We panicked the last time terrorists struck, and we will panic the next time. The 11 September 2001 and 7 July 2005 attacks were pinpricks compared to the destruction that could be caused by an atomic bomb in a suitcase or by an anthrax epidemic. The next major attack may kill tens of thousands, and if it does it will cause a political tidal wave likely to leave behind repressive legislation far more drastic than anything we have yet seen.
We are in a downward cycle. After each terrorist attack, politicians come up with a raft of new repressive laws that ease our anxiety by promising greater security - only to find that a different terrorist group strikes in a new way a few years later. This new disaster creates in turn a demand for more repression, and on and on. Even if the next half-century brings only three or four attacks on the scale of 9/11, the cycle will prove devastating to civil liberties.
It is a dismal prospect, and the time to confront it is now - before the next attack. We need to think ahead in an atmosphere of relative calm and ask ourselves how we should react to future crises without becoming ever more authoritarian, without permanently compromising our civil liberties; without, in other words, destroying many of the things that make our society worth defending. What I propose is that we agree, now, on an "emergency constitution" that could come into force at times of extreme danger, and which would give politicians the means to respond to that danger but prevent them from exploiting panics in ways which cause lasting damage.
I do not reach this conclusion lightly. The British constitution, and that of the United States, have proved their capacity to cope with crisis and it is right that we should pause - and pause again - before taking so important a step. Nevertheless, the recent rounds of repressive legislation in both countries should serve as a catalyst for reflection. Above all else, they suggest that we should make a clear distinction between short-term and long-term responses to terrorist attacks.
My proposal is this: that prime mini sters, and presidents, should be granted broad emergency powers in the immed iate after math of a major attack, but that these powers should last only for a week or two unless a major ity in par liament expres sly authorises their extension for a two-month period. After those two months the government would have to return to parliament for a further authorisation, and this time a majority of 60 per cent of MPs should be required; after two months more, the majority should be set at 70 per cent; and then 80 per cent for every subsequent two-month extension. Except in the most extreme circumstances, this "escalator" would terminate the use of emergency powers within a relatively short period.
I am seeking here to develop ideas and practices that are already in use. Temporary emergencies are often declared in response to natural disasters, as the Hurricane Katrina episode reminded us. What is more, South Africa already has an escalator of majorities built into its constitutional arrangements: any state of emergency in that country must be renewed at three-month intervals by a vote of at least 60 per cent of members of parliament. South Africans had a bitter experience with "emergencies" under the apartheid regime. They have learned a constitutional lesson from this experience, and so should we.
The free market in death
South Africa operates a parliamentary system, suggesting that an emergency constitution is compatible with the basic premises of the Westminster model. To be sure, the British constitution would theoretically permit a prime minister to repeal the emergency constitution by a simple majority vote, but this prospect should not be taken too seriously as a prac tical matter. Operationally, the "supermajority" requirement gives opposition parties a veto over an extension of the state of emergency, and any panicky effort at repeal would be bound to generate their passionate protests, endangering national unity at a moment the prime minister could ill afford. The political costs of repeal would rise further once the emergency constitution began to prove its value in practice, and the country came to appreciate how it served as a bulwark against episodic panics.
Defining the scope of emergency power is a sensitive business. At its core, it involves the short-term detention of suspected terrorists to prevent a second strike. Nobody should be detained for more than 45 days, and then only on reasonable suspicion. Once that time has elapsed, the government must satisfy the standards of evidence that apply in ordinary criminal prosecutions. And even during the period of detention, judges should be authorised to protect suspects against torture and other abuses.
Why two-month extensions to the emergency powers, and not three? Why 45 days of preventive detention, and not 60? Such matters are obviously open to debate. My point is that an arrangement of this kind is necessary, and we can have it only if we start discussing it now.
