Thursday, October 22, 2015

The Government Jumps off a Cliff in Lockhart v. U.S., and Why All Textualism is (Constructive) Intentionalism

Lockhart v. United States, set for argument on November 3, is a case about two dueling canons - the ancient rule of the last antecedent, and the newly discovered "series-qualifier" rule, which made its first appearance by name in Justice Scalia's 2012 book, Reading Law.  A recidivist sentencing provision in a federal child pornography statute provides that, if an offender has been convicted of one of a number of federal sex crimes, 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, or sex trafficking of children," he is subject to a ten-year mandatory minimum.  

The defendant in Lockhart argues that "involving a minor or ward" modifies the entire series of generic offenses that precede that phrase.  The government argues it only modifies the last antecedent - "abusive sexual conduct."  (Antecedent is a misnomer here, strictly speaking; a noun is not an "antecedent" of an adjective that modifies it.)  Five circuits, including the Second Circuit below in an opinion written by Judge Katzmann, agree with the government; the Eighth Circuit, in a couple of opinions, has assumed Lockhart's interpretation, although never in an outcome-determinative way.  A cert grant on a defendant's petition from such a weak circuit split strongly suggests, I think, some doubt about the government's position, which the government's briefing in Lockhart, I'm afraid, likely has unnecessarily compounded.

Currently, the state of the art on statutory interpretation of lists followed by a modifier - A, B, and C that XYZ - is more of a muddle than ever before, in a good way.  Grammatically, "A B and C that XYZ" is ambiguous as to whether XYZ modifies A, B and C or just C.  However, until fairly recently, courts usually resolved such ambiguities with the rule of the last antecedent, which links the modifying XYZ to only the last potentially modified word/phrase, C.  The rule of the last antecedent is not a rule of grammar - again, sentences of this kind are grammatically ambiguous - but it does turn out to be a decent guess about usage, as research shows that English speakers have a tendency towards "low attachment" -- i.e. reading modifiers, in cases of ambiguity, to modify (in linguistics-speak, attach to) less rather than more.  That said, it's only a good guess some of the time, as the very able federal public defenders who represent Lockhart explain in their brief.  To give a couple of their examples, a sign at a park that permits swimming, boating, or fishing before sunset only doesn't permit swimming and boating all day; a menu offering a special of a cheeseburger, hamburger, or turkey burger with fries probably doesn't mean that only the turkey burger comes with the fries.

Though the last antecedent rule is a woefully coarse estimate of drafters' intent, congressional understanding, or public expectations, courts could have left well enough alone and hoped that Congress would legislate against it, as a misguided but predictable background rule.  Background-rule defenses of canons, however, assume that Congress knows about them and cares enough to draft against them, which recent empirical research calls into question.  So in his recent book, Justice Scalia pioneered a counter-canon to the last antecedent rule, the series-qualifier rule, which essentially says that where a modifier seems just as applicable to the first words in a series as to the last, it modifies the whole series.  Justice Scalia didn't invent the series-qualifier rule, though he gave it its name; the rule makes an appearance in a 1920 Justice Brandeis opinion (see page 348), and in at least one later opinion of the Court.  But it was largely forgotten until it resurfaced in Reading Law, at which point it came to play an important role in a circuit split over restitution to child pornography victims, and in the Court's ultimate resolution of that split a year ago in Paroline.  We are now, then, in a spot where courts, under the guise of applying two canons, really simply ask what Congress likely or apparently intended, with a thumb on the scale in favor of low attachment. This seems exactly right, although it can leave courts groping for answers.

The lesson observers should be drawing from the collapse of the rule of the last antecedent (in the series-and-modifier context) is that canons, to say the least, aren't gospel.  But predictably, what we instead see happening in Lockhart is courts and practitioners attempting to arbitrate a dispute between two countercanons with more canons.  Specifically, the government and multiple circuits have reasoned that in the child pornography sentencing provision, the rule of the last antecedent must prevail because the series-qualifier rule would generate an interpretation that falls afoul of the most powerful linguistic canon of all, the rule against surplusage.*  Unfortunately for the government, Lockhart, in a stunning bit of jujitsu, conceded partial surplusage, leading the government to bite on his concession and attack it in such a way as to propose a surplus construction of their own that's completely absurd.

