Saturday, October 06, 2007

Santa Clara dilution conference: Consumer Perceptions, Part 1

Ronald R. Butters, Duke University, Department of English (emeritus)

A Linguistic Look at Dilution

Dilution is barely noticed within linguistics even among consulting experts. If dilution is Bigfoot, as McCarthy suggested, then linguistics is what my dog thinks about Bigfoot: If he sees it, he’ll bark, but he hasn’t seen it.

Courts have accepted forensic linguists as trademark experts, but not so much in dilution. There are possible applications for linguistics:

Fame: Applied linguistics/lexicography. Historically, lexicographers have needed to be able to define trademarks, as terms of art. Dictionary makers have needed to develop concepts akin to TM fame, since dictionaries must rank words according to lexicographic importance and choose which to include. But it’s not simple; special consideration is given to arcane words that might nonetheless be encountered – brand names of drugs – as well as to words everyone knows. So TMs should neither get extra or less consideration because they’re in the dictionary.

Dictionaries in fact contain few TMs; the American Heritage Dictionary online has only 300 of 70,000 entries as TMs. Some of these are famous/arguably generic: Band-Aid, Frisbee, Xerox. Some are unfamiliar: Pentothol, Bondo. Omissions include: Microsoft, Burger King, Apple, and transparent omissions like American Airlines. Still, one could construct an evaluative metric for degree of fame from usage studies. Lexicographic methodology involves surveying large samples of text, which could be adapted for fame.

Genericness: Cases in which linguistic evidence was used: Beanie/Beanie Babies v. Screenie Beanie; Steakburger claimed as a common law TM by Steak & Shake; kettle for potato chips; zinger for spicy chicken menu item by Florida Ale House. Widespread use of a term in generic fashion may be incorporated into dictionaries: the American Heritage online includes Band-aid, Frisbee, Mace, Xerox, and notes that the marks sometimes appear in lower case as nouns and are used figuratively. The AHD doesn’t use the term “generic” because it doesn’t want to draw a legal conclusion (read: a C&D letter) even though that’s what it means.

Identity/similarity: This is related to the likelihood of confusion inquiry. Butters recalled the Lexis/Lexus case in which the court of appeals majority relied on testimony by an English professor about the supposed linguistic differences in pronunciation. Every linguist he knows disagrees with that testimony, but it’s still an instance in which such testimony influenced the outcome.

Jacob Jacoby, New York University Leonard N. Stern School of Business
The Who, What, When, Where and How of Measuring Dilution

He was the plaintiff’s expert in the Lexis/Lexus case, by the way.

We need data to measure dilution.

Among who? The general consuming public. He’s not convinced niche fame is dead. General and consuming are the same; why the redundancy? Must mean consumers of the product, which might not be nationwide or might not be demographically divided, like MODERN MATURITY (older consumers), MONISTAT (women), ISOTONER (colder climes).

When? After the mark is famous, which means brand owners need benchmarks.

How? You can use surveys to measure fame, as in the Pebble Beach case, in which his survey asked golfers to name the “most famous” golf courses in the US. The question asks the respondent think about fame. It uses the product category, because marks are famous for something. It uses unaided recall, not aided recall or recognition – the toughest possible test. More than one answer is permitted. Out of 13,000+ regulation golf courses in the US, 87% of golfers named Pebble Beach, while 25% named Pinehurst, the other plaintiff.

Similarity: Though no one’s done this yet in litigation, we could measure similarity through multi-dimensional scaling/conjoint measurement.

Association that impairs distinctiveness: this has to do with uniqueness or particularity. Jacoby used Pebble Beach and a case between Intel and Intell Management. In both cases, he measured change over time by surveying in 2 waves. In Pebble Beach, 14 days apart: asked them, before you heard about Tour 18, did you think that a Pebble Beach hole could only be played at Pebble Beach? Before learning about Tour 18, 73% said no; after, 85% said yes. (There are some pretty significant issues with this question, notably whether the statement that one can play a “Pebble Beach” hole at Tour 18 is in fact true, which depends on what it means to have a “Pebble Beach” hole.)

