BROOKE GLADSTONE:: When it comes to intelligence failures in the lead-up to the Iraq war, Washington tosses around blame like a hot potato. But everyone agrees on the need to connect the dots better in the future. And so this month, intelligence officials launched the Open Source Center. It will be devoted to gathering all sorts of freely available information, from foreign media reports to websites to - and this is directly from the New York Times - "slogans on tee-shirts in countries of interest." By piecing these info tidbits together, the government figures it can learn a lot about the plans of potential adversaries without having to rely on spies and satellite images. But if we can do it, so can they. So the government tries to keep our bits under wraps, especially in Freedom of Information lawsuits, by invoking the "mosaic theory," a trusty old tool, often used in the Reagan years. According to the theory, individual bits of seemingly innocuous information considered alongside other bits of information can help paint a picture that's greater than the sum of its parts and expose important secrets to the enemy. But in the next issue of the Yale Law Journal, law student David Pozen argues that the "mosaic theory" by its very nature is especially prone to abuse.

DAVID POZEN:: The "mosaic theory," I think, has been kind of coded over time in National Security law as this dangerous trope. It conjures up the idea of unknown adversaries exploiting information in unpredictable ways. And after the terrorist attacks of 9/11, that prospect seems especially frightening. Moreover, there's a kind of limitless character to the "mosaic theory" reasoning, that while it's true that some information, though independently innocuous, can become more dangerous in the context of other information, that argument could be extended to apply to pretty much anything.

BROOKE GLADSTONE:: Can you think of a case where you think the "mosaic theory" was applied egregiously?

DAVID POZEN:: Well, I think perhaps the most egregious example was a case called North Jersey Media v. Ashcroft. In that case, the government moved to close deportation hearings to the public and the press. These deportation hearings have traditionally been open to both. The government defended this action on "mosaic theory" grounds, that information that might come out in these deportation hearings could threaten national security. What I think was particularly notable about North Jersey Media is that the government and then the court applied the "mosaic theory" to justify closing off all of these special interest deportation hearings rather than looking at the individual issues arising in a given person's hearing. And some language in the opinion is particularly interesting. In paraphrasing the government's arguments, the court says that it's unable to see the mosaic here. And this I found a kind of crazy statement, because why should the court be less able to see the mosaic of dangerous information potentially arising out of these hearings than any other body?

BROOKE GLADSTONE:: On the other hand David, one could say that if there's anyone in a position to know which information might be particularly susceptible to a mosaic reading, it would be the government agencies whose job it is to generate and analyze this stuff. So shouldn't judges, who don't necessarily have any expertise in national security matters, be deferring to the experts?

DAVID POZEN:: I think courts should be deferring but they shouldn't be abdicating their review role. This is why we have these courts in there at all, because we want to strike a balance between national security needs and the public's right to access this information. I think that there's certainly a way in which courts can respect government arguments that are plausible and reasonable but still call the government on it when an argument just isn't a realistic threat.

BROOKE GLADSTONE:: I guess that as tricky is it is for the government to argue its case in favor of the "mosaic theory," it may be more difficult to argue against the "mosaic theory" in a given case.

DAVID POZEN:: That's a big concern. The government is arguing that adversaries out there at large could use this information in some perhaps unknown way to threaten national security. And it's pretty much impossible to rebut that claim, you know, precisely because "mosaic theory" often cannot be proved, so to speak, the way a normal argument might be. It can become an attractive argument for agencies to gravitate to when they know their argument for withholding is tenuous. And because of that kind of potential for abuse, overbreadth and opportunism, I think courts need to be especially vigilant when they see a "mosaic theory" claim. Also, I think courts should realize that mosaic-making is ubiquitous. Humans always make mosaics when they receive information. We always are looking for these inferences and these leaps between pieces of information to get something new out of them. So that being said, every case is, at heart, a "mosaic" case, and courts should realize this and not treat these cases differently than other FOIA cases just because the mosaic metaphor is invoked.

BROOKE GLADSTONE:: Is it true that in all the times that lawyers have tried to rebut the "mosaic theory," only one case have they won?

DAVID POZEN:: In FOIA, that is the case. The government's track record is excellent at winning when it invokes the "mosaic theory," although it should be said that the government's track record is excellent generally in national security litigation.

BROOKE GLADSTONE:: Can you tell me about the FOIA case in which the "mosaic theory" actually lost?

DAVID POZEN:: Sure. That was a recent case called Gerstein v. the Department of Justice. There, this reporter wanted summary statistics on the FBI's usage of Section 213 of the Patriot Act that allows the FBI to get search and seizure warrants without immediately notifying the warrant's target, and the government rebutted with the "mosaic" argument. It said that basically if adversaries were to learn about the usage patterns of Section 213, then they could revise their counter-efforts accordingly and better be able to get around the usage of these search and seizure warrants. There, the court actually said that this argument just doesn't hold up, and so it ordered the release of that information to the reporter. And I think that this case actually provides an interesting template for other courts to look at. The court did its own independent analysis to say that while we generally defer to agency opinion, this is just not a claim that's strong enough to win under FOIA.

BROOKE GLADSTONE:: David, thank you very much.

DAVID POZEN:: Thanks very much for having me on.

BROOKE GLADSTONE:: David Pozen is in his second year of studies at the Yale Law School. His article on the "mosaic theory" appears in the Yale Law Journal next month. [MUSIC UP AND UNDER]

BOB GARFIELD:: Coming up, big trouble at the L.A. Times, and a new advertising campaign for a down-market drink. Got beer?

BROOKE GLADSTONE:: This is On the Media from NPR.

copyright 2005 WNYC Radio