* Anti-surplusage maintains its power despite recent empirical research showing that legislative drafters are well aware of the rule and don't buy into it.

How does the surplusage argument work in Lockhart?  There's a good surplusage argument for the government and a suicidal one.  The good argument, made by the Sixth Circuit in an unusual 14-1 per curiam en banc, goes like this.  The series in Lockhart is "aggravated sexual abuse, sexual abuse, or abusive sexual conduct involving a minor or ward."  Aggravated sexual abuse is clearly a subset of sexual abuse; suppose, then, that sexual abuse is a subset of "abusive sexual conduct" (though this is not at all clear).  If "involving a minor or ward" modifies "aggravated sexual abuse" and "sexual abuse," not just "abusive sexual conduct," those phrases becomes pointless, because the inclusion of abusive sexual conduct involving a minor or ward necessarily includes sexual abuse/aggravated sexual abuse involving a minor or ward.

That's an okay argument if you buy its premises, although even the government has to concede that "aggravated sexual abuse" is surplusage on its reading; the statute includes all sexual abuse, they say, so it necessarily includes all aggravated sexual abuse.  The government's explanation of that bit of surplusage is rather weak - they say it serves the "important purpose" of clarifying that the enhancement applied to "particularly serious types of 'sexual abuse' offenses," not just to . . . all sexual abuse offenses - and once they give away surplusage as to the first item on the list, it opens the door to the possibility that the entire list is a surplusage stack: A which is a subset of B, B which is a subset of C, and C.  Nevertheless, it's not a terrible argument.  One could very plausibly read the statute to say that abuse and aggravated abuse as to all victims enhances sentences, while a less serious kind of offense - "abusive sexual conduct" - only enhances sentences if children are victims.  That is, so long as one saw a way to distinguish sexual abuse from abusive sexual conduct, which isn't easy.

Unfortunately, that's not the argument the government makes, because of a (ingenuous?) concession Lockhart made.  Lockhart acknowledged in his opening brief that sexual abuse and abusive sexual conduct are just about the same thing, and argued that it therefore makes sense for "involving a minor or ward" to modify both. (This submission has particular force because the phrase "relating to" precedes the series and blurs whatever fine boundaries there may be between sexual abuse and abusive sexual conduct.)  On Lockhart's view, then, abusive sexual conduct involving a minor or ward doesn't add anything beyond what sexual abuse involving a minor or ward adds.  The government, transfixed by that concession of surplusage, accepts Lockhart's concession that sexual abuse and abusive sexual conduct mean the same thing, says he's right, and then argues that given the equivalence, Lockhart's interpretation creates impermissible surplusage.  As Lockhart reads the statute, the government says, the statute means: "synonym A involving a minor or ward, or synonym B involving a minor or ward," where synonym A and B are synonymous.  What's the point of including synonym A?  But as the government reads the statute, it means "synonym A, or synonym B involving a minor or ward."  Now there's a point to synonym A!  It serves to show that only synonym B requires the involvement of a minor or ward.

Of course, as Lockhart argues in his reply brief, this is a comically self-defeating argument.  If the government accepts that "sexual abuse" and "abusive sexual conduct" are synonyms, and argues that "involving a minor or ward" modifies only the latter, they've painted themselves into a corner where involving a minor or ward is meaningless because the government can always get an enhancement through the unmodified synonym.  And while it's possible enough for a skeptic of the rule against surplusage (like me) to read the series as a stack of subsets, where abusive sexual conduct carries the weight of the whole provision, no one could read the minor/ward limitation to effectively limit nothing because the phrase it modifies is overridden by an unmodified synonym of that phrase.  The government could only have avoided this conundrum by rejecting Lockhart's concession and arguing that sexual abuse is a subset of abusive sexual conduct, in which case "involving a minor or ward" would limit a set of offenses not completely captured by the other items in the series.

