Did India and Pakistan nearly fight a nuclear war in 1990? In a provocative 1993 article, Seymour M. Hersh claims that they did. During a crisis with India over the rapidly escalating insurgency in Kashmir, according to Hersh, Pakistan "openly deployed its main armored tank units along the Indian border and, in secret, placed its nuclear-weapons arsenal on alert." As a result, "the Bush Administration became convinced that the world was on the edge of a nuclear exchange between Pakistan and India." Hersh quotes Richard J. Kerr, deputy director of the Central Intelligence Agency in 1990, as saying: "It was the most dangerous nuclear situation we have ever faced since I've been in the U.S. government. It may be as close as we've come to a nuclear exchange. It was far more frightening than the Cuban missile crisis." Robert M. Gates, President George Bush's deputy national security adviser in 1990, reportedly told Hersh that "Pakistan and India seemed to be caught in a cycle that they couldn't break out of. I was convinced that if a war started, it would be nuclear."(1)
Hersh's account of the 1990 Indo-Pakistani crisis has acquired the aura of conventional wisdom in both popular and scholarly circles. William E. Burrows and Robert Windrem essentially re-tell Hersh's story and add that "Indian nuclear forces" were also "on alert."(2) Political scientist Scott D. Sagan acknowledges that Hersh's thesis is "unconfirmed," but relies on it nonetheless to help demonstrate the "perils of proliferation."(3) In this article, I present a different interpretation of South Asia's 1990 crisis. My research suggests that a fourth Indo-Pakistani war was indeed a possibility, but that Hersh misinterprets the nuclear dimension of the crisis. I have found little evidence to corroborate his core description of events, much of which appears to be based on interviews with one anonymous U.S. analyst.(4) To the contrary, a number of senior South Asian and U.S. officials have categorically denied Hersh's report. I will argue below that India and Pakistan were deterred from war in 1990 by each side's knowledge that the other was nuclear weapon-capable, and therefore that any military hostilities could have escalated to the nuclear level.
My case study of South Asia's 1990 war scare is framed by several important issues in the scholarly analysis of nuclear proliferation. At one level, it is embedded in a lively debate over the consequences of proliferation extending back to the early 1960s.(5) The orthodoxy in this debate, which I call the "logic of nonproliferation," considers the spread of nuclear weapons to be extremely dangerous: more of them in more hands around the world increases the likelihood of nuclear explosions, either intended or unintended.(6) The "logic of nuclear deterrence" rebuts this perspective, suggesting instead that proliferation has stabilizing effects: nuclear weapons have deterred war between their possessors and will continue to do so.(7) At another level, the case study is theoretically situated in the analysis of "opaque proliferation."(8) This pattern of unconfessed proliferation has characterized every emerging nuclear weapon state since the mid-1960s, but its implications have only recently been appreciated.(9)
With the end of the Cold War, the possibility of a regional conflict escalating to a global conflagration has vanished. Why, then, should the rest of the world care if India and Pakistan fight a nuclear war? The main reason is a humanitarian one. Indians and Pakistanis comprise about 20 percent of the world's population. Many are clustered in heavily populated cities well within the reach of the adversary's nuclear-capable bombers. Even a limited exchange of rudimentary fission weapons would kill millions of people. A second reason is the continued maintenance of the "nuclear taboo" that has developed in world politics since the United States dropped atomic bombs on Japan in August 1945. For five decades, the military use of nuclear weapons has been steadily de-legitimized. Since the end of the Cold War, this process has taken material form, with the draw down of U.S. and former Soviet nuclear forces. Breaking the nuclear taboo would disrupt this positive trend, with potentially grave consequences for international security.
The remainder of this article is organized as follows: in the next section, I outline a theoretical basis for believing that nuclear deterrence can operate between opaque proliferants. First, I analyze the major area of disagreement between the logic of nonproliferation and the logic of nuclear deterrence, which concerns the likelihood of preemptive escalation between new nuclear powers.(10) In doing so, I pay special attention to the influence of opacity on the prospects for crisis stability, a factor whose influence is still underestimated and often misunderstood. Contrary to the conventional wisdom, I argue that the prospects for preemptive nuclear attacks between new proliferants are low, and that opacity makes them even more remote. Then I briefly explore the logical nexus between opacity and "existential deterrence,"(11) and the mechanics of nuclear deterrence between opaque proliferants. The credibility of existential deterrence under conditions of opacity rests on the complementary processes of tacit bargaining and inadvertent "transparency-building" by the international community, which together demonstrate the proliferants' capabilities and resolve. Next, in a study of the 1990 Indo-Pakistani crisis, I argue that New Delhi and Islamabad were deterred from war by their recognition of each other's nuclear weapon capabilities. In South Asia at least, the logic of nuclear deterrence has been closer to the mark than the logic of nonproliferation. The concluding section briefly explores the implications of my research for regional stability and the broader phenomena of nuclear deterrence and nuclear proliferation.