Let us look at the alternative. Words are the lifeblood of constitutional life, and we are off to a bad start when we talk of a "war on terror". This war talk is preparing the way for terrible abuses after the next attack. Terrorism is a technique - the intentional attack on innocent civilians - but war isn't a technical matter: it is a life-and-death struggle against a particular enemy. We made war against Nazi Germany, not the V2 rocket. Once we allow ourselves to declare war on a technique, we open up a dangerous path, authorising the government to lash out at amorphous threats without the need to define them. By calling it a war, we also frame our problem as if it involved a struggle with a massively armed power. But modern terrorism has a very different genesis: it is more a pro duct of the unregulated market place than one of state power.
We are at a distinctive moment in history: the state is losing its monopoly over the means of mass destruction. And once a harmful technology escapes into the black market it is almost impossible for governments to suppress the trade. Think of drugs and guns. Even the most puritanical regimes learn to live with vice on the fringes. But when a fringe group obtains a technology of mass destruction, it won't stay on the fringes for long. The root of our problem is not Islam or ideology, but this free market in death. If the Middle East were transformed into an oasis of peace and democracy, fringe groups from other places would fill the gap. Osama Bin Laden wasn't behind the bombing of the Alfred P Murrah Federal Building in Oklahoma City - this was the work of Americans. Every country has such people, and they will do all they can to get hold of suitcase A-bombs.
This problem is not illuminated by war talk. Even the greatest wars have come to an end - when Winston Churchill or Franklin D Roosevelt asserted extraordinary war powers, everybody recognised that they would last only until the Axis was defeated. But the black market in weaponry at the source of the "war on terror" will never end. Whatever powers are conceded to the government in this war, it will have for ever.
With terrorism, everyone acknowledges, even the best preventive measures will sometimes fail. The question is not: Can the security services prevent another attack? It is: How often will they slip up - once in every ten threats, once in a hundred, once in a thousand?
This, too, is obscured by war talk. Real wars don't come out of nowhere because the government has slipped up. They arise after years of highly visible tension between sovereign states, and after the failure of countless efforts at diplomacy. Even Pearl Harbor was preceded by years of rising tension. But when terrorists strike we do not know, at first, whether we face a tiny group of fanatics or a serious organisation with real staying power. War talk leads us down a misleading path suggesting that not only are "the terrorists" numerous and well organised, but that they have powers comparable with those of nation states. Above all, war talk invites us to suppose that we should give our government powers that might be appropriate when fighting a Third World War.
Prevention is better than war
My idea of an emergency constitution is predicated on a more accurate description of our situation. We are reeling from a surprise attack and we don't know whether the terrorists were just lucky, whether they have the capacity to organise a second strike, or whether they are in it for the long haul. So let's buy time. The short-term problem is the second strike, so we should grant our government the extraordinary powers needed to pre-empt it. This is the danger of the moment and we should focus our energies on preventing it, and not on launching a never-ending war on terrorism.
This is precisely the point of the escalator of majorities. While the country might go on emergency alert for two months, or even six, the escalator ensures a return to normality if the security services manage to disrupt the conspiracy, or if the terrorists prove to be a passing threat. Without a constitutional framework of this kind, leaders will inevitably respond by calling on us to sacrifice more and more of our freedom in the effort to win an unending "war".
Full and free debate is a liberal society's best hope against extremism, and it must be protected in moments of crisis. The emergency constitution should hold firm to fundamental legal doctrines forbidding prosecution for advocacy of obnoxious ideas. People should be punished for what they say only when their words have an instrumental role in concrete criminal conspiracies.
The government will face a great temptation to abuse its emergency powers by restricting standard political activities such as large demonstrations, but this would not only damage the checks and balances on which the emergency constitution relies - it would also undermine the larger project of a speedy return to normality. Rather than repressing political expression, the emergency authorities should be obliged to give demonstrations of political opinion special security priority.
Another important point concerns the treatment of the principal victims of the emergency regime - the suspects swept up into emergency detention. Most, we know from experience, will turn out to be innocent. Even if released after 45 days, they will have been done grievous harm. Six weeks may be a short time, but it is long enough to disrupt a life, to create enormous anguish, and to stigmatise detainees in the wider community. The emergency constitution should take aggressive steps to remedy these harms. Detainees should be legally guaranteed their previous employment, and each innocent suspect should be provided with a substantial payment for every day his life has been disrupted for the greater good - say, £250 a day. This is elementary justice, but it would also have a desirable impact on the bureaucratic mind: given that detentions will be very expensive, there will be a strong incentive to avoid detaining people needlessly.
