On 3/7/17 4:18 PM, Harper, Cynthia wrote:
>  Perhaps scanning the indexes would
> be legally defensible because they constitute less than 10% of the
> work...

Do keep in mind that figures like 10% as a maximum threshold for a Fair 
Use come from various consensus documents developed by groups of 
stakeholders related to certain media formats, but they have no actual 
legal standing.  There are many factors that play into Fair Use, and 
proportion of the work is just one of them.

That said, if you had a collection of OCRd indexes and built a search 
engine told you which works contain a mention of a term in the digitized 
index, a court might well consider this a transformative fair use -- 
akin to HathiTrust's digitization of in-copyright works -- and not find 
you guilty of infringement.  But I'm not an attorney, so this should not 
be construed as legal advice.

Kevin