I will pause in answering the large backlog of reader questions from several months ago to address a general comment just now made by Eric M. Jones, who has been making very interesting contributions to the discussion:
Okay, Fred; I too have been guilty of Googling. But I propose that the massive trove of Google Books and similar digitizations is the best way to discover the root of sayings. How else can one do it?-by searching manually? By asking ‘experts’? By checking other ‘authoritative’ sources? When Google has digitized all printed materials, we can toss out the authoratative references and begin again with REAL research. … My guess is that you have a book to defend. My hope is that you toss out all the earlier stuff and Google like mad for new answers.
My response, which I also posted as a comment, was:
My book does not need to be defended against a superior Google Books-searching alternative. First of all, I used Google Books extensively in researching my book. The Yale Book of Quotations was the first reference work to make extensive use of Google Books. Second, because Google Books has serious problems with the quality of its metadata, it is limited in its usefulness for this kind of research. There are other databases, such as ProQuest Historical Newspapers and Newspaperarchive, that may be more powerful than Google Books, and I used them a lot as well. There are also, believe it or not, nonelectronic methods that will often provide better information than Google will, and I used them too.
Google Books is a fantastic tool for historical research. Its coverage becomes more enormous every day. But to regard it as the be-all and end-all of research is a serious mistake, and to uncritically accept its metadata about dates and editions, or even to assume that the metadata hasn’t been linked to a completely irrelevant book, can lead to whoppingly wrong assertions about the origins of words or quotations. “Experts” in Google Books searching know that if a result is surprisingly early, the explanation may be that Google Books has erred in dating the book, or in some other way, and that it may be necessary to-gasp-look at the printed book in a research library to verify the result. Much of the authority of the Oxford English Dictionary is derived from the fact that they verify citations of word usage in the original printed books in research libraries.
and 
. You can also watch it on 
Even if the edition and date are not in the metadata, aren’t they at least on a scanned page of the book?
Terry asks, “Even if the edition and date are not in the metadata, aren’t they at least on a scanned page of the book?”
Most of the books published after 1923 are only available in “snippet view”, so you can’t easily see when the book says it was copyrighted. Sometimes you can, but it’s difficult to be sure.
You have to be careful about “full view” books, too. Sometimes there are multiple items that were bound together or that were accidentally scanned together. Google Books indexes by just the first item.
A really subtle one is this book: http://books.google.com/books?id=cEcSAAAAYAAJ&pg=PA85&dq=%22eternal%20vigilance%22
It looks like it has “the price of liberty is eternal vigilance” from 1809 on page 85. If you look at the front of the book it says it’s printed in 1809. If you look at the next page it says that it’s a reprint of that book. It’s easy to miss that a new section starts on page 71 that was written in or after 1836! Arrgh!
But to regard it [Google Books] as the be-all and end-all of research is a serious mistake
Surely regarding any single source as the be-all and end-all of research would be a serious mistake.
It used to be standard grad school practice in the liberal arts to teach future researchers to evaluate the strengths and weaknesses of sources, as well as to check facts against multiple sources where possible. Is it different in economics?
I doubt Shapiro is an economist. Isn’t he a librarian?
Anyway, the metadata for Google Books is scary-bad. It’s a shame.
“Evil is a little man afraid for his job” is a line spoken by Roy Scheider in “Blue Thunder”. Is that the first appearance of this saying?