LONDON: With the release of the 918-page report of the Iraq Survey Group, many people are again asking how the intelligence agencies in Britain and the US could have got it so wrong.

Why did they express confidence in a proposition - that Iraq had ample weapons of mass destruction - without a scrap of evidence?

I think the answer turns on how the issues came to be framed. This is important for future intelligence analyses, as well as for historians.

In the UK, the Butler report echoed the Hutton inquiry that exonerated the government from charges of lying but concluded that the joint intelligence committee might have been unconsciously influenced by propinquity to 10 Downing Street; in the US the Senate intelligence committee blamed "collective group think" and a shortage of human intelligence assets.

"Group think", a lack of "humint," "unconscious political pressures" - all seem plausible candidates for blame. But how could these phenomena account for the fact that the intelligence agencies of France, Germany, and Russia concurred in the erroneous assessment, even as their political masters swerved away from the policy of regime change? They can't all have been unconsciously pressured, as their governments disagreed.

I believe this intelligence failure can be traced to our ideas of what constitutes proof, and what happens to an issue when the "burden of proof" shifts, asymmetrically. Many people think of proof - at a trial, for example - as a fact that stands apart from judgement. But there is no fact that determines a murderer (as opposed to the killer), or the existence of God, or whether someone loves you. It is always and only a matter of judgement applied to proof, and in Iraq judgement had been decisively structured by earlier intelligence failures.

In the summer of 1995, UN inspectors were only weeks away from a report giving Saddam Hussein a clean bill of health on the charge of pursuing WMD. This report was important because investigations in 1991 after the first Gulf war had stunned western intelligence agencies with the discovery that Iraq was just a year from deploying a nuclear weapon.

The significance of this intelligence failure could have been catastrophic; it would have meant that Iraq possessed a counter-deterrent, a capacity to preclude a force like the Gulf war coalition, should Saddam make a new lunge for his neighbour's oilfields. So one can imagine the consternation when Saddam's son-in-law, Hussein Kamal, defected and brought knowledge of new clandestine WMD programmes.

Kamal's debriefing left no doubt that the UN inspectors and intelligence agencies had been fooled; he even named the person running Iraq's deception operations. And he stated that deployable Iraqi stocks of WMD had been destroyed on his orders.

These revelations had a decisive effect on the intelligence agencies. In the search for evidence inside a totalitarian society, they had twice been fooled. Given the stakes - preventing Iraq from dominating the oil region with nuclear weapons - they stood on the brink of an historic failure.

The effect was to shift the burden of proof from the requirement of evidence of the presence of WMD programmes to that of their permanent dismantlement. It was now necessary to find convincing evidence not just of the destruction of stocks, but that no clandestine programmes were under way. But how could the intelligence agencies disprove the proposition that somewhere Saddam was busily working at a WMD programme, especially in light of an avalanche of intercepts that was evidence of calculated, ongoing deception?

Historians will find that there were many complex judgements to be made before the authorizing of action to change the Iraqi regime. Some depended on assumptions about the interim period before a legitimate regime was in place. It has not gone well.

It may be that here, too, the decision procedure - preparing for only one scenario, mistakenly regarded as the likeliest - holds the key to our failures. Paul Bremer's statement this week that far too few troops were provided seems to confirm this conjecture.

Decisions about the future can seldom be settled by facts in the present. Suppose Hans Blix established that there were no WMD. The ISG report, based on interviews with Saddam's inner circle, concludes he would have resumed a nuclear programme once sanctions ceased.

If Blix had found weapons, that would have been decisive; but there is no fact, discoverable by inspectors or anyone else, that could assure us that Saddam would not simply use the $11bn we know was skimmed from the oil-for-food operation to buy nuclear weapons once the inspectors left. These are, ultimately, matters of judgement. -Dawn/The Guardian News Service.

(The writer is a law professor at the University of Texas at Austin and a former director for intelligence programmes at the US national security council).