The legal profession is afflicted by systemic problems, some urgent, some chronic. But here’s one you might not be aware of: it appears that many judges are ill-equipped to perform basic textual interpretation.
Such a generalization has to be grounded in specifics, so let’s look at a recent opinion issued by the Minnesota Court of Appeals, State v. Khalil, No. 27-CR-18-4880 (Minn. Ct. App. 27 Jul 2020) (here). After we consider it, I’ll return to my general point.
State v. Khalil involves an appeal by an appellant convicted of third-degree criminal sexual conduct. For our purposes the facts are irrelevant, as are the grounds for appeal, except for one: the appellant argued that the district court erred in instructing the jury that a complainant may be “mentally incapacitated” through voluntary consumption of alcohol or narcotics.
Here’s the statute at issue (Minn. Stat. § 609.341, subd. 7):
“Mentally incapacitated” means that a person under the influence of alcohol, a narcotic, anesthetic, or any other substance, administered to that person without the person’s agreement, lacks the judgment to give a reasoned consent to sexual contact or sexual penetration.
The question was whether “administered to that person without the person’s agreement” modifies just what immediately precedes it, namely “any other substance,” or whether it modifies every item in the list (“alcohol, a narcotic, anesthetic, or any other substance”). So we’re dealing with syntactic ambiguity, which arises principally out of the order in which words and phrases are used and how they relate to each other grammatically.
The latter interpretation, advanced by the appellant, would mean (in the words of the court) “a person must be administered alcohol, a narcotic, or anesthetic without that person’s agreement before that person may be considered mentally incapacitated.”
Canons of Construction
The district court judge chose the former interpretation. So did the Court of Appeals. It offered reasons based on semantics and the statutory context, but what’s of interest to us is the discussion of “canons of construction”—principles of interpretation courts rely on to resolve disputes over meaning.
The court says that the statute “potentially evokes competing grammar rules,” namely the rule of the last antecedent and the series-qualifier canon.
According to the court, quoting another case, the rule of the last antecedent “instructs that a limiting phrase ordinarily modifies only the noun or phrase that it immediately follows.” The court mentions what I call the comma test under the rule of the last antecedent—the principle that a comma before a closing modifier suggests that it modifies all preceding elements. The court held that presence of a comma wasn’t conclusive but that it did call into question application of the rule of the last antecedent. (By contrast, one of the Court of Appeals judges, concurring in part and dissenting in part, relied on the comma test to attribute one of two possible meanings to the statute.)
The court goes on to say that the series-qualifier canon applies when the natural construction of the language demands that the modifying clause be read as applicable to all. The court considers the implications of the any before “other substance,” quoting Antonin Scalia and Bryan Garner’s book Reading Law: The Interpretation of Legal Texts 149 (2012) to the effect that a determiner such as any tends to preclude the modifier from reaching further back up the sentence. The court decided that use of that any called into question application of the series-qualifier canon.
Ultimately, the court decided not to pick one canon over the other: “Overall, the weight of authority shows that grammatical rules are not absolute and are easily defeated by context.” It held on other grounds that the language at issue was subject to only one possible interpretation.
The court’s discussion of the competing canons of construction is dismaying.
The court misunderstands the function of canons of construction. They aren’t at all “rules of grammar.” Instead, canons of construction is a grandiloquent term for expediency-driven tiebreakers that courts invoke to resolves disputes based not on the often-elusive intention of the parties but on generalized and arbitrary notions of what the utterance in question might have meant.
So the court didn’t attempt to determine what a reasonable reading would be. Instead, it immediately entered the hall of mirrors that is canons of construction. Relying on canons of construction to resolve syntactic ambiguity is unpromising: if text is sufficiently confusing that people are willing to fight over what it means, it’s arbitrary to nevertheless use canons of construction to attribute a meaning to it.
That’s borne out when you look more closely at the canons of construction the court invokes. According to the most thoughtful article I’ve encountered on the subject, the rule of the last antecedent “has little weight or value.” Joseph Kimble, The Doctrine of the Last Antecedent, the Example in Barnhart, Why Both Are Weak, and How Textualism Postures, 16 Scribes J. Legal Writing 5 (2015) (here).
As for the comma test under the rule of the last antecedent, I cheerfully demolished a recent attempt to explicate it. See Kenneth A. Adams, Bamboozled by a Comma: The Second Circuit’s Misdiagnosis of Ambiguity in American International Group, Inc. v. Bank of America Corp., 16 The Scribes Journal of Legal Writing 45 (2014–15) (here).
Regarding the series-qualifier canon, I defer to Neal Goldfarb, who says, in this tweet, “The most serious problem is that the Series-Qualifier Canon is made up. There was no such canon previously.” According to Goldfarb, it was made up by Scalia and Garner in Reading Law: The Interpretation of Legal Texts. See his blog post on the subject here.
And the court’s attempt to attribute significance to the any before “other substance” is unconvincing. That’s not surprising, as the entire concept is unconvincing. In quoting Scalia and Garner on the effect of determiners before postpositive modifiers, the court omits their understated caveat, “but that effect is not entirely clear.”
The starting point for interpreting confusing language is seeing whether it’s possible to attribute reasonable meaning to it.
In the case of the Minnesota statute, the closing modifier suggests syntactic ambiguity. The only prospect for avoiding that ambiguity is the commas before and after “or any other substance”. They’re what I call “delimiting commas in coordination”; see this post. They suggest that “administered to that person without the person’s agreement” modifies not just what precedes it, namely “any other substance,” but also the other items in the list. (There’s no basis for invoking the comma test under the rule of the last antecedent and attributing significance to just the comma before “administered to that person without the person’s agreement”, as I demonstrate in my Bamboozled by a Comma article.)