These ideas seem strange, so do we really need them? Let us remind ourselves of the British experience of the past few years, which provides vivid evidence of the cycle in which terrorist attacks repeatedly generate permanent losses of liberty.
Within three months of 9/11 the Blair government rushed through a statute that made the US's Patriot Act seem mild by comparison. Under current British law, there is no need to accuse a suspected terrorist of a crime to strip him of his liberty. Recent terrorism legislation dispenses with trial by jury and the need to prove guilt beyond reasonable doubt; the government need only convince a judge - not a jury - that the detainee is probably a terrorist. Worse, the suspect can't learn of all the evidence against him. The government can persuade the judge to hide key assertions in a confidential file, which is kept from both the suspect and his lawyer. To safeguard the defendant's interests, the law creates a very small corps of special advocates, with security clearances, who can see the file but can't tell the suspect what they have seen. Little wonder that a couple of these advocates have resigned in disgust: it is awfully hard to defend your client when he can't help you pick holes in the state's evidence.
As compensation for this utter failure of due process, the accused "terrorist" gains one advantage over the typical criminal defendant: he can't be thrown in jail. But he can be confined under house arrest, with his movements and access to the larger world kept under tight control; or, if the government is in a gentle mood, it may merely place him under curfew, or limit his access to means of communication, or deny him contact with other suspects, or subject him to electronic tagging.
This goes on for six months; but the six-month term can be extended again and again, so long as the procedural charade is repeated. These draconian measures were in place before 7 July 2005, yet their existence did not prevent the government from demanding yet another round of repressive legislation. The Blair government's most recent initiative transformed a broad range of political activities into crimes against the state, and while Muslims are the initial targets, the net will be cast more broadly if the next incident involves, say, a hypernationalist group blasting away at some citadel of authority. To be sure, Blair overreached when he demanded authority to detain suspects for 90 days without producing any evidence of guilt, but the back-bench rebellion only obliged him to cut the period of arbitrary detention to 28 days. Blatant violations of due process are becoming a normal part of British law.
If this is the response to a relatively small attack, how will some future government respond to an assault that kills thousands at a single blow? Remember that existing law already authorises continuing house arrest for anybody who is "probably" a terrorist. After the next attack, a panicky public might prove willing to support indefinite detention on the basis of "reasonable suspicion". And why stop at house arrest? Why not throw suspected terrorists into prison, without giving them a chance to establish their innocence? Where will it end?
We may be lucky: perhaps there will be no next attack. Or perhaps, when it occurs, our leaders will refuse to succumb to the political dynamics of fear and repression. But perhaps not. No constitutional design can give us a guarantee against the very worst case, and no constitutional design is needed for the best of all possible worlds. But there is plenty of room in the middle, which is where human beings generally live out their lives. This is where the emergency constitution can make a big difference.
Bruce Ackerman is Sterling Professor of Law and Political Science at Yale University, and the author of the recently published book "Before the Next Attack: preserving civil liberties in an age of terrorism" (Yale University Press, £18.99). This essay is derived from the Ashby Lecture he delivered at Cambridge in May
Terror and UK law
Research by Sohani Crockett
2001 Anti-terrorism, Crime and Security Act: Introduced as emergency legislation after 9/11 to increase the powers of the police to deal with those suspected of planning or assisting terrorist attacks within the UK. Went from introduction to royal assent in 32 days. Key feature: resident foreigners can be detained without trial if they cannot be deported without breaching British human-rights laws
2005 Prevention of Terrorism Act: Allows the Home Secretary to impose "control orders" on foreign and UK citizens he suspects of involvement in terrorism. Removed 790-year-old principle of habeas corpus
2006 Terrorism Act: Drafted in aftermath of 7/7, creating new offences related to terrorism and amending existing ones. Made it illegal to glorify terrorism and distribute terrorist publications; also increased the maximum duration that a suspect could be held without trial from 14 to 28 days