In Intel, the case was against a developer of upscale retail, housing, etc. projects; Jacoby surveyed experienced market participants. In-person groups were exposed to Intell and modified Ibell promotional materials. Seven days later, there was a phone interview in which they were led to believe a different survey was being conducted (Jacoby being IRB-free in private practice). The number who associated “Intel” with both a chip manufacturer and a real estate company in the test cell was 11.4-12.0%, and in the control cell it was 0%. That’s low, but in terms of whittling away it’s quite a big slice (… of real estate professionals aware of Intell). Intell is now Extell.

In conclusion, Jacoby discussed tarnishment, in the form of Anheuser-Busch v. Balducci, where a “Michelob Oily” ad affected respondents’ perception of Anheuser Busch, such that 15% more of the test group than the control group which saw a regular ad said the ad made them less likely to buy Michelob.

Rebecca Tushnet, Georgetown University Law Center
Dilution & Cognitive Science

I gave my standard spiel. Despite the language of science with which dilution proponents clothe their explanations, intuition is the basis of the cause of action, and we should acknowledge that rather than scientizing our presumptions.

I contrasted the messiness of the real world, where marketers constantly bombard us with ads, competing not just with direct competitors but with everyone else for our attention, and the tidy models of “owning a place in the consumer’s mind.” In the model, interference with attention is a form of trespass – except that we recognize many legitimate types of interference, for example Pepsi’s interference with Coke’s mental space. Why be concerned about dilution-based interference anyway?

I discussed the argument promoted by Jerre Swann that cognitive psychology justifies dilution through evidence that high-frequency words with many associations are harder to remember than high-frequency words with few associations. For low-frequency words, more associations don’t hurt and may help, but the argument is that dilution moves a mark from the “few, controlled associations” category to the “many, hard-to-remember associations” category. There are a number of problems with this argument, but the main one is: famous marks – even Apple and Ford -- are low-frequency, like almost every other word. The definition of frequency used in the underlying research is not intuitive (and our heuristics make us think that famous marks are high-frequency, when really it’s words like “day” that are). So the theory offered by dilution’s proponents doesn’t match the underlying research.

What about practice? In the lab, it’s possible to produce delays and errors in recognizing brand category membership and linking brand with attributes (e.g., GODIVA is chocolate, GODIVA has rich taste) by showing people ads for dilutive products (e.g., DOGIVA dog biscuits). Over three tested brands (Godiva, Heineken, Hyatt) respondents in the test cell were about 6.4% less accurate than control respondents who saw an unrelated ad when asked whether the major brand was a member of its category (chocolate, beer, hotels). Interestingly, because Godiva was so much less well-recognized than the other two brands, merely reporting the mean of all three makes the dilutive ads’ effects on accuracy look much bigger – an example of how statistics can be manipulated. It’s also notable that 6.4% is below the usual threshold for granting relief in TM cases.

Likewise, dilutive ads in the lab produce a response delay averaging 23 milliseconds in correctly recognizing the brand as a category member. This number is distorted because Hyatt Legal Services improved the recognition speed for Hyatt just as much as an ad for Hyatt hotels did. Response delays of several hundred milliseconds may seem trivial, but it’s certainly reasonable to think they matter, since people make affective judgments quickly, and a delay in response may make it less likely that consumers will choose that brand over others. So Dogiva dog biscuits could theoretically worsen Godiva’s competitive position against Ghirardelli, even though they don’t compete. But the Hyatt result shows that, at most, we don’t know enough to tell when a “dilutive” use will help or hurt. (Interestingly, another study used Hyatt tattoo parlors as a test, and they improved respondents’ preferences for Hyatt.) The best analogy for our mental associational models may not be trees with branching paths of conceptual linkages, but a series of one-way streets.