Another fragment of evidence supports that interpretation. In “under the influence of alcohol, a narcotic, anesthetic, or any other substance,” you first have a mass noun (alcohol). Then comes a count noun (narcotic) with a determiner (a). The count noun that follows (anesthetic) shares the determiner used with narcotic. Normally the third count noun in the series (other substance) would also share the determiner, but saying under the influence of alcohol, a narcotic, anesthetic, or other substance would be awkward, because you’d expect an or after alcohol instead of a comma. Instead, whoever drafted the statute didn’t add that or and instead inserted any before other substance, disrupting the narcotic, anesthetic, and other substance grouping. That the drafters didn’t put the determiner an before anesthetic suggests that adding any might have been a careless later edit.
The only reason I can think of for this structure is that the drafters wanted the four nouns to constitute a series. And the only reason I can think of for that is that the drafters wanted “administered to that person without the person’s agreement” to modify the entire series.
So instead of the determiner any constituting evidence that the modifier modifies only the final item in the list, here it suggest that the drafters might have wanted it to modify all items in the list.
But given the awkward drafting of the statute, it’s hard to confidently attribute meaning based on such hints. As an additional example of that, consider the comma after “administered to that person without the person’s agreement”. According to notions of standard notions of comma usage, that comma is unnecessary. That suggests that whoever drafted the statute wasn’t in command of their commas.
Given that the evidence is too subtle to rule out syntactic ambiguity, I can’t fault the court for basing its decision on other grounds. But I do fault the court for mistaking canons of construction for rules of grammar, for being credulous in its discussion of the rule of the last antecedent (including the comma test) and the series-qualifier canon, and for not considering the readily accessible scholarship calling into question those canons of construction.
The Broader Picture
My subject is how to draft contracts clearly. One benefit of clearer contracts is that you’re less likely to end up in a dispute. I’m in the business of avoiding disputes, not resolving them.
But I keep half an eye on caselaw, because court opinions are a good source of cautionary tales—examples of bad drafting that landed the parties in court, examples of how courts misinterpret contract language. Mostly I learn about interesting court opinions from reader tips (that’s how I learned about the Khalil opinion) and news items. That doesn’t happen terribly often.
Considering only the kind of ambiguity at stake in Khalil, in the past year I’ve written five blog posts about syntactic ambiguity (including this one). Of the other four, this post, this post, and this post describe how a court botched its analysis of syntactic ambiguity. By contrast, this post highlights the syntactic ambiguity at issue and otherwise has nothing to say about the court’s analysis.
So four out of my five blog posts on syntactic ambiguity in the past year describe major shortcomings in a court’s analysis of syntactic ambiguity. I selected those court opinions because they dealt with syntactic ambiguity and not based on whether the court’s analysis was sensible, so my blog posts are evidence—albeit anecdotal evidence of the crudest sort—that courts are prone to misunderstanding syntactic ambiguity.
Another kind of structural ambiguity that I’ve written about extensively is the ambiguity associated with plural nouns and the words and, or, each, any, and all. (I call this kind of ambiguity “the ambiguity of the part versus the whole.”) It too is reliably mishandled by courts; in the past year, I’ve written two blog posts about that, here and here. I haven’t written any other blog posts this year about caselaw relating to this kind of ambiguity.
To generalize even more broadly, after 20 years of writing about such cases I’ve come to expect that courts are more likely than not to mess up analysis of either kind of ambiguity.
Instances of syntactic ambiguity and ambiguity of the part versus the whole can be identified and assessed independent of the broader meaning of the text in question, so it’s relatively easy to spot when courts go wrong. As such, those kinds of ambiguity might be a sign of a broader problem—if courts routinely mishandle them, it might be unrealistic to expect courts to be more competent in handling other aspects of textual interpretation.
Training Judges and Their Clerks
How to make judicial interpretation of legal texts less erratic? Relevant commentary exists. For one thing, my book A Manual of Style for Contract Drafting contains what is, to my knowledge, the most thorough examination of syntactic ambiguity and ambiguity of the part versus the whole available in any literature.
But it’s not realistic to expect judges and their clerks to find their way to, and read, relevant materials. One could offer training, but that too offers unpromising options. In particular, a standard continuing-legal-education presentation is unlikely to get anyone far enough along the learning curve to make a difference.
The training I offer in contract drafting might offer a model to emulate. Recently I launched a course called Drafting Clearer Contracts: Masterclass. It’s built around eight live, hour-long online sessions held once a week and supplemented by reading, quizzes, and short assignments. It’s intended for small groups—up to around 20. All materials are available in a convenient online curriculum. It’s very much a work in progress, but the combination of self-study and online engagement should help participants become more familiar with the issues, and more quickly, than they would otherwise. This format might help those involved with interpreting legal texts gain a better grasp of the process.
Ideally any training would receive enough institutional support to allow it to be offered widely—even internationally—to any court that’s interested.
But we shouldn’t expect judges and their clerks to morph into linguists. Perhaps we should be satisfied if they learn enough to recognize when they’re out of their depth.
Judges could acknowledge reality by admitting expert testimony on ambiguity. As things stand, it’s generally inadmissible—courts in the United States consider resolving ambiguity to be a matter of law. (See this 2009 post.) There’s no sign of that changing. For one thing, admitting expert testimony on ambiguity would likely further bog down proceedings.
Training Litigators Too
Another constituency could step into the breach—litigators. The next best thing to expert testimony on ambiguity would be for litigators to incorporate expert analysis into their pleadings.
But for that to work, litigators would need semantic acuity, instead of doing what many lawyers habitually do when wrangling prose—overestimate their own facility and rely on pseudoscholarship and misbegotten conventional wisdom.
So litigators too would benefit from training in textual interpretation.