I got another tip from a participant at the recent “Drafting Clearer Contracts” seminar in Houston, although I didn’t have the wit to note his name.
He suggested that I check out the 2004 opinion in SuperGuide Corp. v. DirecTV Enterprises, Inc., by then Circuit Judge (now Chief Circuit Judge) Sharon Prost of the United States Court of Appeals for the Federal Circuit; for a PDF, go here. (The Federal Circuit is the only appellate-level court with the jurisdiction to hear appeals in patent cases.) This opinion involves the semantics of at least one of X and Y, which is something I hadn’t previously had occasion to look into.
A Patent Case?
But SuperGuide is a patent case! Patent claims aren’t contracts, so why am I bothering to look at it? Consider what the SuperGuide court said:
There is a “heavy presumption” that the terms used in claims “mean what they say and have the ordinary meaning that would be attributed to those words by persons skilled in the relevant art.”
So the SuperGuide court considered the “ordinary meaning” of at least one of X and Y. (The business about “would be attributed to those words by persons skilled in the relevant art” is irrelevant in this context, as we’re dealing with everyday English and not technical or scientific terminology.) It wouldn’t be surprising if a court considering the same issue in the context of a dispute over contract interpretation were to consider what SuperGuide has to say. That makes SuperGuide potentially relevant for purposes of contracts.
The Court’s Analysis
Let’s consider the opinion. I won’t summarize the facts, as that would take too much of your time and mine. Instead, I’ll focus on the semantics.
Here’s part of the language at issue:
An online television program schedule system comprising:
first means for storing at least one of a desired program start time, a desired program end time, a desired program service, and a desired program type; …
Does this refer to storing at least one unit within each of the four categories, or storing at least one category?
Here’s what the court said (footnotes and citations omitted):
In interpreting this phrase, the district court concluded that the term “a desired,” which precedes “at least one of,” is repeated for each category and because the final category in the criteria list is introduced by “and a desired,” the list is conjunctive. The court also concluded that accepting SuperGuide’s position that “at least one of” refers only to one category of the criteria would contradict the purpose of the invention as described in the written description, as depicted in Figure 4a and recited in claim 1. Thus, the court construed the phrase “at least one of … and” as meaning “at least one of each desired criterion; that is, at least one of a desired program start time, a desired program end time, a desired program service and a desired program type. The phrase does not mean one or more of the desired criteria but at a minimum one category thereof.”
On appeal, SuperGuide contends that the claim phrase “at least one of” unambiguously requires the selection and storage of one or more of the four listed criteria (start time, end time, service channel, or type) and does not require storing all four criteria. …
DirecTV counters that the district court’s construction is supported by the patentee’s use of the conjunctive word “and” and by the grammatical rule requiring that the phrase “at least one of” be applied to each category in the list. …
We agree with DirecTV. The phrase “at least one of” precedes a series of categories of criteria, and the patentee used the term “and” to separate the categories of criteria, which connotes a conjunctive list. A common treatise on grammar teaches that “an article of [sic] a preposition applying to all the members of the series must either be used only before the first term or else be repeated before each term.” William Strunk, Jr. & E.B. White, The Elements of Style 27 (4th ed. 2000). Thus, “[i]n spring, summer, or winter” means “in spring, in summer, or in winter.” Applying this grammatical principle here, the phrase “at least one of” modifies each member of the list, i.e., each category in the list. Therefore, the district court correctly interpreted this phrase as requiring that the user select at least one value for each category; that is, at least one of a desired program start time, a desired program end time, a desired program service, and a desired program type.
The court also suggested that the outcome would have been different if the language at issue had used as a conjunction or instead of and. Specifically, in footnote 10 the court said, “SuperGuide does not articulate its argument that ‘at least one of’ means ‘one or more of the four listed criteria’ without using the term ‘or’ to separate the four listed categories.” It then noted that an earlier Federal Circuit decision had interpreted a claim constructed as “at least one of A, B, or C” as covering only A, only B, only C, or any combination of the three.
The court’s analysis makes no sense. First, the passage in The Elements of Style is irrelevant. All it does is say that as a matter of writing style it would be awkward to say “in spring, summer, or in winter” instead of one of the two alternatives the court quoted. It has nothing to do with identifying ambiguity.
And second, of course the phrase “at least one of” modifies each member of the list. That’s not the issue. Instead, the issue is whether “one of” refers to an item within a category or to the category itself.
So the “Therefore” that begins the final sentence of the extract above is unearned—nothing has been demonstrated.
The meaning suggested by SuperGuide is the more natural reading, but I could imagine a not-particularly-skilled drafter wording it that way even if they had intended to express the DirecTV meaning. So I could imagine a court holding that the language at issue conveys the SuperGuide meaning or holding that it’s ambiguous. The one conclusion that’s implausible is the one the court arrived at—that the language at issue conveys unambiguously the meaning sought by DirecTV.
And there’s the court’s suggestion that the outcome would have been different if the language at issue had used or instead of and. That too makes no sense. For one thing, as described above, the possible alternative meanings of the language at issue don’t arise from use of and.
Furthermore, and is actually the logical choice. Let’s explore how this plays out, first by considering use of at least one of X and Y with units, a use that wasn’t at issue in SuperGuide:
[1a] at least one of Tom, Dick, and Harry
[1b] at least one of Tom, Dick, or Harry*
The reference is to at least one of a group, making and the appropriate choice, as in [1a]. Using or instead would suggest that one is referring to at least one out of one of the three. That doesn’t make sense. Option 1[b] is illogical—hence the asterisk.
Now let’s consider use of at least one of X and Y with categories—in other words language analogous to that at issue in SuperGuide, but written to make it clear whether what is being referred to is the categories or units in the categories.
Here are examples that refer to the categories (analogous to the meaning sought by SuperGuide):
[2a] at least one of the following categories: nuts, bolts, and screws
[2b] at least one of the following categories: nuts, bolts, or screws*
Again, the example using and is the logical choice. The example using or is the illogical choice, as it suggests that one is referring to at least one out of one of the three categories. It follows that the SuperGuide court is doubly mistaken in suggesting that SuperGuide could have expressed the intended meaning only by using or.
Now let’s consider examples that refer to units in the categories:
[3a] at least one unit in each of the following categories: nuts, bolts, and screws
[3b] at least one unit in each of the following categories: nuts, bolts, or screws*
[3c] at least one unit in the category nuts, the category bolts, or the category screws*
[3d] at least one unit in [at least] one of the following categories: nuts, bolts, and screws
Example [3a] is analogous to the meaning sought by DirectTV. Here, too, using or, as in [3b], is the less defensible choice, and for the same reason that [2b] is less logical than [2a]. You could use or to convey a different meaning, as in [3c], but it would be clearer to use and to convey that same meaning, as in [3d].
One commentator on English usage has in the past offered an analysis that conflicts with my analysis of [1a] and [1b] (and, by extension, the subsequent examples). In the second edition of Bryan Garner’s A Dictionary of Modern Legal Usage, the entry for later of [date] or [date]; later of [date] and [date]—a construct analogous to at least one of X and Y—says as follows:
Drafters frequently debate whether the proper conjunction in this phrase is or or and. The better idiomatic choice is or—nine of every ten lawyers believing it is the proper choice.
True, and has logic on its side. If we paraphrase by saying the later of two dates, it becomes clear that the sense must be plural (conjunctive and), not singular (disjunctive or). But the wording with and sounds as pedantic—and as wrong—as a number of people was there.
This analysis exhibits three shortcomings. First, recommending usages based on their popularity sounds alarmingly descriptive coming from the essentially prescriptive Garner. Second, to my ear using and sounds just fine, and my ear has served me well over the years, but more to the point, it’s a cop-out to invoke the “sound” of a usage and nothing more. And third, the comparison with a number of people was there is unhelpful, like any comparison of apples and oranges. So I’ll go with the logical alternative, and. It’s perhaps not surprising this entry didn’t make it into the third edition of what is now called Garner’s Dictionary of Legal Usage. [Correction 7 February 2017: Actually, it’s in the third edition.]