The logic of nonproliferation's main foundation is the belief that new nuclear nations will be especially prone to preemptive nuclear escalation, a condition Thomas C. Schelling dubbed "the reciprocal fear of surprise attack."(12) From this perspective, the low survivability of emerging proliferants' second-strike forces, and their unsophisticated command, control, communications, and intelligence ([C.sup.3]I) capabilities could breed miscalculation of adversary actions or intentions, and lead to unnecessarily hasty decision-making. To reduce the vulnerability of their weapons, new nuclear states might adopt launch-on-warning procedures, which in turn could promote hair-trigger reactions to perceived threats. As Steve Fetter argues, if one proliferant believed war to be inevitable, it might "try to preemptively destroy the other side's vulnerable but valuable weapons of mass destruction. Even if both sides prefer not to preempt, each may fear that the other side will; consequently, both may decide to launch at the first (perhaps false) indication of an attack."(13)
Analysts have typically compared regional security environments unfavorably with the Cold War nuclear arms competition. Brad Roberts writes: "There are many reasons to think that the emergence of stable deterrence in the East-West context had to do with unique cultural and geostrategic circumstances not found in the Middle East, South Asia, or East Asia. . . . the political, technical, and situational factors in these regions differ sharply from the Cold War nuclear framework."(14) Bruce G. Blair argues that the superpowers were themselves more vulnerable to crisis instability than is commonly believed, and that aspiring proliferants' lesser technological capabilities make them even more subject to preemptive war pressures. Blair maintains that because U.S. and Soviet leaders knew that their [C.sup.3]I systems were vulnerable to disruption by even a few incoming warheads, Washington and Moscow delegated alert and launch authority to lower levels in the chain of command and shortened response times to perceived attacks. Sophisticated early warning networks created an extremely time-sensitive interaction between the two [C.sup.3]I systems, which generated intense escalation pressures. Blair writes that "classic preemptive instability resulted, and the command and control systems became prone to overreaction to erroneous indications of enemy attack." He believes that emerging nuclear powers face the "same dilemmas as those faced by their predecessors during the Cold War. A similar evolution of their nuclear postures and an attendant increase in the risk of inadvertence are predictable."(15)
The logic of nuclear deterrence downplays the likelihood of preemptive war between new nuclear states. For Kenneth N. Waltz, preemption is viable "only if the would-be attacker knows that the intended victim's warheads are few in number, knows their exact number and locations, and knows that they will not be moved or fired before they are struck. To know all of these things, and to know that you know them for sure, is exceedingly difficult." Also, because nuclear weapons are easy to hide and move, creating uncertainty for the attacker does not require advanced technology.(16) John J. Weltman generally agrees, but admits that there are "some regions where short distances . . . combine with economic and industrial constraints to suggest that local powers will never be able to achieve levels of survivability relative to one another approaching those achieved by the superpowers." He adds, however, that "high population densities in small areas" also characterize these regions: "Failure to eliminate even a single deliverable weapon would thus be to risk catastrophe; and short distances mean that no great sophistication in means of delivery is required for a successful countervalue response."(17)
Past practice indicates that in the area of crisis stability, the logic of nuclear deterrence is more robust than the logic of nonproliferation. The Gulf War suggests the difficulties any country would need to surmount in launching a successful first strike. Although the January 1991 allied bombing target list included only two Iraqi nuclear installations, UN inspectors after the war discovered over twenty Iraqi nuclear weapon facilities. Moreover, 1000 hours of allied air strikes left much of the Iraqi nuclear infrastructure untouched.(18) The coalition failure to locate and destroy Iraqi SCUD missiles illustrates the virtually unlimited possibilities for deception in concealing nuclear delivery systems.(19) The Cuban missile crisis provides another example. Former White House aide McGeorge Bundy recalls that President Kennedy had decided to proceed with a naval quarantine of Cuba, "subject only to a final discussion of the [alternative] air strike [option] with the responsible air commander." In a review of the "requirements and prospects" of a "massive" air strike against the Soviet missiles in Cuba, the U.S. Air Force "could not promise to destroy more than 90 percent of the Soviet missiles" in a first strike. Nor could the Air Force "guarantee beyond doubt that no Soviet missile would be fired in reply by a local commander." As Bundy recalls: "discussion was quickly over, and the president's decision [for the quarantine] became final."(20) Even more telling, in both of these examples the [potential] attacker deployed the world's most sophisticated weaponry and intelligence capabilities, and enjoyed a decidedly lopsided advantage over the defender. These conditions would not hold in future crises between new nuclear states, making the prospects for first strike success even bleaker.
If history discloses an unblemished record of political leaders resisting the temptation to decapitate their enemies' existing nuclear forces, opacity enhances this extreme caution. After all, opaque nuclear forces are even less attractive targets for first strikes than transparent ones, because they are even more shrouded in ambiguity and secrecy. How many weapons does the opponent have? Are they assembled? If so, where are they located and in what mode? Are they mobile or hidden? Which are real and which are dummies? If the weapons are unassembled, where are the various components stored? Complicating this murky picture are the limited intelligence capabilities that opaque proliferants can bring to bear in their efforts to answer these questions. In sum, all that is necessary to deter the launching of a preemptive strike is "first strike uncertainty," or the planting of a seed of doubt in the minds of the potential attacker's leaders about whether it is possible to destroy all of the victim's nuclear weapons before it can retaliate. The same technological back-wardness that is said to make new nuclear forces vulnerable also implies that they will be nowhere near sophisticated enough to achieve first strike reliability. Although analysts blithely speak of new proliferants getting in the first blow, "so as partly to disarm the opponent and to minimize damage,"(21) even a 99 percent success rate could well be suicidal. Viewed from the attacker's rather than the defender's perspective, regional balances of terror do not appear "delicate"; indeed, consideration of this whole issue is distorted rather than clarified by the inordinate preoccupation with huge margins of weapon survivability and "windows of vulnerability" during the U.S.-Soviet nuclear competition. Although survivability in any regional nuclear balance will be lower than that between the superpowers, it will hardly be negligible.
