Note: thanks to the several people who sent this to me! In an opinion citing your humble correspondent, the court of appeals affirmed the lower court ruling, discussed here, that Havana Club rum is not falsely advertised by virtue of its name because no reasonable interpretation of the label as a whole could mislead consumers.
There’s a lot of history and a lot of litigation behind this, but the court offers a capsule summary: Before the Cuban Revolution, the Arechabala family made Havana Club run in Cuba and exported it to the US. The Communist Cuban government expropriated the business, and then the US enacted a trade embargo against Cuba. Despite this, the Cuban government registered Havana Club as a trademark for rum in the US in 1976. In 1994, the Cuban government assigned its interests, including the registration, to a joint venture of which Pernod’s parent is a member. However, [edited to correct facts] the U.S. Treasury Department's Office of Foreign Assets Control ("OFAC") then retroactively revoked the trademark transfer. In 2006, OFAC declined permission to renew the trademark.
In 1994, Bacardi also applied for a US trademark for Havana Club on rum and sold a nominal amount of rum imported from the Bahamas under that mark. The joint venture sued, and Bacardi bought from the Arechabala family any rights they had in the mark and the goodwill. The joint venture’s case was dismissed after the Cubans revoked permission for transfer of the mark. Bacardi’s application is still pending, though the USPTO denied registration of Bacardi’s applications for "Havana Select," "Habana Clasico," "Old Havana," "Havana Primo," and "Havana Clipper," for rum on the basis that they were primarily geographically deceptively misdescriptive.
Days after the Cubans’ federal registration expired, Bacardi began selling rum in Florida under the Havana Club brand name. It was distilled in Puerto Rico and made using the Arechabala family recipe. The Treasury Department’s Alcohol and Tobacco Tax & Trade Bureau approved the labeling (it’s supposed to stop misleading statements on alcohol labels).
The phrase "Havana Club" appears on the bottle in large stylized letters, followed by the word "BRAND" in much smaller letters. “Below that, in letters of prominent though slightly smaller size than those in the brand name and in a different font, the words "PUERTO RICAN RUM" appear.” Havana Club is repeated several times on the bottle. The back includes a statement in “clearly legible type”: "Havana Club Rum is a premium rum distilled and crafted in Puerto Rico using the original Arechabala family recipe. Developed in Cuba circa 1930, [it's good stuff]."
Pernod sued for false advertising. It presented “unrebutted survey evidence that approximately eighteen percent of consumers who looked at the Havana Club rum bottle were left thinking that the rum was made in Cuba or from Cuban ingredients.” Still, the district court found in favor of Bacardi, ruling that it had a First Amendment right to accurately portray where its product was historically made. Because the label truthfully disclosed the rum’s origin, there was no need to analyze actual or likely deception, including the survey evidence.
The court first said that the false advertising case was a proxy “for the real fight the parties want to have, which is over the right to the exclusive use of ‘Havana Club’ as a trademark.” Numerous times, the court said it wasn’t deciding what should happen if trademark rights were involved.
The court recited the general rules that likely deception is presumed if a plaintiff proves unambiguous and literal falsity, and that a literally true or ambiguous statement requires proof of actual deception or a tendency to deceive, which can be done with a survey.
But not always! Pernod argued that the question of whether an ad implies an inaccurate message “virtually demands” a survey because it centers on consumer perception and memory. Bacardi argued that the first step was to determine what message was conveyed by an ad, which can sometimes be done just by looking. The district court did a bunch of odd things, including holding that “origin” in §43(a)(1)(A) didn’t mean the same thing as “origin” in §43(a)(1)(B), so that while “origin” for the former means place of manufacture, it could mean recipe or heritage in the latter. But that didn’t matter because Pernod lost anyway: the label clearly and truthfully provided the geographic origin of Bacardi’s rum (the court also found it persuasive that the relevant federal agency approved the label). Quoting the no-good Mead Johnson decision, the district court held that survey research doesn’t determine the meaning of words. As to heritage, Havana Club rum had a Cuban heritage, and saying so wasn’t deceptive. Pernod’s survey evidence was necessarily irrelevant because of Bacardi’s First Amendment right to advertise that heritage.