Dilution is, as an empirical matter, not proven. The scientific explanations are facades over the real concept of unfairness, which is not an empirical issue. Everything dilutes – ad clutter, competition (since one of the main things people think of when prompted with Pepsi is “Coke”), commentary, etc. Why protect TM owners from this behavior in particular? I’m prepared to argue that it’s in fact unconstitutional – an irrational regulation of a subset of commercial speech. Truthful commercial speech can’t be regulated if the scope of the regulation has nothing to do with the posited harm. Because everything dilutes, the FTDA/TDRA are the equivalent of taking a thimbleful of water out of an overflowing bathtub, and current commercial speech doctrine says you can’t do that even if the thimbleful is not itself very valuable.

David Welkowitz: Jacoby says the results for Pinehurst, 25%, are obviously proof of fame. But that’s less than half!

Jacoby: People were asked what they thought were famous courses; Pinehurst was 4th among golfers out of 13,000.

Welkowitz: But that just means it was #4 among golfers.

Jacoby: That’s up to you guys – it’s above my pay grade.

Q for Butters: Aren’t dictionary makers afraid of being sued, so the content isn’t exogenous to law?

Butters: Yes. The American Heritage is rather daring in showing pseudo-generic uses. Most dictionaries disclaim any legal statements. They get C&D letters all the time. Kitty Litter, it turns out, is a brand for greeting cards; one of his friends got harassing letters from the TM owner and took the term out completely. There’s also pressure from in-house lawyers even though there’s no cause of action against a dictionary. Art and Craft of Lexicography is a good book on this.

Mark Lemley for Jacoby: Your interpretation of “general consuming public” as only a limited public is flatly inconsistent with congressional intent, and is a strained reading besides. Many members of the public don’t make purchasing decisions – kids, people in nursing homes – the redundancy, if any, is to communicate the concept and Congress’s seriousness about it.

Jacoby: Maybe, but then why does Congress elaborate on geographical distribution in the factors?

Lemley: That’s consistent with the idea that fame can’t be geographically limited.

Jacoby: That’s above my pay grade (though kids are consumers and make decisions!).

My comment: “Above my pay grade” is an insufficient response. Jacoby is claiming that science provides answers which can then be translated into law. That’s not good enough, because we have an obligation to question the science, and because courts have historically responded to identifications of “harm” to expand the scope of TM law.

Question: Jacoby’s interpretation of general consuming public raises a larger point. Words have different meanings in context. Jacoby has techniques that measure things, then assigns legal names to them. What is the underlying theory that motivates the measurements? There must be a theory of relevance.

Jacoby: He’s done this in his 2001 Trademark Reporter article. There is widespread cognitive science agreeent about what goes on in the mind, confirmed with fMRIs. The associations we’ve talked about exist in the mind. His results reflect our best thinking right now, though science is always revisable. (Comment: of course his survey results are reflected in the mind; it would be surprising if they were not. The question is whether looking at an fMRI tells us anything about law.)

Comment: It’s important to emphasize, given the attention people are likely to pay to this specialized conference, that Jacoby’s interpretation of general consuming public is inconsistent with the legislative history, specifically the House Report on the TDRA. The president of INTA, which drafted the statute, testified that INTA rejected the concept of niche fame.

Question for Jacoby & me: Perhaps your conflict is not as severe as it seems. There are different things going on – Jacoby can measure indentations in skin, not the death of a thousand cuts. He’s identified something, but it’s de minimis, contrary to the assumptions of dilution doctrine.

My answer: Yes, plus there’s the anomaly of Hyatt – what we call dilution may sometimes help marks, and it’s a mistake to adopt a theory that uniformly posits harm.

Jacoby: The trademark owner wants the right to control the mark, even if the other use isn’t harmful.

Me: That’s a copyright theory – understandable why TM owners want the right to control, but dilution was sold on the premise that it was harmful. It’s “death by a thousand cuts,” not hugging by a thousand hugs.

No comments: