Lockhart: The Canon of Reading Statutes to Make Sense

This is a post about canons, and why, although lawyers should know them, they should not really rely on them to interpret text.

In a case to be argued Tuesday, Lockhart v. United States, the question presented is basically this.  When a statute says, regarding a mandatory minimum provision in a child pornography statute:

....or under the laws of any State relating to aggravated sexual abuse, sexual abuse, or abusive sexual conduct involving a minor or ward, or the production, possession, receipt, mailing, sale, distribution, shipment, or transportation of child pornography....

are the three mandatory categories "aggravated sexual abuse," "sexual abuse," or "abusive sexual conduct involving a minor or ward," or aggravated sexual abuse (involving a minor or ward)," "sexual abuse (involving a minor or ward)" or "abusive sexual conduct involving (a minor or ward)?"  I think the answer is the latter, and I will explain why.  But first I want to take a shot at the canons themselves, as most recently celebrated by Bryan Garner and Antonin Scalia in their book Reading Law: The Interpretation of Legal Texts.

In their book, Scalia and Garner argue that statutory interpretation can always be gleamed from the text (which I generally dispute but am not taking issue with here, exactly).  More specifically, they point to 57(!) canons of interpretation which can be used to discern the meaning of any statute, and posit that even when the Court gets to the right result in interpreting statutes by looking outside of the text, the canons would produce the same result.

Judge Posner has already done a remarkable job eviscerating this book.  But I want to make a somewhat different criticism of canons.  

While the canons appear to capture how we actually interpret law (really, how we interpret language), they aren't prescriptive, but descriptive.  That is, they are just an attempt to try and break down how we actually use language into a series of rules.  And that this is ultimately descriptive rather the prescriptive makes sense, because language itself is descriptive rather then prescriptive.  Or, in plain English, language is how people use and understand it, not a set of unbending rules.  Ain't isn't a word, until people use it and understand what it means.  And then it is!

And so using a set of descriptive rules that approximate how we understand language, in place of how we actually use language, is never going to be as good as just actually trying to understand the language directly.  Tests are sometimes necessary, sure.  But they should never prevail over the real thing, when they fail to accomplish their intent.  And Posner hits on this, if indirectly, in his criticism:  if the meaning of something is obvious, like a "No dogs, cats, and other animals allowed," sign in a shop that would obviously include bears, a a canon that says "no, it must be an animal like a dog or a cat, i.e., a pet, to be prohibited," should be avoided.*

In other words, in easy cases, you don't need the canons, and occasionally the canons will lead you astray.  And in hard cases, the canons are of no help, either.  This is for two reasons, both of which I think Lockhart illustrates.

The first reason is that some of the canons are wrong.  That is, they are not descriptive of how people write or read at all.  The most famously wrong canon is the canon against surplusage.  This canon is probably the most cited of all the canons!  It holds that we must give each word, phrase, or clause independent meaning.  But, as the Supreme Court reminded us in King v. Burwell, the “preference for avoiding surplusage constructions is not absolute."  And the reason for this is because it is wrong as a descriptive manner:  People repeat themselves all the time.  Even people, like lawyers  (especially lawyers!), who draft legislation knowing the rules upon which people interpret them fail to live up to these rules.  Lockhart is a good example.  Regardless of whether "minor or ward" modifies all three terms or just "abusive sexual conduct," the first two terms are " "aggravated sexual abuse" and "sexual abuse."  But any "aggravated sexual abuse" will also be "sexual abuse," rendering it surplusage.  

The other reason is that hard cases are hard for a reason.  The fifty-seven canons describe how we interpret language, and in cases where it is difficult to interpret language, the canons don't help, usually because they contradict.  As two other blogs posting on the subject have already well-documented, the two canons at issue here, contradicting themselves, are the "series qualifier" canon and the "last antecedent" qualifier.  And, in fact, there's nothing special about these canons contradicting themselves, as they are almost entirely opposite rules.  (This is a famous critique.)  Many textualists argue that, when they contradict each other, you can use context and other interpretive guides to figure out the meaning.  But what good is that advice?  In easy cases, you don't need the canons.  In hard cases, they're not helpful.**  The canons should be left for law reviews and linguistics classes.  They provide no real help to drafters and judges who are trying to do the serious work of interpreting statutes based either on their intent or their commonly-understood meaning.  (I am not wading into that debate here.)  

So let's get rid of them.  Once we get rid of them, Lockhart is an easy case.  The statute at issue is a child pornography statute; the code section title is "Certain activities relating to material involving the sexual exploitation of minors."  The reasonable interpretation is that the enhancement should relate to minors and wards.  For this reason, the better interpretation is to include the qualifier for all three listed crimes.

I will discuss another case about statutory interpretation, canons, and a holistic make-sense-of-the-statute in another post.  That case, Torres v. Lynch, it turns out, is much closer.  And again, canons do not help.

*This reminds me of a case I recently litigated in which the other side argued that something should be classified as a regulatory taking instead of a per se taking.  (If you don't know or care about takings law, feel free to ignore this footnote - it isn't necessary for understanding the post.)  The other side argued that, sure, it was *just like* a taking substantively, but because they used a different process, it should be treated as a regulatory taking, and then proceeded to argue it was not a taking under the regulatory takings test.  But the regulatory takings test is specifically designed to determine when a regulation is substantively the same as a taking, so there was no point in actually going through the test.

** People also refer to the "rule of lenity" as a canon.  The rule of lenity says we should interpret criminal statutes narrowly.  While substantively it has the effect of canon, it is less about how we interpret language and more about fairness.  Fairness dictates that people should be aware of the acts which will subject them to criminal consequences, and so we should not interpret statutes in a way that people might not have understood, because that's not fair.  (There are criticisms of the rule, such as the factually correct criticism that most people don't actually read these laws and so the basis does not make sense, and the more philosophical question of whether the criminal code is a guide, or instead is the real guide "don't be bad," with the code simply a way of breaking that down bureaucratically.  I won't go into these criticisms except to note that this debate pales to the problem of overcriminalization.)