This reasoning has broader significance: deriving our expectations about the strategic calculus between emerging nuclear nations solely from the U.S.-Soviet experience is not analytically fruitful. Contrary to ingrained Western thinking, sub-superpower nuclear weapon states may actually be less subject to preemptive instability than were the superpowers. During much of the Cold War, Washington and Moscow embraced counterforce targeting strategies, which placed a high premium on detecting and responding hastily to signs of enemy attacks; any hesitation on the part of the defender would have left it vulnerable to an attack that might have degraded its ability to command and control a counterforce response. The threat of preemptive war between the United States and the Soviet Union stemmed from the counterforce doctrines they followed, not from any inherent logic of nuclear strategy.(22) In this respect, the superpower nuclear arms competition was sui generic: without exception, every subsuperpower proliferant has embraced countervalue, not counterforce, nuclear doctrines.(23)
Countervalue doctrines are theoretically more crisis-stable, because they have only one essential requirement: that a small number of nuclear warheads can be delivered against the enemy even after his attempted first strike. Establishing adversary first-strike uncertainty is relatively easy. Prompt responses are unnecessary if one is not targeting the opponent's own nuclear weapons. With a counter-city strategy, the defender can wait to see if an attack actually materializes before responding. As a former Indian Chief of Army Staff (COAS) notes, a "very highly sophisticated, highly responsive" [C.sup.3]I system that "functions in real time" is not necessary for an effective countervalue strategy.(24) U.S. nuclear planners' preoccupation with robust [C.sup.3]I was related to expanding their range of "flexible responses," a menu of options in which the second-generation proliferants have shown little interest. The assumption that regional nuclear arms competitions will simply mimic the dynamics of the superpower arms race is not useful. It obscures the fact that there can be more than one road to nuclear deterrence, and the possibility that non-U.S. defense planners may have understood and accepted the fundamental principles of the nuclear revolution better than their U.S. counterparts, who struggled mightily to overcome these principles through "rational" planning.
Opaque proliferants pursue nuclear capabilities at least in part for the deterrent effect they will have on adversaries. This raises an interesting puzzle: how can new nuclear powers deter aggression when they deny deploying nuclear weapons? The logical nature of nuclear deterrence under opacity is "existential deterrence," a concept invented by McGeorge Bundy. Bundy argued during the Cold War that any nuclear conflict between the superpowers would be fraught with "terrible and unavoidable uncertainties" which have "great meaning for the theory of deterrence":
They create what I will call existential deterrence. My aim in using this fancy adjective is to distinguish this kind of deterrence from the kind that is based on strategic theories or declaratory policies or even international commitments. As long as we assume that each side has very large numbers of thermonuclear weapons which could be used against the opponent, even after the strongest possible pre-emptive attack, existential deterrence is strong. It rests on uncertainty about what could happen, not what has been asserted.(25)
According to Bundy, existential deterrence was "strong in every major crisis between the superpowers since `massive retaliation' became possible for both of them in the 1950s. As everyone closely involved recalls, such deterrence was particularly powerful during the Cuban missile crisis."(26) President Kennedy's Secretary of Defense, Robert McNamara, concurs with this assessment. He relates that during the October 1962 crisis, Washington was deterred from "even considering a nuclear attack by the knowledge that, although such a strike would destroy the Soviet Union, tens of their weapons would survive to be launched against the United States. These would kill millions of Americans. No responsible political leader would expose his nation to such a catastrophe."(27)
Under conditions of opacity, the role of existential deterrence is even more pronounced. Since each side in an opaque nuclear arms competition has only limited information about the other side's nuclear forces, any deterrence derived from nuclear capabilities will logically be existential. In other words, even more than in the U.S.-Soviet nuclear balance, mutual deterrent calculations rest not on relative capabilities and strategic doctrines, but on the shared realization that each side is nuclear-capable, and thus that any outbreak of conflict might lead to nuclear war. Of course, there are stark differences between the huge U.S.-Soviet thermonuclear standoff and the comparatively tiny atomic bomb balance between India and Pakistan. To the extent that Washington and Moscow had no doubt about each other's assured second-strike capabilities, these assets were "visible" and helped make mutual deterrence credible. The important question for opaque proliferants is: can deterrence be achieved with less potent nuclear capabilities? Following the earlier assessment of preemptive logic, the answer is yes. The sine qua non of nuclear deterrence is the survivability, not size, of one's nuclear forces. As the superpower arms race spiraled into ever-larger forces, U.S. and Soviet strategists came to equate survivability in part with redundancy. Logically, however, large forces are only required to ensure survivability against preemption by other large forces. The absolute size of one's nuclear forces is epiphenomenal; the real key to maintaining first-strike uncertainty is a very rough equivalence in forces, which ensures that one side cannot simply overwhelm the other in a massive attack.
