The preamble of the Lanham Act Section 2 is followed by six lower-case–lettered sections, the fifth of which contains four separately numbered “grounds” on which it is permissible to refuse registration. No form of the word “generic” is to be found anywhere in any of them. How, then, does the USPTO prevent registration of generic names of the goods for which the marks are intended to be registered? . . . We all know that a thermos keeps liquids warm or cold substantially longer than if the soup or soda were kept in a non– vacuum-insulated container. Many apparently believe that THERMOS is one of the leading brands of thermos. This seems nuts.
John Welch links to this fantastic article in the Trademark Reporter by Anthony L. Fletcher, discussing the critical — and, like so much in trademark law, vaguely described — legal distinction between generic and descriptive trademarks (or rather, in the case of genericness, would-be trademarks):
Most trademark lawyers are aware that generic names are unavailable for trademark protection under any circumstances, while merely descriptive terms are available for trademark protection if, and only if, they have acquired secondary meaning in the marketplace. “Secondary meaning” arises from use of a descriptive, or otherwise unprotectable, word or phrase in such a way, for a sufficient duration, and with enough frequency, that it becomes uniquely associated with the user and/or the user’s product(s). It is “secondary” in that it refers to an acquired, rather than to an inherent, meaning. Accordingly, it is critical to understand the “generic”–“descriptive” dichotomy for trademark management and planning purposes. Promoting a name that can never become a trademark as if that name were a trademark is an exercise in futility. While promoting a generic name as a trademark might result in a degree of trademark recognition by the public, when that begins to happen, competitors can step in and, with impunity, appropriate to themselves and/or their products whatever degree of de facto trademark meaning might have been created. . . . Unfortunately, the consensus of United States legal authority on what is generic rather than descriptive, and vice-versa, has become somewhat discordant. . . . On the face of things, it appears that generic words are most likely to be nouns, while descriptive terms are more likely to be adjectives. On the face of things, there are also hundreds of judicial and administrative holdings on what is generic and what is not. But what is needed is guidance for determining whether a term is generic or descriptive. That is not to say it hasn’t been attempted.
And here I break off for two reasons (copyright infringement not being one; I’m nowhere near the fair use threshold). One is that you, dear reader, need a break. And the other one is that what follows is worth its own separate block quote:
Perhaps the most colorful, and surely one of the most widely known, attempts at such guidance is: “A generic name is one that refers, or has come to be understood as referring, to the genus of which the particular product is a species.” Sadly, while the formulation has literary merit, there is little reason to believe that jurists, jurors, or attorneys are adept at dividing intended brand names between genera and species.
Sweet. 
Fletcher’s words here made me think of an unrelated intellectual property debate, including his allusion to literary merit. I mean the decision in Cariou v. Prince, sadly not reported here in a timely fashion, which raised the topic, especially in the dissent, of what business and how much business the courts have in moving into literary criticism to decide the question of fair use base on transformation in copyright infringement cases. Related, but too tangential for now. Let’s revisit that one later. I actually thought Fletcher’s criticism of this formulation would be, though, it was a tautology: How do we know if something is generic? Why, the name refers to the genus of the thing! But I think that point would border on the tautological as well — in fact, the phrase “tautological standard in trademark law” is probably not only a tautology itself, but is probably, um, “generic,” along with “unpredictable standard in trademark law.” Or have you heard about this LIKELIHOOD OF CONFUSION thing? Anyway, the article is great, and you have to read it (and I have to read it more, and better.) Fletcher’s conclusion:
Quite simply, the problem is that it is impossible to foresee with tolerable accuracy where courts, or the USPTO, will draw the line between generic names and descriptive terms. We can probably all agree that “car” is generic and “fast” is descriptive, but what is a “sports car” or “muscle car”? Are they generic or descriptive-generic? “Car” is plainly generic. But if “light beer” is generic, there is no apparent reason why “sports car” can’t be.
He goes on to suggest a handful of approaches to the problem. They do not, at first blush, simplify matters. But perhaps, if adopted, they could clarify them — an end sorely to be desired in trademark law. If it’s going to be this busy, this expensive, this profitable, can’t it be a little more predictable? Or have I missed the point — again?