The court’s opinion in SuperGuide is unfortunate in two respects. For one thing, with respect to this part of the opinion at least, SuperGuide was deprived of justice. But more generally, this opinion has set an unhelpful precedent by suggesting that the outcome would have been different if the language at issue had used as a conjunction or instead of and. In fact, the seminar participant who told me about SuperGuide mentioned that he and his colleagues assume one should always use or with at least one of.
Here’s what I recommend: Use only at least one of X and Y. Don’t use or. To avoid problems, pay attention to how you use this phrase. If you’re referring to units, as in [1a], there’s nothing to consider—the meaning is clear and furthermore SuperGuide doesn’t apply. If you use it regarding categories, make it explicit that you’re referring to categories, as in [2a]. If you use it regarding units within categories, make it explicit that you’re referring to units, and furthermore make it clear whether you’re referring to units in all the categories, as in [3a], or fewer than all, as in [3d].
And by the way, you can convey the same meaning as at least one of X and Y by saying one or more of X and Y. The same analysis applies to both formulations. I prefer the latter, just because it seems a bit simpler.
As I explain in this 2009 post, courts generally hold that (1) no expert testimony of any kind is admissible to determine whether contract language is ambiguous and (2) in the case of contract language that isn’t ambiguous, expert testimony is admissible only to determine technical meaning.
We see how that worked out in SuperGuide. And although the semantics failure on display in SuperGuide is particularly apparent, there’s no shortage of opinions in which judges mishandle analysis of ambiguity. For two relatively recent examples from appellate opinions, see this forthcoming article about an opinion of the U.S. Court of Appeals for the Second Circuit and this article about an opinion of the Third Circuit.
The explanation is simple enough—diagnosing ambiguity, or alleged ambiguity, requires semantic acuity. Even though the language in SuperGuide was actually straightforward, it was more than the SuperGuide court could handle.
So what I say in my 2009 post holds true:
Just because a judge drives a car, that doesn’t make the judge an expert in automobile engineering. Similarly, the fact that a judge speaks and writes English doesn’t mean the judge is wise to the many ways that ambiguity can manifest itself in contracts.
To help avoid messes such as SuperGuide, judges should accept that much of the judiciary isn’t equipped to analyze ambiguity and that expert testimony can provide guidance. (And by the way, citing Strunk & White is not a good sign; see this takedown by Geoff Pullum.)
And litigators should recognize that they, too, are generally not well equipped to analyze ambiguity. If expert testimony on ambiguity isn’t admissible, litigators should make sure they have on hand appropriate expertise, so it informs their pleadings.
[Updated June 10, 2015: In this June 5, 2015 post I expand on the point I make in the final paragraph.]
12 thoughts on ““At Least One of X and Y”: The Semantics Train Wreck That Is the Federal Circuit’s 2004 Opinion in SuperGuide Corp. v. DirecTV Enterprises, Inc.”
Ken, I have nothing constructive to add, but this makes my head hurt. My first strong impression was that the court got it right. But your very precise analysis has forced my eyes to counter rotate, and now I am a helpless, throbbing blob of goo.
As is so often the case, I am reduced to the realization that I am clueless, which is not helpful to someone who has a transactional practice…and is prideful about it.
‘Helpless, throbbing blob of goo’? Sounds like one of Ken’s pet names for EDGAR.
I feel your pain. Given that this opinion escaped scrutiny for ten years, I can only assume that plenty of others have shared your first impression.
1/ Chinese menu analogy: ‘One from column A….’
2/ Ken, I’m surprised how far you veered from your usual analysis of cases: ‘I don’t care whether the court was right or wrong, I care only how good drafting could have avoided the dispute’.