Another question arises in this context: is existential deterrence limited to thermonuclear weapons? Is the qualitative difference between atomic and hydrogen weapons so great that the effects of existential deterrence are negated in atomic arms competitions? In the context of the U.S.-Soviet balance, there is reason to believe that, had the two sides stopped short of developing hydrogen bombs, the effects of existential deterrence might have been limited. The United States is, and the Soviet Union was, a huge land mass with a widely dispersed population. It is conceivable that an exchange of atomic weapons, if terminated quickly, might not have devastated the two countries. This would imply perhaps some increased willingness to use nuclear weapons during a crisis. For Robert Jervis, Bernard Brodie's 1946 argument that "military victory in a nuclear war was not possible" was "certainly incorrect" when he made it. "It took the hydrogen bomb to bring about the world Brodie had foreseen."(28)
This distinction between fission and fusion weapons is less meaningful on the Indian subcontinent, which has one of the world's highest population densities. Even small fission weapons would kill millions of Indians and Pakistanis. In the event of a nuclear exchange, no South Asian leader can have any illusion that the use of atomic weapons would result in an "acceptable" number of deaths. Moreover, each country's industrial and administrative assets are concentrated in a few cities, implying that several atomic bombs could severely disrupt each society. Other factors, too, suggest that atomic existential deterrence should exercise a powerful impact on nuclear decision-making. Heavily congested areas of India and Pakistan are in close proximity, raising the possibility that radioactive fallout could blow across national boundaries to poison the aggressor's own population. Also, given these countries' limited emergency management capabilities and health care infrastructures, recovery from a nuclear exchange would be extremely problematic.(29) In sum, the distinction between fission and fusion weapons may have been significant in the U.S.-Soviet case, but it is less so in South Asia.(30)
Existential nuclear deterrence is a structural condition affecting the behavior of all nuclear weapon states, but its power is in turn enhanced by the behavior of those states. Deterrence of any kind depends on the adversary's perception of one's capabilities and one's resolve to use them.(31) This raises an important question: can opaque proliferants deter aggression without overt demonstrations of their nuclear prowess and direct nuclear threats against adversaries? The answer is yes: like all nuclear weapon states, opaque proliferants signal resolve to one another through a process of strategic bargaining, which runs along a communication spectrum from formal negotiations to the transmission of intentions via deeds rather than words. Signaling that falls toward the latter end of this spectrum is called tacit bargaining, "in which communication is incomplete or impossible."(32) As Neil Joeck observes, opaque proliferants are forced into this mode of discourse by the imperatives of secrecy.(33) Formal negotiations require, as a basis for discussion, exchanges of detailed information that opaque proliferants are loath to provide. Under opacity, the signaling of resolve takes a variety of tacit forms: intentions can be transmitted indirectly by, for example, passing them through intermediaries. Or they can be signaled through state-controlled or state-influenced media organs. Schelling terms this "passive deterrence," which is achieved by "just letting it be known, perhaps through an innocent leak of information, that a government . . . simply had nuclear weapons, letting every potential addressee of this `deterrent threat' reach his own conclusions about what kind of misbehavior, if any, might provoke nuclear activity."(34) There is not an absence of communication between opaque proliferants, but a different type of communication: less formal, less direct, but mutually intelligible nonetheless. Over the years, this discourse establishes certain deterrent understandings that may not be as clear as those between the transparent nuclear powers, but which are compelling all the same. As these understandings develop, it becomes exceedingly unlikely that decision-makers in opaque nuclear states will fail to understand the possibilities that confront them.
While deterrent intentions can be signalled by tacit bargaining, capabilities must also be credibly demonstrated. How are the opaque proliferants' nuclear capabilities made "visible"? Paradoxically, the nonproliferation community has inadvertently "solved" this dilemma. One means by which governments and private "watchdog" organizations try to pressure recalcitrant proliferants into nuclear chastity is by publicizing their nuclear transgressions. For example, senior U.S. policymakers have repeatedly stated over the past few years that India and Pakistan can build and deliver nuclear weapons quickly in the event of a crisis.(35) These pronouncements, intended to pressure opaque proliferants into reversing their course, instead stamp their nuclear programs with a seal of credibility that they would otherwise lack.
This section has established a theoretical basis for believing that nuclear deterrence can work in regions like South Asia, albeit in a much different way than traditionally conceptualized. My argument rests on two main pillars: that the chances of preemptive escalation between new nuclear powers are much lower than most analysts believe, especially given the profound "first-strike uncertainty" imposed by opacity; and that existential nuclear deterrence is made credible by the twin processes of tacit bargaining between opaque proliferants and inadvertent "transparency-building" by the nonproliferation community.
The case study that follows is framed by the debate between the logic of nuclear deterrence and the logic of nonproliferation. A stiffer challenge for nuclear deterrence could scarcely be devised: in 1990, two opaque proliferants made preparations for war over a territory that had already caused two of the three wars between them. The weaker of these states, Pakistan, had grounds to fear a conventional invasion that might have "liberated" a large chunk of its territory, for which the 1971 Bangladesh war provided a vivid precedent. The stakes could not have been any higher for Pakistan in 1990. Furthermore, both India and Pakistan were technologically inferior to other nuclear weapon states, which the logic of nonproliferation suggests should lead to pressures for the preemptive use of nuclear weapons. Despite these conditions, war did not break out in 1990. I examine why below.
The late 1980s marked the dawn of a new era in South Asia's international relations. Global, regional, and domestic developments sharply intensified the India-Pakistan rivalry. These three political streams reached their confluence in the spring of 1990, with the eruption of a major crisis over the disputed territory of Kashmir. Analysts disagree about the nature of the 1990 crisis, in particular whether or not New Delhi and Islamabad were on the brink of war and what, if any, nuclear dimension shaped their behavior. It will suffice at the outset to state that there was a real possibility of war between India and Pakistan in 1990, and that no decision-maker in Islamabad or New Delhi could rule out the slimmer, but not negligible, chance that such a war might involve the use of nuclear weapons. A brief discussion of the context shaping South Asian politics in 1990 sets the stage for an analysis of the crisis.
At the global level, the end of the Cold War had serious repercussions for India and Pakistan. A fundamental characteristic of South Asian international relations since the 1950s had been various regional manifestations of the global U.S.-Soviet competition. Beginning in the mid-1950s, Pakistan was linked with the United States in two regional containment alliances and a third, bilateral pact that pledged the United States to help Pakistan resist communist aggression. The 1979 Soviet invasion of Afghanistan ushered in nearly a decade of unprecedented bilateral cooperation, which netted Pakistan billions of U.S. dollars in military and economic assistance. As for India, its special relationship with the Soviet Union began with Stalin's death in 1953, and culminated in the 1971 Indo-Soviet Treaty of Peace, Friendship, and Cooperation. Indian leaders were privately disturbed by the Soviet war in Afghanistan, but New Delhi continued to buy vast quantities of Soviet weaponry during the 1980s. Super-power involvement in South Asia changed with the U.S.-Soviet thaw in the mid to late 1980s. One of Mikhail Gorbachev's main international objectives was to stanch the "bleeding wound" of Afghanistan; Moscow finally withdrew its forces in February 1989. The Soviet defeat cut the legs out from under the U.S.-Pakistani strategic partnership, allowing submerged policy differences to resurface. The most important of these disagreements concerned Pakistan's continued development of nuclear weapons, which replaced the Afghan war effort as the foremost policy issue for Washington. India's relationship with the Soviets also suffered. As Moscow's rivalry with Washington evolved into a tentative partnership, its attention turned inward, and India lost its value as a bulwark of pro-Moscow sentiment in the Third World. As the new decade began, both New Delhi and Islamabad were losing their international moorings.
Regionally, the new decade brought the re-emergence of the Kashmir dispute, which quickly plunged Indo-Pakistani relations into a tailspin. Both India and Pakistan have claimed sovereignty over the former Indian princely state of Jammu and Kashmir since the hasty British withdrawal from the subcontinent in 1947.(36) Their dispute over the Muslim-majority territory was the root cause of two Indo-Pakistani wars, in 1947-48 and 1965; Kashmir has been divided between India and Pakistan since the first war. Neither country has been willing to compromise over Kashmir, partly for strategic reasons, but mainly because this would threaten the legitimating ideology on which each modern state was founded. Pakistan's "two-nation theory" held that the subcontinent's Muslims could safeguard their legitimate political rights only through the formation of a separate nation-state. For Pakistan, the idea of a Muslim-majority state falling within Indian borders was anathema, as it directly repudiated the two-nation theory and thus the entire basis for the creation of Pakistan. India's notion of a secular nation-state rested on the successful incorporation of all minorities, including Muslims, into the Indian political order. A Pakistani Kashmir would be an insult to Indian secularism.
In the 1980s, political agitation grew among the Muslims of Indian-held Kashmir against the central government in New Delhi. By 1989, Islamic militants in the Vale of Kashmir were in open rebellion and, in the years since, a full-blown secessionist insurgency has erupted against the Indian state. A1though some of the insurgents enjoy material support from Pakistan, most analysts, Indian and otherwise, agree that this latest phase of the Kashmir dispute is rooted in the policies of governments in New Delhi and Srinagar, particularly under Prime Minister Indira Gandhi and her son and successor, Rajiv Gandhi.(37) Particularly destabilizing was the Congress (I) party's incessant meddling in Kashmiri politics during the 1980s. By ousting a popularly elected state government in 1984 and rigging state elections in 1987, the Congress Party helped to create a critical mass of Kashmiri Muslims who were politically estranged from Indian democracy. In 1989, as the Vale of Kashmir descended into anarchy, New Delhi accused Islamabad of waging an unconventional war with India by arming and training Kashmiri Muslim "terrorists." Islamabad responded that it provided only moral support to the Kashmiri "freedom fighters." In 1990, the escalating war between Indian security forces and Kashmiri militants radically transformed India-Pakistan relations by bringing the two governments close to war for the first time since the Bangladesh war of 1971.
Domestically, the elections of prime ministers Benazir Bhutto in 1988 and V.P. Singh in 1989 were positive strides for democracy on the subcontinent. Bhutto's ascent to power was a big step in the restoration of democracy in Pakistan, begun cautiously in 1985 and continuing, unsteadily, to the present. Singh's triumph over Rajiv Gandhi marked only the second time in history that India's dominant Congress had been removed from office, and the uneventful transfer of power in New Delhi was widely interpreted as signaling the vitality of Indian democracy. Below the surface, however, the domestic political picture in South Asia was less rosy. After 1988, Pakistan was effectively ruled by a troika of leaders of whom the inexperienced Bhutto was the weakest. The other two power bases revolved around the president, Ghulam Ishaq Khan, and the COAS, General Mirza Aslam Beg. While they were content to let the charismatic Bhutto represent Pakistan for public diplomacy purposes, she chafed under their dominance of national security issues. In India, V.P. Singh's position was also tenuous. His government was a fragile coalition, forced to rely on the rightist Bharatiya Janata Party (BJP) for support in parliament. Once in power, it was immediately buffeted by a welter of contentious issues, including Kashmir, that severely weakened its authority.
Hovering over this tense situation was the prospect of nuclear war. By 1990, the Indo-Pakistani nuclear arms competition was two decades old. India exploded a nuclear device in 1974; Pakistan had resolved after its 1971 defeat in the Bangladesh war to pursue nuclear weapons, and the Indian test increased its determination. In 1986, a U.S. Special National Intelligence Estimate concluded that Pakistan could produce weapons-grade fissile material, making it a de facto nuclear power.(38) New Delhi recognized Pakistan's nuclear progress. As the Indian army chief, General K. Sundarji, said in 1986: "There are enough indicators to suggest that Pakistan has achieved or is close to achieving a nuclear weapons capability." The Indian military, he said, was gearing its "organisation, training, and equipment in such a manner that is not only effective in conventional use but in the unlikely event of [use of] nuclear weapons by an adversary in the combat zone, we will limit damage both psychological and physical."(39)
In January 1987, India and Pakistan nearly went to war during a major crisis precipitated by India's "Brasstacks" exercises, the largest military maneuvers in South Asian history.(40) Nuclear weapons seem not to have been a major factor in the Brasstacks crisis.(41) According to former Foreign Secretary Abdul Sattar, Pakistan's nuclear capabilities had not yet "flowered" by the time of Brasstacks. They were, he said, "nascent," but "not yet actual."(42) The Indian perception was also that Pakistan "had not weaponized."(43) In the wake of Brasstacks, Pakistan raised its nuclear profile. In March 1987, President Zia ul-Haq admitted that "Pakistan can build a [nuclear] bomb whenever it wishes. Once you have acquired the technology, which Pakistan has, you can do whatever you like."(44) In response, Indian Prime Minister Rajiv Gandhi said, "We intend meeting President Zia's threat. We will give an adequate response."(45) Soon thereafter, senior Pakistani leaders began openly discussing the role of nuclear deterrence in preventing war on the subcontinent. In 1988, Zia said, "The present programs of India and Pakistan have a lot of ambiguities, and therefore in the eyes of each other, they have reached a particular level, and that level is good enough to create an impression of deterrence."(46) A U.S. expert summed up the consensus view of South Asia's nuclear balance in 1990: "India and Pakistan are currently capable of deploying small nuclear forces comprised of atomic bombs that could be delivered by advanced fighter-bombers, with India's capabilities being considerably greater than Pakistan's. Neither country is believed to have integrated nuclear weapons into its military forces, however, and it is possible that neither has manufactured complete nuclear devices."(47)
By January 1990, the separatist movement had begun to affect every aspect of life in Kashmir. Insurgent-imposed shutdowns and government curfews had brought economic activity to a virtual standstill. On January 20, tension between the militants and security forces exploded into what would be the first of many spasms of violence, with police firing on a crowd of demonstrators who had defied a government curfew. The fighting transformed the Kashmiri insurgency from a mainly Indian affair into renewed Indo-Pakistani conflict. On the day of the massacre, Pakistan's ruling troika met to discuss the situation, and Bhutto loudly proclaimed the Kashmiris' right to self-determination.(48) In early February, opposition politicians called on the Pakistani government to pursue a jihad (holy war) in Kashmir, and one urged it to build nuclear weapons in order to meet the Indian threat.(49) Robert Oakley, the U.S. ambassador to Islamabad at the time, remembers that although the initial popular uprising in Kashmir was "primarily spontaneous," the Pakistani government "began to take a more active role in support of the Kashmiri protests. Training camps of various kinds multiplied. . . . There were more people and more material going across the border from Pakistan into Kashmir."(50)
Several Indian moves in early 1990 led observers to speculate that New Delhi was raising its nuclear profile, perhaps to send a deterrent message to Pakistan. Two V.P. Singh appointments--Raja Ramanna as minister of state for defense, and P.K. Iyengar as chairman of the Indian Atomic Energy Commission--put prominent nuclear scientists into senior government posts. To one news weekly, the Ramanna appointment "can only mean that India has decided to give higher priority to its nuclear weapons and missile-development programmes." In February, Singh "said India would have to review its peaceful nuclear policy if Pakistan employed its nuclear power for military purposes." He also "told newsmen that Pakistan's going nuclear would bring about a radical change in the security environment in the region. If this were to happen, 'we will have to take stock of the situation and act accordingly'." An influential nuclear trade publication suggested that Singh's appointments and public utterances meant that India was increasing its nuclear preparedness.(51)
The Indian Army's chief concern, according to then-COAS V.N. Sharma, was to stem the infiltration of Pakistan-backed Sikh and Kashmiri "terrorists," who threatened to overwhelm local Indian police forces. Sharma told an interviewer in 1993: "Terrorist groups backed by agencies in Pakistan were able to attack railway stations and vital installations which could affect any military movement on our side. . . . Therefore, there was need for the Indian army to go in there to take care of the communication lines and other bottlenecks so that if there was a military flare-up., we could conveniently move our fighting forces from locations deep in the country to the border areas."(52) Indian military planners were also concerned about residual deployments of Pakistan Army forces after a late 1989 military exercise called Zarb-i-Momin. According to Beg, these maneuvers tested a new Pakistani strategy: "In the past we were pursuing a defensive policy; now there is a big change since we are shifting to a policy of offensive defence. Should there be a war, the Pakistan Army plans to take the war into India, launching a sizeable offensive on Indian territory."(53) After the exercise was over, says Sharma, "we found that these troops were not going back to their peace stations, but they were staying on in the exercise area which is quite close to the international border and the cease-fire line in Jammu and Kashmir." New Delhi believed "Pakistan was keeping troops ready as a back up support to the increased terrorist activities, in Indian territory, across the border and could take full advantage of terrorist successes to support military intervention."(54)
Further south, according to Sharma, the Indian army in February sent two new tank units for training at its field firing range at Mahajan, in Rajasthan. With Brasstacks fresh in their minds, Pakistani planners grew alarmed that the Indian armored units at Mahajan were "ginning up another large exercise of that nature, or, indeed, preparing to launch an attack from the training range."(55) Sharma told U.S. Ambassador William Clark that the Indian Army could not launch an effective offensive against Pakistan from Mahajan, and the U.S. embassy staff concurred.(56) Clark's air attache, Colonel John Sandrock, remembers that "what was unusual from our perspective was the deployment of additional troops in Kashmir as a result of the reported cross-border infiltration from Pakistan into Kashmir and then along the border, south through the rest of Jammu and Kashmir and into the [Indian state of] Punjab." According to Sandrock, there was no evidence that this deployment included tanks and artillery, which appeared to corroborate Indian claims that the "buildup of forces on the border was to prevent cross-border infiltration and did not constitute a buildup of forces preparing for any hostile action against Pakistan." U.S. military attaches in New Delhi took the first of several reconnaissance trips in February, confirming their impression that Indian forces were not preparing for an offensive military thrust. U.S. attaches in Islamabad undertook a similar series of fact-finding missions on their side of the border in February, also finding little unusual military activity. Of special importance, one of the attaches noted, was that the two Pakistani strike corps were not on the move, and that the Pakistan Air Force's forward operating bases were not opened.(57)
Meanwhile, the war of words between India and Pakistan escalated. On March 13, Bhutto traveled to the Pakistan-held area of Kashmir, where she promised a "thousand-year war" in support of the Kashmiri militants. V.P Singh quickly responded that India would react decisively against Pakistani intervention in Kashmir. "There should be no confusion," he told the Indian parliament. "Such a misadventure would not be without cost."(58) In early April, the BJP's national executive committee passed a resolution urging the Indian government to "knock out the training camps and transit routes of the terrorists."(59) Congress leader and former Prime Minister Rajiv Gandhi added to the clamor by urging the government to take "some very strong steps on Kashmir." He added: "I know what steps are possible. I also know what is in the pipeline and what the capabilities are. The question is, does the government have the guts to take strong steps?"(60) On April 10, Singh made an extraordinary speech, warning Indians to be "psychologically prepared" for war. Addressing Islamabad, he said: "Our message to Pakistan is that 'you cannot get away with taking Kashmir without a war'. They will have to pay a very heavy price and we have the capability to inflict heavy losses." Singh responded to Bhutto's threat by saying that "those who talk about 1000 years of war should examine whether they will last 1000 hours of war." As if to dispel any notion that Pakistan's nuclear weapon capabilities would give Islamabad a deterrent umbrella under which to carry out offensive operations against India, Singh said that if Pakistan deployed nuclear weapons, "India will have to take a second look at our policy. I think,we will have no option but to match. Our scientists have the capability to match it."(61)
On April 11, Pakistan's army chief convened a meeting of his corps commanders to carry out a "detailed threat assessment." Beg told his subordinates that India had deployed a strike force of up to 100,000 men within fifty miles of the border in Rajasthan. He was referring to the Indian army units that were on winter exercises in the Mahajan area, which Pakistani officials now declared extended. They estimated that the Indian units were deployed in such a way as to "halve India's normal mobilisation time to one week." In addition, Islamabad noted that New Delhi continued to move military forces into Kashmir. One reporter wrote: "The concern in Islamabad is that India might be preparing an attack on Pakistani Kashmir on the pretext of destroying Kashmiri 'freedom fighter' training camps. There is also concern that a simultaneous attack might be launched into Sindh province, where the only road and rail link between north and south Pakistan is located about 40 km from the Indian border."(62)
Indian officials confirmed that they were putting more men and arms into Kashmir. Diplomats in New Delhi "said forces on both sides of the border were on a higher than normal state of alert, but several levels lower than would indicate imminent hostilities."(63) Western military analysts reported no major troop mobilization near the international frontier, but speculated that by extending their exercises, Indian military planners may have positioned tanks and heavy artillery near the border. In the words of one analyst, "everything the Indians have been doing fits under the category of defensive preparedness, but some of it is ambiguous."(64) In the meantime, sentiment was growing among influential Indians for offensive strikes against Pakistan. The Indian home minister argued that war "would be fully justified if the objective of freeing Kashmir from the stranglehold of the secessionists was achieved." BJP leader L.K. Advani warned that Pakistan would "cease to exist" if it attacked India.(65) Despite severe domestic political pressures, however, "the Indian military leadership deliberately refrained from moving armor associated with its strike forces out of peacetime cantonments," and Pakistan "deliberately refrained from moving its two strike corps to the front."(66)
Throughout the spring, Washington had grown increasingly alarmed over the escalating Kashmir crisis. U.S. concern crystalized in the middle of May with the visit to South Asia by Deputy National Security Adviser Robert Gates. In his May 20 meeting with Ghulam Ishaq and Beg, Gates made the following points: Washington had thoroughly war-gamed the Indo-Pakistani confrontation, and Pakistan was the loser in every scenario; in the event of a war, Islamabad could expect no help from Washington; and Pakistan must refrain from supporting terrorism in Indian-occupied areas of Kashmir, avoid threatening military deployments, and tone down its war rhetoric.(67) In New Delhi, Gates met with Singh, Sharma, and other Indian leaders. In these meetings, his message was similar: India must avoid provocation that could spiral out of control. In sum, according to one official, the gist of the message to both sides was that war would be to neither side's advantage. India would win, but even if it did, the long-term costs would exceed any short-term benefits.(68) In early June, India announced that the armor it had sent to the Mahajan range in February would be returned to its normal station. Within two weeks of the Gates mission to South Asia, the crisis passed.
Hersh reports that as the Kashmir crisis intensified, Pakistan responded to India's movement of military forces to Kashmir and Rajasthan by preparing to attack India with nuclear weapons. Specifically, he contends that:
Sometime in the early spring of 1990, intelligence that was described as a hundred per cent reliable--perhaps an NSA intercept--reached Washington with the ominous news that General Beg had authorized the technicians at Kahuta to put together nuclear weapons. Such intelligence, of "smoking gun" significance, was too precise to be ignored or shunted aside. The new intelligence also indicated that General Beg was prepared to use the bomb against India if necessary. Precisely what was obtained could not be learned, but one American summarized the information as being, in essence, a warning to India that if "you move up here"--that is, begin a ground invasion into Pakistan--"we're going to take out Delhi."
