South Korea's foreign minister predicts that Pyongyang will soon return to the six-party nuclear talks, which Beijing says are slated to begin this coming Thursday, but the United States doesn't seem so sure. After reports that North Korea was preparing to conduct a second nuclear test, Washington sent a squadron of stealth fighters to South Korea-supposedly a "routine deployment." Still, talk of a renewed nuclear freeze is wafting through Asian capitals, even as a British newspaper charges that North Korea has been aiding Iran's nuclear program.

Maybe the United States and its allies will strike an accord with Pyongyang next week, but North Korea has been dashing the hopes of Western diplomats for years. With military pre-emption seemingly off the table, despite the arrival of the F-117 Nighthawk fighters, what can be done if the North forges ahead?  The conventional wisdom is to strengthen and extend America's nuclear umbrella.

Umbrella Proliferation

Indeed, Secretary of State Condoleezza Rice responded to recent talk about possible development of a countervailing Japanese or South Korean nuclear weapon by flying to East Asia.  She declared:  "It's extremely important to go out and to affirm, and affirm strongly, U.S. defense commitments to Japan and to South Korea." Those promises were understood to be nuclear.  Tokyo, in particular, responded by disclaiming any interest in going nuclear.

Although America's nuclear umbrella for Japan dates back to the end of World War II, the United States has not limited nuclear guarantees to historic allies. In order to convince Ukraine to disgorge the nuclear weapons that remained on its territory after the break-up of the Soviet Union, Washington reportedly provided Kiev with some security guarantees. Whether they include a promise to use nuclear weapons against Russia on behalf of Ukraine has never been revealed. In any case, Kiev may have given up its ultimate deterrent in the belief that Washington was offering an implicit commitment.

Moreover, Jim Hoagland of the Washington Post wants America to provide nuclear guarantees for the Middle East.  He writes: "Bush should announce that he wants consultations with Saudi Arabia, Kuwait, Egypt, Jordan and other Arab states-as well as principal U.S. allies in Europe-on extending a U.S. or NATO nuclear umbrella over friendly states in the Gulf."  This would, Hoagland contends, "enable Arab states to forgo developing their own nuclear programs, just as the U.S.-Japan bilateral security treaty is intended to keep Japan nuclear-free."

It's one thing to promise to respond to a nuclear attack by a potential global hegemon, the Soviet Union, against a major ally, such as Germany or Japan, especially when Washington has deliberately disarmed them. Very different is to promise to protect Jordan or Kuwait, friendly countries, true, but neither historic nor important allies, against an attack by Iran, a regional power without global reach.  The latter is an extraordinary extension of a doctrine fraught with danger.

The Tripwire Doctrine

The principle behind extending Washington's nuclear umbrella is deterrence. That is, smaller nations, even if evil or aggressive, will not risk American retaliation by threatening friendly states.  Moreover, friendly states, sheltered behind a U.S. guarantee, will avoid taking steps opposed by Washington-most particularly, constructing their own nuclear weapons.

Undoubtedly, security commitments help deter. The possibility of U.S. intervention raises the cost of war, and thereby discourages aggression. If aggression is less likely, then so is the likelihood that countries will adopt extreme defensive measures. Advocates of extended security commitments, and particularly nuclear guarantees, emphasize these effects.

However, though a military guarantee may help deter conflict in this way, it makes conflict more likely in other ways. First, if the U.S. commitment is not credible, there is no deterrent effect. Even a written treaty may not be enough. The famous Chinese challenge-you won't risk Los Angeles to protect Taipei-suggests some doubt in Beijing that the United States would pay the potential price of confronting a nuclear power in order to protect a peripheral geopolitical interest.

Second, if war erupts, U.S. involvement (assuming America makes good on its promise) is automatic. Washington loses the ability to weigh costs and benefits in the particular case at the particular time. For decades the quintessential example of this policy was the forward deployment of U.S. forces in Korea, the so-called "tripwire" that ensured sufficient American deaths in any North Korean invasion to trigger U.S. involvement.  That policy may have reduced the likelihood of war breaking out, but only by ensuring U.S. involvement in any conflict.

