Maria Bonn is the Director of the University of Michigan’s Scholarly Publishing Office. She is responsible for the production of electronic books and journals and for broadly developing the role of the Library in scholarly communication. In an interview by Mary Minnow of the Stanford Copyright and Fair Use website she discussed how the University determines whether a book is under copyright:
On determining copyright status: right now we are being fairly conservative in our copyright judgments. Keep in mind that these volumes are only for sale in the United States, so we are guided by U.S. copyright law. We run an automated analysis on the MARC records to identify all volumes published prior to 1923 and most U.S. government publications. After this analysis, he bibliographic information is sorted by publication date and undergoes a quick manual review to check for obvious errors or bibliographic oddities (such as a record for a book that asserted it had been published in 1099). These volumes are removed from the POD (Print On Demand) stream. There is a wealth of material that is relatively easy for us to identify as public domain, and these are the books that we are currently working on getting out into the world. There are many other books that are probably also public domain but we’ll need to do a more nuanced analysis in order to make that determination.
If we extend our arrangements with POD printers and distributors to allow for sale of books in other countries (we already have that option for the U.K and Germany, but have not acted on it), we will need to construct a similar process of analysis taking into account the copyright laws of the countries in which we are selling the books.