I spent part this evening as part of a panel talking about Google Book Search. 

The publishers take the view that any effect on any prospective market they might want to enter into fatally undermines a finding of fair use.  The logical outgrowth of this position is that because there are innumerable possible markets out there, mass uses of any kind can't possibly be fair -- to the publishers, scale matters enormously.

Translated:  Google makes a complete scanned copy of books.  That copy isn't made available to anyone else (other than the libraries with which Google has arrangements, and their uses are in turn sharply limited.)  That reproduction is fair use, in my view, and doesn't require permission from the publishers.

Why?  Because it's an essential step towards the transformative "snippet" view -- can't get there without making a copy.  And even though it's a complete copy, that doesn't mean the use isn't fair (see the "multiple copies for classroom use" part of Sec. 107).  And the existence of this reproduction doesn't undermine sales of the book.  In fact, the snippet view may actually lead to many more sales.

Well, that doesn't satisfy the people who have sued Google.  They point to the fourth factor in Sec. 107, and say, "Google's possession of this copy is having an effect on the POTENTIAL MARKET for the work.  Google is working itself into a position to be the world's bookstore for e-books.  We won't be able to create our own market along those lines, and so our potential revenues are being undermined."  Q.E.D.

The world is sufficiently unpredictable that anything could happen, right?  So fair uses that threaten any possible secondary market can't exist, according to the publishers.  In effect, they'd like to use copyright law to protect against network effects and first-mover advantages that they can't personally monetize.

I very much hope that Google won't settle this case.  We need these issues decided.