The current debate about terrorism, including its implications for civil liberties (which were recently highlighted by the law lords' denunciation of British ministers for imprisoning foreign nationals without trial), is vitiated by a failure to distinguish between two types of threat. One type I shall call old-style terrorism - with which we have been familiar for decades. It was practised by groups such as the Red Brigades in Italy or Baader-Meinhof in Germany in the 1960s and 1970s, but it is more commonly associated with nationalist struggles, usually involving "nations without states". It is the sort of conflict associated with Northern Ireland, the Basque country, Quebec or Kashmir.
Such terrorist activity is linked to specific, local objectives. The level of violence is often relatively low - more civilians died on the roads in Northern Ireland during the Troubles than from terrorist violence. Old-style terrorism can become more violent and destructive where it shades into something closer to civil war, as in Israel/Palestine or Sri Lanka.
Old-style terrorism dates back at least as far as the rise of the modern nation state in the 18th century. Globalisation has changed its nature. It can now have far-flung networks of support and finance: the IRA got funding from sympathisers in the US, Libya and other countries, and established links with terrorist groups and guerrillas in central and southern America.
New-style terrorism, however, is directly a child of the global age. Globalisation - meaning the growing interdependence of world society - is not just about the spread of markets and the increasing influence of cross-border financial institutions. It is driven primarily by the development of instantaneous electronic communications and mass transportation; and it is political and cultural, as well as economic.
Al-Qaeda and its activities form a prototypical example of new-style terrorism, but al-Qaeda is by no means the only group of its kind. Mary Kaldor, of the Centre for the Study of Global Governance at the London School of Economics, has described al-Qaeda as being run in some ways as if it were a non-governmental organisation such as Oxfam or Friends of the Earth. The analogy cannot be stretched too far because NGOs are open and legitimate, while new-style terrorist groups are covert and illegal. Yet the similarities are striking. Both al-Qaeda-type organisations and NGOs are highly decentralised. They are loose networks, driven by a sense of mission that holds together disparate groups or cells across the world, which quite often act semi-autonomously. Both deploy up-to-date communications technologies to co-ordinate their actions and to promote their messages. They have home bases, but these are in principle mutable and diverse. The home bases of new-style terrorist groups are in failing states, but they may also get covert support from governments elsewhere.
The goals of the new-style terrorists are not local, but more far-reaching. Al-Qaeda's aims, as expressed in Osama Bin Laden's proclamations, are truly geopolitical. Bin Laden wants to see the "Crusader-Jewish alliance" driven out of Arabia, and ultimately proposes the re-creation of a caliphate running from Pakistan through to North Africa and southern Spain.
This is the first difference between new and old terrorists. The aims of the former are wider, but also vaguer, than those of the latter. That is why negotiation, much less settlement, is generally impossible. The second important difference concerns organisational capacity. New systems of communication, allied to rapid travel, allow groups to organise at a distance and to co-ordinate terrorist actions. The third difference is ruthlessness. New-style terrorists are prepared to kill thousands, even millions.
The fourth difference is weaponry. The internet allows anyone with the requisite technical knowledge and resources to develop weapons of high destructive potential. And following the collapse of the Soviet Union, large amounts of weaponry are circulating through the illegal global arms trade. Thousands of nuclear weapons and stocks of weapons-usable materials are held in insecure silos in Russia, vulnerable to theft by those who might sell them to terrorists. Enough nuclear material to build 20 medium-sized nuclear weapons is known to have gone missing in Russia.
The combination of organisational capacity and ruthlessness made 9/11 a landmark in the history of terrorism: it was the first successful large-scale attack on the US mainland since the British invasion from Canada in 1814; and it could have been even more destructive than it was. The hijacked planes were aimed at the three centres of US power: financial, military and political. If the twin towers had collapsed straight away, 50,000 people, instead of 3,000, could have been killed. Flight 77 would have done more damage if it had hit the Pentagon more centrally. And the fourth plane - aimed at either the Capitol or the White House - fell short of its target thanks to the bravery of passengers on board.
It is now commonly argued, especially on the left, that governments overhype the risks from terrorism, and that al-Qaeda is something of a paper tiger. This view was put forward last year, for instance, in two programmes in the BBC series The Power of Nightmares, directed by Adam Curtis. Al-Qaeda, Curtis argued, does not really exist: western intelligence agencies and politicians have turned a few, widely scattered terrorist incidents into a sinister global conspiracy. Two US experts on terrorism, Adam Dolnik and Kimberly McCloud, similarly claim that it is time we "defused the widespread image of al-Qaeda as a ubiquitous, super-organised terror network and call it as it is: a loose collection of groups and individuals that doesn't even call itself al-Qaeda". According to them, Osama Bin Laden never spoke of "al-Qaeda" before 9/11; it had only cropped up, in a marginal way, in the pronouncements of those around him. US intelligence agencies first used "al-Qaeda" in a generic fashion after the 1998 embassy bombings in Kenya and Tanzania. Only later was the title appropriated by the terrorist groups themselves.
