Let's be honest: None of us were in much mood for cool deliberation in the weeks immediately following 9/11. Under the best of circumstances, it would have been a feat to get everything right in a 300-page bill that makes dramatic changes to dozens of complex national security statutes. In the panicked aftermath of a horrific terror attack, it would have been nothing short of miraculous. So you'll have to forgive me if I'm more interested in looking back to see where a better balance could have been struck in light of what we now know.
Let's clear some brush first. Many pundits keep invoking the national security equivalent of Lisa Simpson's tiger-repellent rock: Some doubt the rock's magic powers, but I sure don't see any tigers around! If we actually pay attention to the institutional sources of pre-9/11 intelligence failure, we find that an array of internal structural problems and simple blunders, not some paucity of investigative tools, deserve the lion's share of the blame. Inadequate information sharing and coordination, which could have been substantially remedied even without legislative change, were indeed part of the problem — but I'm mystified by how often some come back to this aspect of the law, which is largely irrelevant to the actual reform proposals on the table. Meanwhile, the successful investigations since 9/11 — only a few of which actually involved "plots" for attacks in the U.S. — appear to have depended a lot more on good human intelligence and informants in the community than broad new surveillance powers.
If we were really engaged in a "fundamental debate" over whether to "squash" Patriot Act authority wholesale, then perhaps an analysis at the magic-rock level would be adequate. But there are many more specific questions: How narrowly should the target description in a roving wiretap application be drawn to avoid issuing broad "John Doe" warrants? What kind of evidentiary showing is appropriate when investigators use national security letters — issued by the tens of thousands each year — to gather financial records or track Internet activity? Does the creation of vast databases of personal information about mostly innocent Americans for the purposes of "pattern analysis" yield an intelligence benefit that justifies the risk to privacy involved? What procedures might avoid the already-evident problem of "mission creep," where extraordinary tools for counter-terrorism efforts end up primarily being used for ordinary criminal cases? What rules should govern the handling of information about innocents?
For the most part, supporters of the Patriot Act have chosen to gloss over these important questions. Take for example one particular reform, though it's not among those still under consideration: Sen. Russell D. Feingold's proposed requirement that national security letters be used to identify terror suspects only when they are the "least intrusive means" to do so. Many worry that this is too "subjective" a standard, and that hapless investigators uncertain about how to apply it will be too cautious. But this is not how it works. The attorney general develops specific guidelines that give content to the more abstract requirements of the statute. Agents are not required to make complex legal determinations on their own, and there is no reason in principle to expect this to be a more serious problem than fleshing out the meaning of "probable cause," or any number of other standards that apply to government search and seizure.
And this brings us to the real bone of contention. The ultimate investigative "flexibility" is a total lack of standards that the Founding Fathers despised. We don't allow this, not only because it is inconsistent with American values but because before we imposed some restraints on intelligence surveillance in the 1970s, we know that domestic spying authority was systematically abused, on a massive scale, for political purposes unrelated to national security. The response of Congress was to impose the very restraints whose dilution you now applaud. The risk that history will repeat itself is transparently real, even if you don't deem the new abuses that have managed to slip through the veil of secrecy sufficiently "dramatic" yet.
What we need is an analysis of the trade-offs. We are so used to talking about the balance of "liberty versus security" that it's easy to forget they're not always in conflict. Some restraints on government power leave us much more free without seriously impeding legitimate investigations.