The court of appeals held that, while it might not agree with “every aspect” of the analysis, it was persuaded that the result was right. In particular, the label as a whole could not mislead any reasonable consumer about where the rum was made, “which means that survey evidence has no helpful part to play on the question of what the label communicates regarding geographic origin.” The court did not address the case law indicating that "reasonable" consumers means target consumers, or otherwise question the factual predicates of the survey, so I guess that means that 18% of relevant consumers are unreasonable. Too bad for them!
The court of appeals did question “whether the District Court should have endeavored to use the modifier ‘geographic’ to expand the meaning of ‘origin’ into the realm of history, heritage, and culture,” given that Dastar specifically talked about false advertising of origin, but didn’t decide the issue.
The issues here are both factual and philosophical: “The central conceptual problem before us is whether language can be clear enough that its meaning is beyond reasonable dispute.” Implication is of course part of false advertising law. Words need to be read in context to find what meaning a reasonable reader could be expected to give them. “[T]here is and must be a point at which language is used plainly enough that the question ceases to be ‘what does this mean’ and becomes instead ‘now that it is clear [ed. note: to whom?] what this means, what is the legal consequence.’”
Mead Johnson demonstrates this principle. “[T]here is a baseline meaning to some words that put them beyond any credible claim of misunderstanding.” Kindly citing me as one of Mead Johnson’s detractors, the court still agreed with its “general proposition that there are circumstances under which the meaning of a factually accurate and facially unambiguous statement is not open to attack through a consumer survey. In other words, there may be cases, and this is one, in which a court can properly say that no reasonable person could be misled by the advertisement in question.”
Comment: I can see how you get to this result. But it seems to me that you then have to own up and say (1) there was something wrong with this survey, because if it was done right then it wouldn’t have found substantial confusion, (2) the law actually does not try to protect actual consumers if a substantial proportion of them fall below the court’s determination of how a reasonable person should react, or (3) some separate normative concern justifies tolerating this confusion (as the district court did with the First Amendment). You can’t just say the determination has to be made from the viewpoint of an ordinary consumer and then say that, despite the best evidence you have of what that viewpoint is, you’re going to find otherwise. This is especially true from the perspective of predictability, because—despite what the court says below—I do not see how a court is supposed to decide when it should ignore a survey that is methodologically sound.
Still, the court was kind to me, pointing to “recent, thoughtful, academic commentary [arguing] that, in a false advertising case, consumer survey evidence should not be necessary to demonstrate consumer confusion if that confusion would be clear, i.e., ‘when other factors strongly favor a finding of likely confusion.’” “While the parallel is not perfect, it is similarly appropriate to conclude that a consumer survey should not be necessary or dispositive when it is plain from an advertisement that rational confusion is not possible. A logical conclusion, and one we reach here, is that if ‘[c]ourts should be more willing to use common sense in finding deceptiveness,’ they should likewise be prepared to use common sense to conclude that an advertisement, when taken as a whole, could not mislead a rational consumer.” (Internal quotes from my Penn article.) If this means that the Third Circuit is relaxing the explicit/implicit division on the side of allowing courts to use common sense to find deceptiveness even without a survey, that would be great, but given that it recited the explicit/implicit division early on in the opinion I don’t think that’s exactly what happened. Sauce for the defendant may not be sauce for the plaintiff. (No pun intended.)
Anyway, here there was a factually accurate, unambiguous statement of the geographic origin of the rum. The label clearly stated on the front that it was “Puerto Rican Rum” and on the back that it was made in Puerto Rico. The court didn’t need to consider a situation in which this statement was in fine print, because the label here was plainly legible and “particularly prominent on the front of the bottle.” No reasonable consumer could be misled by those statements, and the rest of the label doesn’t create doubt as to their meaning (given that Havana Club was not patentently unambiguous the way “Cuban Rum,” “Havana Rum,” or “Rum of Havana” would have been).