3/ The language of set theory may help here. Generally speaking, ‘sets’ have ‘members’. The set ABC, for example, has members A, B, and C.
4/ With that heavy background, the ‘starter’ language at issue can be paraphrased as follows:
(1) ‘at least one of ABC, DEF, GHI, and JKL’
5/ The contending interpretations can be paraphrased this way:
(1a) ‘at least one of [the following sets]: ABC, DEF, GHI, and JKL’
(1b) ‘at least one [member] of [each of the following sets]: ABC, DEF, GHI, and JKL’
6/ I think 1a is the more reasonable than 1b, but 1b is not unreasonable. So the drafter is to be whipped for using language with two fair meanings. The fix is to use language like 1a or 1b, whichever captures the drafter’s intent.
7/ The conjunction is indeed a red herring. Whether you say ‘always avoid A, B, and C’ or ‘never do A, B, or C’, the members of the set ABC are identical.
Regarding your point 2, you of course are correct as to my usual approach. But a court’s analysis can be so misguided that it threatens to find dysfunction, or additional dysfunction, where none exists. That sort of analysis can result in a court’s deeming unclear what is in fact clear, and as such it’s a threat to contract drafters. In the coming days I’ll write more about this.
I look forward to more from you on courts deeming unclear what’s in fact clear. I guess no good drafting goes unpunished.
By the way, I’m a fan of Strunk & White *and* of Geoff Pullum’s takedown. The Elements of Style is not to be worshipped, but even if everything in it is wrong, it’s still right.
Omit needless words. Be concrete and specific. Put statements in positive form. Prefer the standard to the offbeat. Yes, ‘the active voice is usually more direct and vigorous than the passive’. Write with nouns and verbs. ‘Rather’, ‘very’, ‘little’, and ‘pretty’ are ‘leeches in the pond of prose, sucking the blood of words’. Throw the reader a rope. Revise and rewrite. Don’t despair because your manuscript needs major surgery.
How good it would be if the spirit of those injunctions blew steadily over the ocean of academic and judicial prose. (Or even just mine.)
The little book is no Cambridge Dictionary of English Grammar, but Strunk and White would deserve statues if all they had written was ‘omit needless words’.
Though I understand Ken’s position, I believe this had more to do with judges contorting their thinking to arrive at what they subconsciously “feel” is the correct answer is because they have seen Dish having the four categories on all the guide screens on the TV. A case of bad facts making bad law maybe?
Whatever the rationale, the problem is that the court’s nonsense influences how patent claims are written. Perish the thought that it strays into contracts!
The claim drafter should have used “and/or” (as D. C. ducks and runs for cover).
Incidentally, patents are often referred to as contracts, between the inventor / applicant and the U.S. Government: The applicant makes an enabling, best-mode disclosure of his or her invention, and in return, if the claimed invention meets the statutory requirements for patentability, the government gives the applicant a 20-year exclusive right to make, use, sell, etc., the claimed subject matter.
Bad idea — use of ‘and/or’ makes the patent language indefinite (definiteness being one of the requirements of patentability, See 35 USC 112). Examiners were routinely citing the indefiniteness card against claims with ‘and/or’ until Ex Parte Gross happened.
In patent related to Ex Parte Gross, the examiner rejected the claim as indefinite under 35 U.S.C. § 112, second paragraph, asserting that, “Despite the everyday usage of this term [“and/or”], the inherent ambiguity is [sic] creates as to the legal scope of the claims at hand require applicant to either pick ‘and’ or ‘or’ and not both.” The examiner also suggested that the applicant file separate claims to address the desired alternatives.
On appeal, the PTAB reversed the rejection, and agreed with the applicant that “‘and/or’ covers embodiments having element A alone, element B alone, or elements A and B taken together.” According to the PTAB, “and/or” is not wrong, but it’s not preferred. Since, the higher courts have neither seen this construction or ruled for and against such verbiage, I would steer away from it.
This discussion will be in MSCD 4, right?
Yes indeed. FYI, I’m pushing MSCD4 back a bit.