FOR ONCE, the importance of an arms-control deal is not being overstated. But last week's INF agreement is important for a much greater reason than the withdrawal of a few hundred warheads. It marks, I believe, the beginning of a "post-nuclear" era, whose different strategic dangers are being ignored as we argue over the details of verification.

Thousands of nuclear weapons may remain in this post-nuclear world. But they will no longer provide a realistic option for defending Europe. Nor will they offset the weakness of our conventional forces elsewhere in the world. For this reason, we may have reason to regret the passing of the nuclear era, for all its obvious dangers.

The decline in the importance of nuclear weapons does not stem from the development of other, more powerful weapons. Nor does it reflect the recent growth in Soviet nuclear capabilities. Rather, it is a continuation of trends that have been evident for 20 years, ever since NATO replaced the strategy of massive retaliation with one of flexible response. This was the first great reduction in the role of nuclear weapons within the U.S.-Soviet military balance. Nuclear weapons were now to be kept in reserve in the event of an imminent non-nuclear defeat.

Interestingly enough, flexible response was imposed on NATO at the very time when the superiority of U.S. intercontinental nuclear forces was at its peak, by every criterion of measurement. The Polaris forces of 41 submarines with 656 ballistic missiles had just been completed and the United States possessed a force of 1,000 Minuteman missiles as against far fewer and much inferior Soviet sea- and land-based ballistic missiles. The U.S. advantage in manned bombers was even greater. If the balance in the respective destructive capacities had been the determining factor, 1967 should have witnessed a reaffirmation of massive retaliation rather than its final repudiation.

U.S. and NATO strategies have not changed officially in the past 20 years. But the role of nuclear weapons has relentlessly declined, and spending for non-nuclear forces has increased sharply. The question is: Why?

Nothing has changed in the physical realm, but everything has changed because mentalities have changed. Public perceptions of the threat inherent in Soviet military power as unlimited in scope (a threat to "our way of life") have gradually given way to a less-threatening geopolitical conception -- in which such-and-such Soviet forces are seen as capable of invading this-or-that theater of war. Accordingly, the response deemed appropriate these days is also more discrete.

The drift to de-facto denuclearization has exposed the political implausibility of the various U.S. nuclear guarantees, which are intended to dissuade foes from conventional attack by threatening an American nuclear response. Consider some examples:

Until 1985, Quemoy and Matsu -- islands claimed by Taiwan off the Chinese mainland -- were to be held by nuclear counterattacks against non-nuclear Chinese artillery barrages under joint U.S.-Republic of China defense plans. At this writing, the notion that the United States would use nuclear weapons to defend Quemoy and Matsu seems absurd, but it was not so for the U.S. officials who approved those plans in the 1950s -- though they themselves might now look back in disbelief at their thoughts of those days.

The denuclearization drift probably has progressed far enough to rule out the use of nuclear weapons for the defense of South Korea against a non-nuclear invasion from the north. That was still seen as a realistic contingency in the 1960s, and a definite reliance on nuclear weapons was not eliminated from U.S.-Republic of Korea defense plans until well into the 1970s.

Japan is still to be defended by nuclear weapons if its non-nuclear defense fails. Public discourse on the "nuclear umbrella" mostly refers to the dissuasion of a Soviet nuclear attack. But that may not be the real threat, given the acute vulnerability of Hokkaido -- the northernmost Japanese island -- to a non-nuclear invasion from nearby Soviet territory. It is hard to believe that a U.S. president would authorize nuclear attacks against Soviet forces in response to a non-nuclear invasion of Hokkaido, or that such attacks would be acceptable to public and congressional opinion.

Europe remains the great question mark. But the trend is clear enough. There, as everywhere else, the political plausibility of extended deterrence against conventional threats is diminishing. The apparent U.S. readiness to use nuclear weapons against a non-nuclear invasion was greater in the past than it is now, and it is almost certainly greater now than it will be in the future.

The best index of denuclearization at any one time is the size, quality and cost of the non-nuclear armed forces that the United States and the European members of the Atlantic alliance see fit to maintain. In the 1950s, the various potential fronts -- in northern Norway, Italy and eastern Turkey, as well as the central front in Germany from the Baltic to the Austrian border -- were held only by scattered units to be deployed in long, thin lines. They provided a very poor defense against concentrated armored thrusts, but they were ideal to trigger the nuclear bombardment of the Soviet Union, which was the envisaged response to invasion at that time. Since then NATO ground forces have grown in size and armament and their logistic capacity for sustained combat has increased.

We should recognize that the very fragility of the alliance defenses of the 1950s added to the credibility of massive retaliation. Similarly, the continuing increase in the capacity and sustainability of NATO combat capabilities not only marks but also promotes the continuing advance of denuclearization. That's because the "robust" defenses that we are forever asking our European allies to provide would merely slow a Soviet invasion long enough to allow European parliaments and Congress time to debate whether nuclear weapons should be used. And that debate would produce only one answer: the exclusion of any use of nuclear weapons.

