I was just thinking about why I'm a little bit bothered at the idea of indexing works being a transformative use.
In Perfect 10 v. Google, the 9th Circuit said that creating thumbnails of copyrighted images in the course of indexing images for a visual search engine was a fair use. I wholeheartedly agree with this—it doesn't harm the market for the original works, and the purpose of the use certainly seems to be consistent with a fair use. But the 9th Circuit also said that this use was transformative, and that always struck me as odd.
Maybe it's because I'm used to thinking of transformative works as being at least partially congruent with derivative works—and couple that with the idea that usually, you have some element of creativity or creative direction going into the creation of a derivative work. This concept came up in a recent conversation I had about the Kindle text-to-speech / derivative works controversy I recently blogged about.
Michael Kwun at EFF also wrote a blog post on the matter, and he noted that, in addition to the Kindle not creating a derivative work (because the read-aloud text isn't fixed), the Kindle also doesn't create a derivative work, because the work is substantially identical to the text in the eBook—nothing creative has been added that would meet the (relatively low) standard of creativity required for copyrightability.
I left that point out of my blog post because I hadn't poked deeply enough into the issue to be sure that that was right—in defining derivative works, section 101 notes such examples as abridgements, editions, or in some cases even annotations. So I was content to hang my argument on the fact that this wasn't a work.
An aside: translations are also noted as a derivative work, but I think that looking at text-to-speech as a form of translation is a red herring. Translating from text to speech is a format shift—whether or not that's trivial for a computer doesn't change the fact that there is a one-to-one correlation between a written and spoken word. That isn't true by any means when translating from one natural human language to another—there's a massive amount of original thought and creativity in choosing the right word to convey the proper connotations, given different idioms, cultural references, even the sounds and lengths of the words.
But regardless of whether or not a mechanical text-to-speech conversion (which would be missing the interpretation inserted by human inflection and performance) could rise to the level of a derivative work, I don't know that a database or index of thumbnails would, either.
There isn't much creativity going into the creation of an index—in fact, that lack of creativity and the shallowness of the indexer's interaction with the work is part of what makes me think of the use as being less harmful to the copyright holder, and therefore more fair.
Which is why, although the concepts of derivative works and transformative uses are linked, I'm not sure that calling indexing transformative is the best logical fit—even if it might be the best legal fit into the framework of fair use.
In the Google Books settlement, Google retains a copy of the entire body of scanned material—the “research corpus,” which will be made available to some researchers for “non-consumptive” uses. Large bodies of human-written text are extraordinarily useful to linguists and other social scientists, who can use them to trace the usage of particular words or other features of natural language usage. Such studies need not read the texts for their meaning—using a text to locate the earliest instance of the word “chortled,” for instance, doesn't involve using Through the Looking Glass in any way that a consumer of the text would. The text isn't being read for its content; it's being analyzed for its form. The usage is non-consumptive: it's a shallow involvement with the copyrightable aspects—the creativity—of the work.
Indexing a thumbnail is an even shallower interaction with the copyrightable aspects of a work. There is even less human direction and consumption in an automated image crawler and compression utility than in analyzing the form of a text. So while this usage isn't transforming the work, the use is non-consumptive.
One way to rationalize the use of the term transformative is to focus on the fact that the use is what must be transformative, rather than the allegedly infringing work. This is essentially what the 9th Circuit did in Perfect 10; even though the opinion quoted the rule from Campbell as saying that “The central purpose of this inquiry is to determine whether and to what extent the new work is 'transformative,'” it then goes on to apply that rule this way:
Google’s use of thumbnails is highly transformative. In Kelly, we concluded that Arriba’s use of thumbnails was transformative because “Arriba’s use of the images serve[d] a different function than Kelly’s use—improving access to information on the [I]nternet versus artistic expression.” Kelly, 336 F.3d at 819. Although an image may have been created originally to serve an entertainment, aesthetic, or informative function, a search engine transforms the image into a pointer directing a user to a source of information. Just as a “parody has an obvious claim to transformative value” because “it can provide social benefit, by shedding light on an earlier work, and, in the process, creating a new one,” Campbell, 510 U.S. at 579, a search engine provides social benefit by incorporating an original work into a new work, namely, an electronic reference tool. Indeed, a search engine may be more transformative than a parody because a search engine provides an entirely new use for the original work, while a
parody typically has the same entertainment purpose as the original work.
Here, it seems that the opinion is trying to wrestle with the difference between how transformative the new work is and how transformative its use is. Plainly, a parody makes more fundamental changes to the original, even if its use—entertainment—more closely resembles that original.
And the difference between the original and new works is likely what the Supreme Court had in mind in setting out this guideline in Campbell, a parody case. So maybe it's better and more consistent for another doctrine to account for the non-consumptive uses involved in automated processes like indexing. As more and more works are manipulated mechanically—and in ways that range from de minimis and incidental to fairly involved—the extent to which an operation touches the heart of a work's creativity, and not just how human that contact is—should probably be taken into account when evaluating fair use.