Even a small risk of war would be extraordinarily dangerous when dealing with nuclear-armed states. Confronting China, which has global ambitions, or even Iran or North Korea, assuming they develop a capacity to hit the United States, would be far different than attacking Serbia or Iraq. It would be tragically ironic to survive the Cold War without a nuclear exchange and then blunder into one by intervening in a small conflict of limited importance.

Third, offering to lend America's military to a friendly nation reduces the latter's need to develop its own defense and foster its own alliances. This perverse impact of U.S. defense promises and deployments is evident in East Asia today. The primary example is Japan, which only now, six decades after the end of World War II, is debating a more active defense and foreign policy that is commensurate with its abilities and interests. South Korea's behavior has been only slightly less extreme: despite possessing a dramatic economic edge over Pyongyang, the South continues to under-spend on defense because it can count on Washington's protection.

Despite Chinese pressure for reunification, Taiwan, which trusts in somewhat ambiguous American defense guarantees, has allowed spending on its military to slip. The Philippines pushed for a new visiting-forces agreement to strengthen military links with America in part because it is locked in a long, rancorous dispute with China over the Spratly Islands. Yet Manila essentially has a navy that won't sail and an air force that won't fly.

Fourth, providing a security guarantee is likely to change the behavior of the client state. That is, of course, why Secretary Rice visited East Asia:  she wanted to use America's nuclear umbrella as a means to dissuade Japan and South Korea from developing their own nuclear deterrent. That is why Jim Hoagland wants to provide a similar guarantee to numerous Middle Eastern states:  there have been mutterings about Egypt and Saudi Arabia, and perhaps other countries, pursuing a nuclear option if Iran develops such weapons.

Unfortunately, there is no reason to believe that this will be the only effect of the guarantee. A superpower's promise to defend a nation creates potentially wide-ranging and dramatic incentives.  If Taipei, for instance, is convinced that Washington will come to its aid-even willing to threaten nuclear war in Taiwan's defense-then Taiwanese officials are likely to approach China very differently.

The risk is two-fold. Most bilateral relationships involve some compromise and give-and-take. A small nation acting with the full faith and credit of the United States might refuse to make even modest accommodations to a powerful neighbor, irrespective of the legitimacy of the latter's claims. Worse, the allied state might be willing to take enormous risks, far greater than it would accept if on its own militarily, to advance interests that might be in conflict with those of the United States.  Germany's famous "blank check" to Austro-Hungary in July 1914 demonstrates the attendant risks.

The most obvious case today is Taiwan. The vast majority of Taiwanese prefer an identity separate from China. However, since a formal declaration of independence could provoke Beijing to act militarily, most Taiwanese have been willing to accept today's uneasy status quo.  However, that caution could disappear if they believe that U.S. threats and actions would deter the People's Republic of China from responding.

Or imagine Washington attempting to dampen the possibility of a nuclear exchange in South Asia by extending a nuclear guarantee to both India and Pakistan. Would it reduce the pressure of both nations to push the nuclear button? Perhaps. However, such a step almost certainly would make both powers more intransigent, creating increased risks of conflict.

Of course, American diplomats recognize this danger and have, for instance, insisted that Taiwan not provoke the PRC. What if Taiwan believes the United States will act irrespective of the former's responsibility, however? Would Washington, having promised to aid its long-time friend and ally, really leave the latter to its fate, offering legalistic quibbles as its reason? Client states, especially democracies where public opinion matters, are never as easy to control as dominant powers wish.

"Desperate times demand desperate measure", writes Jim Hoagland. But not recklessness. No where is sober judgment more necessary than in designing nuclear policy.

Nuclear proliferation creates enormous risks and therefore is worth enormous effort to thwart, especially in such unpredictable countries as Iran and North Korea. How to respond if the attempt fails is far less clear. For decades the conventional wisdom endorsed the doctrine of extended deterrence. The Bush Administration refuses to reconsider this strategy as Iran and North Korea threaten to become nuclear weapons states. As nuclear weapons spread, however, such a policy risks entangling America in ever more conflicts where the worst-case outcome is a nuclear exchange. The cure would be far worse than the disease.

Doug Bandow is a former Special Assistant to President Ronald Reagan and the author of several books, including Foreign Follies:  America's New Global Empire (Xulon Press).