The color purple

A well-known book; not so well-known as a mark for 3M's abrasive products, it seems. The TTABlog reports on the TTAB's ruling, after a hard-fought battle that included a consumer survey by the opposer, denying registration to 3M on the ground that purple had not acquired distinctiveness and was in any case functional.

The well-advised 3M promoted its products with the following slogans: “Ask for it by color, not by name” and “If it’s Purple, it’s from 3M.” The evidence was, however, that numerous producers of abrasive products used purple. So the slogans attempted to create trademark meaning in consumers' minds (though the TTAB found that 3M had failed in this aim), which is ordinarily acceptable, but especially the latter one strikes me as potentially false advertising. A consumer who believed this would either start being confused about the source of other, existing purple abrasives -- which seems counter to the ordinary purpose of trademark, to reduce search costs -- or would regard the other producers as infringers, a sort of reverse confusion situation.

Trademark relies on shaping consumer understanding, of course, but it seems that the preexisting understanding here raises unfair competition questions when 3M tries to appropriate for itself what others have freely used. Precisely because we're used to word marks, we don't usually require instruction to regard them as marks -- even if a particular term is descriptive. But color, like other forms of trade dress, is usually perceived by default as decorative or functional rather than source-indicating. So creating trademark meaning may require aggressive, explicit instruction of consumers in a way that's uncommon with word marks. And that instruction itself may at least start out making false claims about meaning -- but if it succeeds, those claims are no longer false! This is a largely unrecognized problem in regulating false advertising, one I'm exploring in the paper I'm giving at WIPIP later this month.

(Side note: opposer's survey used orange sandpaper as a control against purple; the survey expert contended that, when the percentage of respondents who claimed that orange was a source indicator was subtracted from the percentage who said the same of purple, the net recognition was 6%, "about the same level you’d expect just based on noise or error.” But that quote is deeply misleading, since the control itself is supposed to account for noise and error. Six percent should not suffice to show secondary meaning for color; but neither should double-counting be rewarded.)