The government weakly anticipated this objection, arguing that adding "abusive sexual conduct involving a minor or ward" clarified that statutory rape offenses were priors for the sentencing enhancement - I suppose on the theory that statutory rape isn't clearly "sexual abuse," but is clearly "abusive sexual conduct involving a minor or ward."  This argument only underscores that the government never should have conceded that "abusive sexual conduct" and "sexual abuse" mean the same thing, as it only works if they at least possibly don't.  In any case, the argument doesn't help, because it only begs the question of why Congress bothered to write "involving a minor or ward" or "sexual abuse."  Had Congress just written abusive sexual conduct with no modifier, it would be just as clear that statutory rape was covered; adding "involving a minor or ward" doesn't make it clearer, as statutory rape still has to be abusive sexual conduct in the first place, which it obviously is.
I don't know if the government can dig out of this hole at oral argument, or if a majority of the Court may find that the government's bottom line has enough appeal that they will ignore the government's concession and read abusive sexual conduct to do some work that sexual abuse doesn't.  I suspect the government is up a creek, and not just for the reasons rehearsed above.  (Note, for example, that every one of the state-offense enhancers following the series in question involves children.)  However easy a case Lockhart may ultimately be, there are two important lessons we can draw from Lockhart

The first is that the Court and the practitioners that know it best have become far too obsessed with surplusage as a universal solvent of ambiguity.  Lockhart is a case where everyone knows the snippet of text in question is partially redundant; aggravated sexual abuse adds absolutely nothing that sexual abuse doesn't cover.  But even though the first item in the series here is indisputably redundant on the second, when Lockhart had the temerity to argue that the second item was in turn redundant on the third and vice versa, the government could not help but pounce on this concession and paint it as an argument for impermissible surplusage, even though in doing so they blew up their own argument.  (Likewise, the circuits that made a better surplusage argument were willing to accept that the first term in the series was superfluous, but not the second.)  

This isn't surprising; incredulity at pretty conceivable instances of intentional superfluity has been a commonplace in the Court's recent opinions.*  This was inevitably going to trickle down into repeat players' briefs, creating a surplusage-obsessed feedback loop.  But it's rather misguided.  To the extent what we're doing, when we interpret statutes, is trying to discern the intentions of Congress, or the understandings of a hypothetical Congressman who voted on a bill, the evidence is that Congress doesn't mind minor redundancies.  To the extent courts seek to ascertain public expectations of how courts would construe statutes - a somewhat circular but appealing theory, in my view - what evidence is there that the public believes Congress drafts with perfect parsimony?

*  A personal favorite - the Bankruptcy Code exempts "retirement funds to the extent those funds are in a fund of account that is exempt from taxation under [certain sections of the Internal Revenue Code]."  An IRA is exempt under one of the listed sections.  So an inherited IRA might seem to be exempt in bankruptcy.  The Court, however, unanimously held last year that an inherited IRA is not exempt because it's not a "retirement fund"; the person who inherits it may, but doesn't have to, keep the money for retirement.  Responding to an argument that every account exempted by the listed sections of the Internal Revenue Code is a retirement fund, the Court said that couldn't be because it would make the retirement-funds "element" of the bankruptcy exemption superfluous.  Of course, people often say "A, modified by B," where B could only sensibly modify A.  If I were to speak of "charitable contributions exempt from taxation under [the sections of the Internal Revenue Code that specifically and only exempt charitable contributions]," have I referred superfluously to charitable contributions?  Isn't "charitable contributions," at least, grammatically non-superfluous because one needs a noun for "exempt from taxation . . ." to modify?  Should an interpreter of my remark attempt to save it from surplusage by proposing some restrictive meaning of charitable contributions?

The second lesson here (which will come as a mere reminder to many) is that textualism is a thoroughgoingly intentionalist enterprise, or at least, a constructively intentionalist enterprise.  Recall that Lockhart is made possible by the textualist Justice par excellence pioneering a counter-canon to the rule of the last antecedent, on which one rejects the rule of the last antecedent where a modifier seems like it ought to modify everything.  This is a concession that neither grammar nor faux-grammatical maxims are up to the task of deciding what modifiers modify, and that what courts should do with modifiers is attempt to decide what they were intended (or look like they were intended) to modify.  

In arbitrating the seemingly hypertextualist dispute that then remains between two canons, which are really two self-canceling rules that jointly point to intent, the parties in Lockhart and the courts below turn to an array of intentionalist arguments.  One of these is surplusage itself, which is really just an empirical claim that Congress doesn't write superfluous words and thus must intend meanings that don't make its words superfluous. 

Another type of argument both parties and the courts below make is intentionalist inferences from context. Lockhart points out that all the other state-offense enhancers in the statute involve children, so Congress must have only wanted state-offense enhancers that involve children, perhaps because it wasn't too sure what sexual abuse offenses involving adults look like in the states.  The government and the Second Circuit counter that the federal-offense enhancers in the statute include sexual abuse of adults and that it would be odd if Congress enacted an enhancement for federal sexual abuse of anyone but only state sexual abuse of children.  Instead, what Congress must have been thinking when it wrote "involving a minor or ward" was making sure they'd cover statutory rape, or some  sub-sexual-abuse crime that victimizes children.  There's more, but the common thread is that in each argument, the parties take some part of the statute and argue that because of it, Congress probably intended the modifier to modify/not modify the whole series.  None of these arguments depend on legislative history, and all of them are in play for the most dyed-in-the-wool textualist.  Nevertheless, they're all claims about intent - not the semantic meaning of what Congress said, which is utterly unaffected by surplusage or the other enhancing offenses in the statute, and which is irreducibly ambiguous.

Now, a textualist would say that he's interested in all these arguments not because he's interested in Congress's intentions, which he either may deny exist or deny are law.  Rather, he will say, he's interested in these arguments because he's interested in what it seems like Congress intended - its "objective intent," as John Manning would say.  After all, if we only cared about semantics, the law would be radically underdetermined, as statutes can have multiple semantic meanings which inferences from context can clear up.  It's a bit of a mystery to me why textualists think the way to disambiguate semantically ambiguous statutes is making inferences about objective intent that neglect evidence of subjective intent (i.e., legislative history).  One possible reason, although I haven't seen textualists argue it, is that because the public inevitably attempts to figure out what Congress "meant" when they read statutes, courts must give them what it looks like Congress meant so as to give them law that fairly tracks public expectations.  But whatever the reasons textualists have for reasoning just like intentionalists, the space between textualists and intentionalists (besides textualists' refusal to look at legislative history) has been vastly overstated.


  1. Abusive sexual conduct would include not just acts of sexual abuse, but any act that would condone the engaging in or affirmation of a sexual act that abuses the intention of the sexual act and would thus make one complicit in promoting abusive sexual conduct. Not all abusive sexual conduct is illegal. While the desire to engage in a demeaning sexual act does not change the nature of the act, the use of force, coercion, or manipulation is always a criminal offense.
    While it is possible for an adult to consent to engage in a form of abusive sexual conduct, a child does not have the ability to consent, thus every act of abusive sexual conduct that involves a child, is sexual abuse that involves the use of force, coercion or manipulation, with statutory rape being the most heinous.

  2. Hi Mr. Steinberg,

    I'm teaching Lockhart in my 1L statutory interpretation course, and your post is so well-written and substantively thoughtful that I plan to use it as optional reading. Here's a question: Do you have any theories about why the SCOTUS and many other lower courts dig in on anti-redundancy norms, when there is plenty of evidence that lawmakers will never cease and desist from surplusage in their rules and regulations? Is this judicial itch to reduce redundancy a flawed effort to capture likely legislative intent? Is it driven by a desire to discipline Congress into writing more precise statutes with a sort of penalty default rule? (If so, as Abbe Gluck and Victoria Nourse note, Copngress is not getting the message). Or is the canon popular simply because it seems easy to apply and yields definitive results? Its durability is a bit of a mystery.

    Rick Hills (NYU Law Prof)

    1. Dear Professor Hills,

      Thanks so much for those kind words. It's an honor to have my post assigned in your class. I think that very few judges, especially lower-court judges, have any aspirations (at least conscious ones) to discipline Congress into writing more precise statutes. I think your two other ideas jointly explain the canon's popularity. Courts and practitioners don't know much about Congress, and I think many sincerely imagine that Congress doesn't use words superfluously. Ease of application and definitive results also work in anti-surplusage's favor. There's also a kind of pseudo-respectful quality about anti-surplusage; we pretend to believe that Congress doesn't carelessly draft and that all of its words have meaning. It's of a piece with all sorts of anti-realist judicial etiquette that you see in opinions; for example, the Court's rhetoric about the states, which you would know much more about than me. Finally, what is there to cause courts to stop using anti-surplusage? The Gluck/Bressman study says it's an inaccurate measure of intent, but my sense is that courts would rather ignore that information, as they don't know what to do with it. Gluck and Bressman don't say that all superfluities are intentional, just that some are, and courts have little way of knowing which are which.