on Red Street.

    To put it another way, by my very non-expert reading, she seems to be arguing that "intends to destroy an X" can mean not only "intends to destroy a specific entity, which is in category X" and "intends to destroy a nonspecific entity, provided said entity is in category X" but also "intends to destroy a specific entity, which may or may not be in category X."

    I can *almost* see this argument, but far from being a natural interpretation of language from which lawyers deviate it feels to me more like a rather outre piece of philosophy of language. You could imagine, for example, a person who has spent their whole life believing that the word "cupcake" describes a pickled herring fillet rolled into a cylindrical shape, often around a savory filling" and that such a person could, when experiencing the desire to eat a pickled herring fillet rolled into a cylindrical shape around a savory filling, could be said to "want to eat a cupcake."

    By the same logic, a person who impersonated a member of the TVA in an environment in which many people could be expected to believe (with or without prompting) that the TVA was a federal agency could be said to be impersonating a federal employee. And I think you can actually make quite a concrete argument that a person who kills a large number of people believing them to be an ethnic group has acted "with intent to destroy an ethnic group" (although here it is hard to see how you would *prove* that somebody perceived a group of people as an ethnic group).

  20. Jonathan Mayhew said,

    September 28, 2013 @ 9:37 am

    I agree her reading of Pierce seems wrong. The case about impersonating a person entitled to vote is quite different. If the TVA is not part of the government, then it is not impersonating a government employee.

  21. Robert said,

    September 28, 2013 @ 9:45 pm

    Legislative intent aside, surely pretending to be, alive, a dead person whose name is still on the voting register is precisely impersonating "a person entitled to vote".

  22. Dan Hemmens said,

    September 29, 2013 @ 4:59 am

    @Robert:

    I think that's exactly the de re vs de dicto distinction that Andrews is talking about. By a de re reading of the sentence, the fact that the person you are impersonating is dead (and therefore not entitled to vote) means that you are not impersonating a person who is entitled to vote, you are impersonating a person who is not entitled to vote, and then *secondarily* falsely claiming that said person *is* entitled to vote.

    The reading by which impersonating a dead person who remains on the electoral register constitutes "impersonating a person who is entitled to vote" is the de dicto reading. I agree (and Anderson clearly agrees) that this reading is more natural and equally literal. I think that's the basis of her argument.

    @Jonathan Mayhew

    I *think* I can sort of see her linguistic argument, but I think it's an edge case rather than something as clear cut as the "person entitled to vote" issue.

    I think you could make a good case that if you impersonate a member of an organisation which you expect people to *believe* is a branch of the federal government, then you are impersonating an employee of the federal government. This strikes me as a direct analogy to your case B above of telling people that you work for group X, and that group X is a branch of the federal government.

    For example, as far as I know the Counter Terrorism Unit that Jack Bauer works for in 24 is entirely fictional (I might be wrong about this, but let's assume it is for the sake of argument), but I'd lay good money that not everybody knows this. If you were to perpetrate some fraud by claiming to be a member of the Counter Terrorism Unit then you would not be impersonating an employee of the federal government de re (because there is no such thing as the CTU in real life, so claiming to be a member of it is merely claiming to be a member of a fictional organisation) but you could still be said to be impersonating an employee of the federal government de dicto because you are claiming membership of a body which you are allowing (and expecting) people to believe is a branch of the federal government.

  23. Jonathan Mayhew said,

    September 29, 2013 @ 9:02 am

    Good point, but the TVA is not fictional, so you could make a finding of fact about that question. For me this is not a primarily linguistic issue.

  24. Jonathan Mayhew said,

    September 29, 2013 @ 9:09 am

    … and, really, the meaning of the word "impersonate" is not at issue in either case. In one case, it is whether the person impersonated is entitled to vote; in the other, it is about whether the person impersonated is part of government.

  25. Jonathan Mayhew said,

    September 29, 2013 @ 9:51 am

    On closer reading, I have a problem with the paper's self-assurance. I think there are substantive issues in the statutes that the author sweeps under the rug. For example, a certain kind of sentence is ambiguous or "opaque," but the paper wants us to opt, always, for one particular interpretation of the language, which they view as self-evident or black and white: "Lawyers’ self conception as masters of rigorous reading may in fact be inhibiting our ability to detect meaning that is before us in black and white." There are a lot of grey areas here that have to be considered, in all the cases they discuss.

  26. Dan Hemmens said,

    September 29, 2013 @ 10:56 am

    … and, really, the meaning of the word "impersonate" is not at issue in either case. In one case, it is whether the person impersonated is entitled to vote; in the other, it is about whether the person impersonated is part of government.

    I think it's about the meaning of "impersonate" in the sense that (by Anderson's argument) the question is not about the person impersonated but the person impersonating (again, I think this is the distinction between the de re and de dicto readings).

    This is fairly unambiguous in the "person entitled to vote" case, in that if you pose as another person in order to vote in an election, you are fairly clearly impersonating a person entitled to vote, even if the specific person you are impersonating is not technically entitled to vote.

    It's more complex in the case of the TVA because since a member of the TVA is not a member of the federal government, impersonating a member of the TVA by itself wouldn't be impersonating a member of the federal government. But if a person *believed* that the TVA was a branch of the federal government, then you could make a case that claiming to be a member of the TVA would constitute impersonating a member of the federal government to that person. Of course this assumes a definition of "impersonate X" which is something like "behave in a manner intended to deceive others into believing that you are X", and I have no idea how impersonation is defined legally.

    But, like you, I disagree with the article's sense of certainty, and Anderson's seeming insistence that the de dicto reading is always the most natural or most correct reading. I think this was the core of my disagreement with her analysis of the Rwandan genocide. It seemed unsupportable to me to use a de dicto reading of a sentence which clearly *does* refer to a specific entity (the entity in this case being the group which the defendant intended to destroy). And it is certainly the case that in none of her examples is interpretation unambiguous.

    Although, to be fair to Anderson, her argument seems to be not that the de dicto readings are unambiguously correct but that they are as literal as the de re readings, and that the courts should not feel restrained to follow only de re interpretations of opaque verbs.

  27. Jonathan Mayhew said,

    September 29, 2013 @ 11:50 am

    I think your last paragraph is key. It is not a question of literal vs. non-literal, but of two prima facie readings that are equally "literal," but in opposing ways. I don't think lawyers read with a "cognitive bias," but that we expect legal language to be ultra precise, so that a far-reaching federal prosecutor could not file witness tampering charges in a case that didn't yet exist when the alleged tampering took place. In other words, "opaqueness" has a significant role in circumscribing the potential (over)sweep of a law.

    Disclaimer: I am neither linguist nor attorney.

  28. Jonathan D said,

    September 30, 2013 @ 12:27 am

    It seems to me that even if the examples Anderson gives have a commonality in the linguistic issues involved (although the lack of examples related to her genocide point in section II is indicative of something), there are some different contributing factors when it comes to why Anderson's preferred reading is ignored by the legal profession.

    The electoral impersonation seems to be the clearest case of misreading, and in some ways it is analagous to impersonating a police officer. I have no idea whether such laws are argued to cover, or be intended to cover, cases where a police officer impersonates (de re) a specific different police officer (I would think that when there is serious motivation there are more appropriate ways to describe the offense.), but the case of a civilian claiming (de dicto) to be police is the one most easily thought of when writing/reading such a law. In contrast, electoral laws are intended to stop people impersonating (de re) a different person entitled to vote, even if the offender is entitled to vote and so arguably not impersonating a person entitled to vote in a de dicto sense. It's clear that impersonating a dead elector is literally against the law in a de dicto sense, but it doesn't really help to just resolve the ambiguity that way – it's hardly surprising that it wasn't. Perhaps the drafting really is to blame, in describing the impersonated as entitled to vote, when the issue is really whether the impersonation (de re) is used in order to vote?

    J. W. Brewer's explanation of Pierce, and even Anderson's own description of Pettibone, make me wonder whether ignoring the de dicto reading really was the issue in either case, and certainly path dependence seems relevant. I'm all for arguing that de dicto readings are literal readings, and avoiding any hassle if this is indeed being ignored (and she is quite clear that the problem she's addressing isn't "insufficiently full prisons"), but the case doesn't seem particularly convincing.

  29. Ginger Yellow said,

    September 30, 2013 @ 8:01 am

    I don't think lawyers read with a "cognitive bias," but that we expect legal language to be ultra precise, so that a far-reaching federal prosecutor could not file witness tampering charges in a case that didn't yet exist when the alleged tampering took place.