It is indeed mistaken to suppose that al-Qaeda is a highly organised global machine capable of wreaking havoc wherever its leaders decide; rather, it is a network of various types of group, some more effective and menacing than others. Yet it is equally mistaken to underplay the threat from these. After all, 9/11 did happen - and it involved careful planning and sophisticated logistics.
Bin Laden and his immediate associates are almost certainly less of a danger than they were - because of the military intervention in Afghanistan, not because their capacities at the time were exaggerated. Their warriors were forced to leave their bases and many were killed in the fighting. At least one top commander, Muhammad Atef, died in the air strikes. Three other senior leaders were captured in Pakistan, including Khalid Sheikh Mohammed, who was allegedly the prime figure behind the 9/11 attacks.
Al-Qaeda probably retains "sleepers" in western countries, over and above the cells broken up by police in the US, UK, France, Spain, Germany and elsewhere. There are other terrorist organisations that share the same fundamentalist outlook as al-Qaeda and have connections with it, such as the Egyptian Islamic Jihad, the Islamic Army of Aden (which originates in the Yemen) or the Islamic Movement of Uzbekistan.
The recent report on homeland security in the United States, produced by a task force guided by Richard A Clarke, distinguishes three "concentric circles" of international jihadist terrorism. Al-Qaeda is the "inner circle", with an estimated membership today of between 400 and 2,000 activists. In the second circle are numerous further terror groups, with between 50,000 and 200,000 members. The outer circle comprises jihadist sympathisers in the global Islamic community, numbering upwards of 200 million.
Those who play down the importance of terrorist threats, or who speak of a gratuitous "politics of fear", are apt to say: "There hasn't been another 9/11; we were told that a terrorist attack on London was almost inevitable, but nothing has happened. Why were you scaring everybody without good reason?" This view is specious, for two reasons. The first concerns the phenomenology of risk. In order to get people to take a risk seriously and respond in the right way, they have to be told how potentially dangerous it is. If the threat does not materialise, those who spoke of it are likely to be accused of scaremongering. But scaring people - getting them to see that the risk is real - may be the very condition of minimising or avoiding danger. One weekend in 2003, the Prime Minister, after receiving intelligence information, decided to surround Heathrow with troops. He was criticised for scaremongering. Yet his actions may have stopped a terrorist attack.
The second reason why it is wrong to talk of the "politics of fear" is specific to new-style terrorism. It is simply the magnitude of the dangers involved. The chances of a terrorist attack on London that kills thousands may be slight, especially considering the precautions in place. It is now difficult to implement attacks using planes. It is hard to deploy chemical and biological weapons on a large-scale basis against civilians. A "dirty bomb" - almost certain to be set off some time, somewhere - would cause widespread panic, but its lethal range is limited. There is no evidence so far of nuclear weapons or materials falling into the hands of new-style terrorists, although no one really knows.
However, the truth is that a highly destructive strike is no longer just a hypothetical possibility. The consequences of old-style terrorism may be horrific, but can be weathered. The same is not true of a large-scale terrorist attack: just one episode could be devastating, and one involving a nuclear device cataclysmic. The IRA's warning, after the bombing of the Conservative Party conference in Brighton in 1984, that "we only have to be lucky once; you will have to be lucky always" has a new and more acute meaning in the age of new-style terrorism. We have to avoid even a single successful attack; but the longer we do so, the more, perversely, critics will proclaim that we are all being scared unnecessarily.
The left has to adjust its attitudes towards terrorism, just as it had to adjust them towards crime. It won't do to say that there are no serious threats. It won't do to blame the troubles of the world on George W Bush or the Iraq war. It is no good pretending that there aren't problems in reconciling civil liberties with adequate protection and security. It is not wrong to say that we have to deal with the social conditions that have helped to produce new-style terrorism - poverty and unemployment, schisms between the Islamic world and the west, the situation in Israel/Palestine. And it is certainly right to say that we need urgent measures to halt further nuclear proliferation.
But as with crime, we cannot think only of the underlying conditions. New-style terrorists are by no means always drawn from the ranks of the dispossessed; and their aims, as in the case of al-Qaeda, may be primarily religious and strategic. We have to respond to the dangers they pose in the here and now.
Anthony Giddens is a former director of the London School of Economics