"Linguists have a name for this kind of analysis"

« previous post | next post »

Gordon Smith is enthusiastic about a recent opinion of the Utah Supreme Court, as he explains in "Corpus Linguistics in the Courts (Again)", The Conglomerate 8/14/2015:

Yes, yes, yes!

The point at issue is important and ubiquitous in legal argumentation, and his blog post explains the reasons for his (well justified) enthusiasm at least as well as I could. So go read it!


  1. “Corpus Linguistics in the Courts (Again)” | Internet Tax Lawyers said,

    August 15, 2015 @ 9:03 am

    […] "Corpus Linguistics in the Courts (Again)": Gordon Smith had this post yesterday at "The Conglomerate" (via "Language Log"). […]

  2. languagehat said,

    August 15, 2015 @ 9:18 am

    Great post, but he should really fix "Goggle News"; it makes him sound like Grandpa Simpson.

  3. Neal Goldfarb said,

    August 15, 2015 @ 9:23 am

    I've only skimmed the decision, but it looks fascinating in at least two respects.

    The first has to do with underlying issue of word meaning — when a statute makes it a crime to "discharge" a firearm, is each shot that is fired a separate offense, or is the entire episode of firing shots a single offense? The linguistic issues that this raises have to do with event structure and aktionsart, as well as (relatedly) what Len Talmy refers to as plexity and configurational structure.

    The second has to do with interpretive methodology, and the appropriateness of judges relying on the techniques of corpus linguistics. Only one of the justices (Justice Lee) relies on corpus analysis, and the majority is critical of his doing so. In response, result, Lee discusses the methodological issue at length. My initial impression is that he has the better of the argument, by far. But then, I'm biased.

  4. Neal Goldfarb said,

    August 15, 2015 @ 9:59 am

    Regarding the question of how to interpret "discharge" as used in the statute: The court held that each shot fired is a separate offense, and Judge Lee concurred in that conclusion. I think that's probably right, but here's something to consider.

    There is a federal statute that requires an increased sentence for anyone who is convicted of certain crimes if, "during and in relation to" that crime they used or carried a firearm. Normally the statute requires that the sentence be increased by at least 5 years, but "if the firearm is discharged" the sentence must be increased by at least 10 years. Suppose that in a case where this statute applies, the gun was fired more than once. Does that require an additional 10 years for each shot? Or is there just one 10-year increase for any given count on which the defendant is convicted?

    The 10-years-per-shot interpretation would be consistent with the Utah decision, but I'd be much less inclined to agree with that interpretation in this context. Does that call into question my view about how the Utah statute should be interpreted? Or does it instead show that meaning is very much dependent on context and that it can be hazardous to rely on an prior interpretation that arose in a different context? I think the latter.

  5. John said,

    August 15, 2015 @ 10:04 am

    "…dueling linguistics experts"

    I knew there was a way to make my degree sound cool.

  6. GH said,

    August 15, 2015 @ 10:04 am

    Perhaps the other justices' concerns about attempting to do original linguistic analyses as amateurs could be alleviated by having qualified linguists provide some guidelines, tips on suitable resources, and warnings about common mistakes in corpus linguistics, all in one readable pamphlet or website?

    Or perhaps such an "instruction manual" for laypeople already exists?

  7. John said,

    August 15, 2015 @ 10:08 am

    @Neal Goldfarb There doesn't need to be any conflict there. "If the firearm is discharged", not "For every time the firearm is discharged". Even if a "discharge" is interpreted as a discrete shot, you can still apply an if/else test of "Has the firearm been discharged (at least once), or has it not?".

  8. Neal Goldfarb said,

    August 15, 2015 @ 10:38 am

    John: While I agree with you about the conditional semantics of the federal statute, your analysis doesn't distinguish (or at least doesn't fully distinguish) that statute from the Utah statute.

    Just as the federal statute doesn't say "for every time a firearm is discharged", neither does the Utah statute say that. And although the Utah statute doesn't use an expressly conditional structure (it says "A person may not discharge", not "If a person discharges"), but in applying the statute in a particular case, the issue comes down to the same yes/no question: did the defendant discharge a firearm? In neither case does the statute specify whether the focus should be on the individual shots or on the overall incident.

  9. Jon said,

    August 15, 2015 @ 11:13 am

    The alternative interpretation, that to discharge is to unload, doesn't make sense (not that I wish to imply that laws necessarily make sense). If you fire a shot from a single shot firearm, you have unloaded it. If you fire 19 shot from a 20-shot magazine, you have not unloaded it. It seems unlikely that the law was intended to make the first an offence, but not the second.

  10. John Roth said,

    August 15, 2015 @ 11:21 am

    "Each shot fired" doesn't quite cover it, since there are firearms that shoot multiple rounds from a single press of the trigger. Granted that's a corner case. Would a firearm that's set to shoot three rounds for a single press of the trigger count as three discharges or one, since the shooter clearly only intended one three-shot burst.

  11. Gordon Smith said,

    August 15, 2015 @ 11:32 am

    Thanks, Mark.

    Justice Lee has been using corpus linguistics for some time, and it is interesting to see his colleagues on the Utah Supreme Court soften toward this approach. Chief Justice Durrant's concurrence expressed an openness that was simply not there the last time this issue arose. I think we are going to see a steady expansion of this sort of reasoning in judicial opinions.

    Thanks to languagehat for spotting the typo! It's fixed now.

  12. Guy said,

    August 15, 2015 @ 1:36 pm

    I haven't read the opinion yet, but the linked post says "The majority opinion also relied on a close reading of the statute, which refers to a firearm as 'any device . . . from which is expelled a projectile by action of an explosive,'" – with emphasis on the indefinite articles.

    The conclusion seems reasonable, but the above argument is so bad I would be embarrassed to make it even as a fallback.

  13. Guy said,

    August 15, 2015 @ 4:04 pm


    "Fire" versus "unload" is a red herring. It's a common feature in English that a verb referring to a discrete action can refer to a single performance of that action or multiple performances in succession, depending on context. "He knocked on the door" can mean one knock or several knocks, with several being pragmatically more likely. Also compare "John mowed the lawn" as an answer to 1) "what chore did John do yesterday?" and 2) "who was in charge of mowing the lawn when you were kids?", where the number of occurrences has to be inferred contextually. The number of occurrences is usually indicated explicitly or with aspectual marking when ambiguity needs to be avoided.

  14. ohwilleke said,

    August 17, 2015 @ 7:16 am

    Where the judge is taking judicial notice of facts not presented by the parties from Google News, the judge ought to at least give them notice of his intent to do so and allow them to offer briefing in relation to it.

    There are many linguistic corpi, and many ways to analyze them, and not all of them would necessarily lead to the same answer, yet judicial notice of facts not established by the parties is supposed to be limited to fact that are not reasonably subject to dispute. The Judge Lee may see his approach as modern and scientific, but the concerns raised by the other justices are genuine and appropriate.

  15. Bill W said,

    August 17, 2015 @ 2:42 pm

    While I'm all in favor of corpus linguistics for resolving legal issues of statutory construction, and I think it's a far better approach than reliance on a dictionary, I agree with ohwilleke. In this case, the defendant (i.e., the defendant's lawyer) never had an opportunity to challenge Justice Lee's methodology or his results. That seems unfair. The concerns expressed by the other justices were quite valid here.

RSS feed for comments on this post