In isolation, the words Havana Club might be misleading. But they weren’t in isolation, as they might be considered in a trademark registration case. Viewed in context, “any thought a consumer might have that the words ‘Havana Club’ indicate the geographic origin of the rum must certainly be dispelled by the plain and explicit statements of geographic origin on the label.” (Again, the “must certainly” flies in the face of the survey, unless there’s more to be said about the survey, which did feature the whole bottle.)
Under the circumstances, the district court could properly disregard survey evidence is immaterial, “because, by definition, § 43(a)(1) does not forbid language that reasonable people would have to acknowledge is not false or misleading.” However, the court noted, it was not holding that literally truthful claims can’t be misleading. Instead, “false advertising claims occupy no special niche separating them from the standard rule that, once something--for example the meaning of text--is beyond reasonable dispute, there is no longer a question of fact; there is only the question of how the law applies to established fact.”
The probative value of a survey is highly fact-specific and a court may place such weight on it as the court deems appropriate. “A consumer survey does not trump the district court's role as arbiter of the reliability and trustworthiness of expert and hearsay evidence based upon statements of consumers reacting to an advertisement.” The same judge who coined the “now widely-accepted proposition” that “the court's reaction [to an advertisement] is at best not determinative and at worst irrelevant,” American Brands, Inc. v. R.J. Reynolds Tobacco Co., 413 F.Supp. 1352, 1356-57 (S.D.N.Y.1976), also observed that a court “is obliged to judge for itself whether the evidence of record establishes that others are likely to be misled or confused. In doing so, the court must, of course, rely on its own experience and understanding of human nature in drawing reasonable inferences about the reactions of consumers to the challenged advertising.” McNeilab, Inc. v. Am. Home Prods. Corp., 501 F.Supp. 517, 525 (S.D.N.Y.1980). The court of appeals here added that, “when the meaning of words is beyond all reasonable dispute, there is no issue of fact on the question of meaning.”
But compare that to what the court said about the placement of the same words—which would of course have the same meaning, “beyond all reasonable dispute”—in fine print on the same label. If that might be deceptive, which the court here suggests (and the caselaw is pretty solid that you can't take back in fine print what you said in large), then it’s not meaning in an abstract sense but reception that’s at issue. It may well be that, when consumers receive a geographic origin message from a brand name, a substantial number of them don’t bother searching out more information on the rest of the label, even if it’s in large type. That is, at least, what the survey results here suggest. Now, one can take the position that this is not deception worth correcting based on other considerations—materiality possibly among them—but it’s still deception.
The court also relied on puffery cases for the proposition that ad text can be clear enough “that it simply cannot be challenged as misleading.”
Unwilling to give too much license to trial courts to reject surveys, the court
hasten[ed] to add that cases like the present one should be rare, for one hopes that a case with truly plain language will seldom seem worth the time and expense of contesting in court…. A word of caution is nevertheless in order, so that our holding today is not taken as license to lightly disregard survey evidence about consumer reactions to challenged advertisements. Before a defendant or a district judge decides that an advertisement could not mislead a reasonable person, serious care must be exercised to avoid the temptation of thinking, ‘my way of seeing this is naturally the only reasonable way.’ Thoughtful reflection on potential ambiguities in an advertisement, which can be revealed by surveys and will certainly be pointed out by plaintiffs, will regularly make it the wisest course to consider survey evidence.I applaud the call for epistemic humility in general, though there are definitely problems in figuring out where that humility should end.
In another way of framing the warning to district courts, the court of appeals pointed out that a decision to disregard survey evidence will be reviewable de novo (as opposed to the deference given to weighing it), because it’s a legal conclusion based on underlying facts. The relevant facts here: the bottle has an unambiguous, prominent statement of origin and also another statement that, in isolation, could be an ambiguous geographic reference. Applying those facts to the ultimate question of reasonable confusion is a matter of law “and, thus, requires no consideration of survey evidence.” (I’m a bit confused about how that follows. Why isn’t the presence of the survey, or the results of the survey if you prefer, also a fact that needs to be considered in applying the law?)