Let a Thousand Googles Bloom

Last month, Google announced a partnership with major research libraries to scan 20 million books for inclusion in Google's search database. For those works in the public domain, the full text will be available. For those works still possibly under copyright, only snippets will be seen. The potential of this project is only beginning to be understood -- it is likely to bring about the most dramatic changes in the nature of research and the spread of culture since the birth of Google itself.

But the excitement around Google's extraordinary plan has obscured a dirty little secret: It is not at all clear that Google and these libraries have the legal right to do what is proposed. For work in the public domain, the right is clear enough. But for work not in the public domain, Google's right to scan -- to copy -- whole texts to index is uncertain at best, even if it ultimately makes only snippets available. When permission has been given by the copyright holder, again there's no problem. But when permission has not been secured, the law is essentially uncertain. If lawsuits were filed, and if Google and its partner libraries were found to have violated the law, their legal exposure could reach into the billions.

Google, to its credit, has decided to accept these risks. It can afford to fight the lawsuits, and the benefit to society and Google from such access apparently outweighs its potential costs.

But not everyone is Google. Not every library could afford the risks that Google can. And so before we accept a world where only a Google can build valuable, network-based digital libraries, we should ask whether the system that produces these profound uncertainties is a system that we should change.

The basic problem is simple. A copyright is a property right. Yet our particular system of copyrights is insanely inefficient. Rights get created easily enough -- a copyright is automatic; you need do nothing to secure it. But tracking who owns the rights created is astonishingly difficult. There is no Google for determining what works are protected by copyright; there is no Google for tracking down current copyright holders. The law creates a property right, but leaves it practically impossible to respect that property right for older, out-of-print works.