In the hands of a Hitler or Pol Pot, nuclear weapons could certainly retain a larger role in overall military balances than has ever been the case for U.S. nuclear weapons -- because adversaries would believe that such leaders would actually use them, regardless of the consequences. Moreover, in potential nuclear confrontations, Stalin and Mao conceded much less to the risk of nuclear reprisals than many other leaders might have done. Stalin ventured the 1948 Berlin blockade, and Mao attacked U.S. forces in Korea on the largest scale in the winter offensive of 1950.

Soviet threats to use nuclear weapons since the early 1950s may have been more credible than American, because the Kremlin has been perceived as less inhibited. But Soviet nuclear capabilities, too, have also been broadly circumvented. Pakistan, for example, has not been deterred by Soviet nuclear weapons from supporting a war against Soviet forces in Afghanistan.

The decline of nuclear weapons in U.S. strategy has been replicated by our friends and allies. Britain once had its own policy of massive retaliation (outlined in a 1957 white paper), under which the conventional forces were drastically reduced. Thirty years later, the British nuclear force is once again a small component of predominantly non-nuclear military forces. British nuclear capabilities certainly were wholly circumvented in the Falklands-Malvinas conflict.

China is another nuclear power that recently has fought a post-nuclear war -- against Vietnam -- and rather unsuccessfully to boot. Facing the nuclear might of the Soviet Union without the nuclear guarantee (albeit eroding) that NATO enjoys, only an imprudent Chinese leadership would depart from Beijing's present policy.

Finally, the Soviet Union still keeps tactical nuclear-delivery means (if not the warheads themselves) within the unit structure of every army division and once had a published doctrine of prompt nuclear use. But the Soviets' huge and indeed unprecedented accumulation of non-nuclear combat forces is the best possible indicator of non-nuclear intentions. That, of course, is what makes the U.S. drift toward denuclearization possible -- and also so difficult to reconcile with continued security in post-nuclear conditions.

The danger for the West is that the Soviet Union may be better prepared for the post-nuclear era than we are.

In Europe notably, without recourse to nuclear firepower, we will have difficulty countering either the Soviet army's coercive potential or its actual invasion capacity. Checking the Soviets in a non-nuclear framework will require a fundamental change in NATO forces, whose purpose since the 1940s has been to supplement the various forms of nuclear protection. Their composition, deployment and institutional priorities are all ill-suited to sustained non-nuclear defense. The same is true of the armed forces of the United States.

We should not expect, therefore, that we can easily "fix" the conventional balance, in Europe or outside it, by adding new high-technology weapons or by other incremental enhancements. Instead, a complete post-nuclear audit would be needed to uncover a myriad of tacit nuclear-use assumptions. Each then would require structural or operational remedies, ranging from the decentralization of stores and repair depots (now all highly centralized for efficiency under the nuclear assumption of all-or-nothing bombardment) to the restoration of predominantly reservist ground armies to oppose similar Soviet forces.

We will also have to change the way we think -- and talk -- about conflict. For today's military officers, who have grown up in the nuclear era, "sustained" combat implies weeks, not years; "mobilized" forces implies the augmentation rather than the multiplication of the standing armies; and "large-scale" implies dozens of divisions at most, not hundreds.

The emerging post-nuclear era also means that our current arms-control and nuclear-weapon-acquisition policies are becoming directly contradictory.

In the past overall nuclear-force goals could be legitimately pursued by acquiring some new capabilities while limiting other avenues of development by agreement or unilaterally. What mattered was to maintain an advantageous balance between nuclear acquisitions and nuclear limitations. Now, in contrast, arms-control measures that deliberately accelerate denuclearization must be balanced by conventional-force acquisition, or by other, non-nuclear arms-control measures that reduce operational imbalances in the various theaters of interest.

As for nuclear weapons, the issue is no longer their efficiency or compatibility with arms-control goals -- but rather their very purpose. If we accept the advent of post-nuclear conditions, then the complex of current nuclear programs can no longer be justified. If we intend to resist denuclearization as much as possible, however, then our policy must change drastically, and the priority assigned to defensive as well as offensive "strategic" nuclear capabilities perhaps should be increased.

The worst outcome would be to slide into a post-nuclear world with nuclear forces that are largely irrelevant and non-nuclear forces that are structurally inadequate. A fundamental decision at the level of grand strategy is thus required to determine whether the strategic decline of nuclear weapons is to be resisted or accelerated. Only then can congruent arms-control and military policies be formulated. At this stage, however, even the nature of the problem has yet to be recognized in its full strategic implications.

This essay is adapted from a longer version that will appear in The Washington Quarterly.

Edward Luttwak holds a chair in strategy at the Center for Strategic and International Studies. He is the author of "Strategy: The Logic of War and Peace."