Administrative Law

(posted by Peter Hirtle)

Lorcan Dempsey has an interesting blog posting about new publishing ventures in libraries.  He mentions in passing an article in the Economist discussing the Google Book Settlement that opens with an example of two "orphan works."  One is The Appalachian frontier: America's first surge westward by John Anthony Caruso; Lorcan found three editions of this in WorldCat.  The other work was Blunder out of China.  There are two works with that title in WorldCat: one by Geraldine Townsend Fitch and Theodore H. White published in 1946, and one by Walter Henry Judd published by GPO in 1947.

The inclusion of these two examples as "orphan works" in the Economist article illustrates a number of interesting points.  First, it shows how sloppy some people have been when they have referred to "orphan works" and the Google Book Settlement.  Neither of the titles that the Economist cites could be considered to be an "orphan work."  Second, it also points out that metadata problems aren't confined to Google, and how difficult it is to determine copyright status just based on electronic records.  Finally it is a good example of how the Google Book Settlement is the only way that the full text of most books can be made available.

If one checks the Copyright Office's database of registrations and renewals, there is no evidence that copyright in the 1959 edition of the Caruso book was renewed in its 28th year, which puts the text clearly in the public domain.  There is a copyright registration for the new matter added to the 2003 edition, and the claimant for that copyrighted matter is the University of Tennessee Press.  The press can be found with relative ease, so you can't claim that the 2003 edition is an orphan work.

The Blunder out of China example is somewhat more complicated since we don't know which edition Paul Duguid had in his office when he was interviewed for the article.  The Economist doesn't help much.  It refers to the book as a "seminal text," but I wouldn't describe a pamphlet implying that Theodore H. White was a communist as fitting that description.  For copyright purposes, though, it doesn't matter which of the two books in WorldCat was digitized since they are not different works.  The Judd work is an offprint of the Congressional Record of 26 July 1947, where Judd reprinted Fitch's pamphlet in its entirety.  (The difference in page count between the two editions is due to the small type that the Congressional Record used for congressional "extensions of remarks.")  There is only one copyrighted work: Fitch's pamphlet.

So let's look at the metadata problems.  First, the WorldCat record gives Theodore H. White as an author.  Authorship can be important in determining copyright status.  White was not the author of the pamphlet, however, but rather the subject.  The pamphlet is an attack on his book, Thunder out of China, which was published in 1946.  Second, the WorldCat record gives a publication date of 1946.  Publication date can be important when checking renewal records.  I have not been able to confirm the date of publication of the pamphlet, but the Library of Congress record has [1947?].  Cornell's copy was received on 26 May 1947 according to a penciled notation in it, and the fact that a Congressman felt the pamphlet was important enough to include in its entirety in the Congressional Record in late July, 1947, suggests to me that the publication date was 1947, not 1946.  (If someone really cared about this, they might want to check Judd's files on the republication of the work found in his papers at the Minnesota Historical Society.)

Now on to the physical book.   A check of the volume shows that there is no copyright notice in the volume, which would have injected it into the public domain as soon as it was published.  Furthermore, there is no registration record for the pamphlet in either 1946, 1947, or 1948, nor is there a renewal record for it, further confirming that it entered the public domain when it was published.

The one last issue would be whether copyright could have been restored.  For that, we need to know where the item was published and the nationality and domicile of the author.  Geraldine Fitch was, I believe, an American married to a missionary and living in China at the time the pamphlet was printed, so there is the possibility that copyright in it could have been restored.  The pamphlet, though, states clearly that it was "printed in the United States," so copyright restoration is not an option.  This work, too, is in the public domain.

Why does this experiment suggest that the Google Book Settlement is needed?  Look at the effort it took to establish that these two works were in the public domain.  The same effort would have been required to determine that the books were still protected by copyright but out of print.  One would then have to make a reasonable effort to locate the copyright owner before the work could be considered to be an orphan work.  (Only Brewster Kahle, as far as I can tell, believes that an out-of-print work is an orphan work.)  The expense of doing this would be horrendous.

If we want to make the full-text of in-copyright but out of print books available without massively altering copyright law, there is only one cost-effective option: the settlement.