    Is that not a cognitive bias in itself? I'm not a lawyer, but I've reported on many cases involving contractual construction (said contracts having been extensively lawyered by the best paid lawyers in the world) and several involving statutory interpretation. In my experience, legal language is not in fact anywhere near precise – there's a whole category of contractual terms which are deliberately vague (eg "reasonable efforts") and often attempts to create precision instead lead to more ambiguity or irreconcilable interpretations.

  30. J. W. Brewer said,

    September 30, 2013 @ 4:56 pm

    I have now skimmed Prof. Anderson's earlier Yale L.J. piece on the ADA (also linked to up above), which may or may not be easier for a non-specialist to follow (although it does have a Mark Liberman quote). My takeaway from that is that it is truly impressive for Congress to have written a statutory definition that may plausibly have nine different "literal" readings (depending on the various combinatorial possibilities for how multiple ambiguities are resolved), with four of them considered plausible in context by Prof. Anderson (I'm not sure I'm competent to judge the quality of her work opining which four to leave on the table and which five to take off). I suppose where I differ from Prof. Anderson is her notion that judges could and should (and she acknowledges this would require a culture change) both identify these sorts of ambiguities and then figure out which of the multiple possible readings to go with, rather than just throw up their hands and tell Congress to write a less ambiguous statute. If in fact she is correct that lawyers and judges are consistently not very competent at recognizing and resolving de re v. de dicto ambiguities, then legislators should be discouraged from writing statutes that require that mostly-aspirational skill set to be developed before they can be interpreted correctly, and a constructive sort of scholarship would be one that demonstrated how to rewrite this sort of ambiguous statute in ways that would unequivocally mandate various of the possible interpretations lurking in the current ambiguity.

    A separate issue with that article is that she writes not merely as a dispassionate scholar but as an advocate and activist – she is clear in her own sympathies as to what a good and just set of outcomes of the sort of cases the statute deals with ought to be. So that means she is good at coming up with examples where the reading adopted by the courts seems to lead to perverse results, but one would like to see a brief for the other side — i.e. some other advocate's attempt to demonstrate that Anderson's rival interpretation of the language would or could lead to equally perverse results in some different set of actual or hypothetical-example cases. (I should be clear that I don't fault her for the advocate/activist aspect of the piece, and appreciate her forthrightness about it; it is just that in complex areas of the law there is usually a decent case to be made for the other side of any given argument and often one can't and shouldn't expect a single writer to make both sides.)

  31. hbuutr said,

    March 18, 2014 @ 11:14 am

    A separate issue with that article is that she writes not merely as a dispassionate scholar but as an advocate and activist – she is clear in her own sympathies as to what a good and just set of outcomes of the sort of cases the statute deals with ought to be. So that means she is good at coming up with examples where the reading adopted by the courts seems to lead to perverse results, but one would like to see a brief for the other side

  32. hbuutr said,

    March 19, 2014 @ 8:09 pm

    coming up with examples where the reading adopted by the courts seems to lead to perverse results, but one would like to see a brief for the other side

  33. hbuutr said,

    March 21, 2014 @ 9:21 am

    I have now skimmed Prof. Anderson's earlier Yale L.J. piece on the ADA (also linked to up above), which may or may not be easier for a non-specialist to follow (although it does have a Mark Liberman quote). My takeaway from that is that it is truly impressive for Congress to have written a statutory definition that may plausibly have nine different "literal" readings (depending on the various combinatorial possibilities for how multiple ambiguities are resolved), with four of them considered plausible in context by Prof. Anderson (I'm not sure I'm competent to judge the quality of her work opining which four to leave on the table and which five to take off). I suppose where I differ from Prof. Anderson is her notion that judges could and should (and she acknowledges this would require a culture change) both identify these sorts of ambiguities and then figure out which of the multiple possible readings to go with, rather than just throw up their hands and tell Congress to write a less ambiguous statute. If in fact she is correct that lawyers and judges are consistently not very competent at recognizing and resolving de re v. de dicto ambiguities, then legislators should be discouraged from writing statutes that require that mostly-aspirational skill set to be developed before they can be interpreted correctly, and a constructive sort of scholarship would be one that demonstrated how to rewrite this sort of ambiguous statute in ways that would unequivocally mandate various of the possible interpretations lurking in the current ambiguity.

RSS feed for comments on this post · TrackBack URI

Leave a Comment