Washington subsequently increased its satellite coverage of South Asia, which "sometime in May" yielded "photographs of what some officials believed was the evacuation of thousands of workers" from Pakistan's uranium enrichment installation at Kahuta. One analyst told Hersh that he thought that in the event of an Indian ground thrust into Sindh, Pakistan would "cut it off with a nuke. We thought they'd go for Delhi." This analyst continued: "We thought the reason for the evacuation of Kahuta was that they expected a retaliatory attack by India." Soon thereafter, U.S. intelligence produced signs of a truck convoy moving from a suspected nuclear storage site to an air force base in the western Pakistani province of Baluchistan. Eventually, the same analyst told Hersh, U.S. intelligence indicated that Pakistan "had F-16s pre-positioned and armed for delivery--on full alert, with pilots in the aircraft. I believed that they were ready to launch on command and that the message had been clearly conveyed to the Indians. . . . These guys have done everything that will lead you to believe that they are locked and loaded." CIA official Kerr told Hersh: "There's no question in my mind that we were right on the edge. This period was very tense. The intelligence community believed that without some intervention the two parties could miscalculate--and miscalculation could lead to a nuclear exchange." Indeed, according to Hersh, the Gates mission "defused what looked to be inevitable warfare."(69)
Since the publication of Hersh's article, his account of the crisis has been roundly criticized. Most damaging to his thesis have been the categorical denials of many of the key elements of his story by U.S. diplomats and military attaches posted in Islamabad and New Delhi during the spring of 1990. In particular, the U.S. ambassadors involved in the crisis decision-making, Oakley in Islamabad and Clark in New Delhi, directly contradict Hersh's central claims. Moreover, they strongly imply that Hersh manipulated the information he collected to highlight evidence that supported his thesis, and discount that which contradicted it. Other senior officials involved in the 1990 decisionmaking--U.S., Indian, and Pakistani--agree that, at a minimum, Hersh exaggerated the nuclear dimension of the crisis. Hersh also has his defenders, though, who contend that the dangers he describes were real, and that those who disagree with his assessment are themselves guilty of what one U.S. intelligence analyst calls "historical revisionism."(70) Several questions arise.
DID THE PAKISTANI NUCLEAR PROGRAM "CROSS THE LINE IN 1990, AND DID PAKISTANI LEADERS MAKE PREPARATIONS TO USE NUCLEAR WEAPONS AGAINST INDIA? There is little doubt that the Pakistani nuclear weapon program crossed an important threshold sometime during spring 1990, although exactly when is uncertain. In 1989, the new Bhutto government secretly made two promises to Washington: first, that it would stop enriching uranium to weapons-grade; and second, that it would not convert its existing stock of weapons-grade uranium from gas to metal, which could then be machined into nuclear bomb cores.(71) Oakley confirms that in early 1990, "the freeze on the Pakistan nuclear program was . . . removed. And the program began to move forward again. This is what led eventually to the application of the Pressler Amendment," cutting off U.S. aid to Pakistan.(72) By the time the Gates mission reached South Asia, Oakley remembers, "we had ascertained beyond a shadow of a doubt that the promises that Mrs. Bhutto had made and kept during 1989 . . . had been broken and the nuclear program had been reactivated."(73)
However, Hersh's allegations about Pakistani preparations to deliver nuclear weapons against India are directly refuted by the ranking U.S. diplomats in Islamabad and New Delhi in 1990. Clark says, "From my side of it, I'm not aware of any of . . . them." Oakley and Clark deny knowing of any NSA intercept that Hersh claims may have indicated that the Pakistani leadership had authorized the assembly of nuclear weapons. Oakley says that invoking the Pressler Amendment "had nothing to do . . . at least so far as we knew, with the preparation or deployment of nuclear weapons. It had to do with other factors, which were required for certification." On the supposed evacuation of the Pakistani nuclear facility, Oakley says: "we know nothing about any evacuation of Kahuta." He further maintains: "So far as I can recall, we never had any credible evidence that the F-16s were fitted out to deliver a nuclear device; that Pakistan had a nuclear device that could be delivered by an F-16.... Nor did we know anything about any nuclear devices being moved from point 'x' to point 'y,' if there were any." Colonel Don Jones, the U.S. air force attache in Islamabad at the time, is even more vehement in his denial of Hersh's claims, particularly those relating to F-16s on strip alert, which he calls "the silliest allegation I read in the article, and there were a lot of silly things in the article.... Some of the things he said were just on the face of it, ridiculous. It's not true." Jones adds: "The Pakistan Air Force F-16s were sheltered. They were all in sheltered revetments. You'd have to have a camera capable of seeing through three feet of concrete to know what was underneath. They did not leave their alert birds out in the open. They were all sheltered. So from any aspect that you care to look at it, that particular statement simply was not--does not hold water. It's so silly that it--that it found its way into print astounds me, really."(74)
Senior Indian and Pakistani decision-makers have also refuted the Hersh thesis. General Beg, who Hersh claims authorized the assembly of nuclear weapons, "has staunchly denied" Hersh's version of events. According to a Pakistani physicist, "He claimed that Pakistan did not possess a usable nuclear device at that time. Therefore his country could not have been poised to use such a weapon against India. Moreover, in his opinion, such readiness was unnecessary because Pakistan had not faced a critical or desperate situation. Furthermore, 'There was a solid fear of massive retaliation from India,' he recalled, 'as they [the Indians] have a stockpile of more than a dozen warheads'."(75) Beg's Indian counterpart, General Sharma, also scoffs at the notion of the Pakistanis preparing for nuclear war: "There is a lot of bluff and bluster from Pakistan. It is different to talk about something and totally different to do something. In return it is bluff and bluster from India that we would do this and that. In hard military terms your capability is not judged by the bluff and bluster, but by what you have in your pocket and what you can do with it."(76)
The evidence that has come to light so far does not corroborate Hersh's thesis that Pakistan was preparing in spring 1990 to launch nuclear weapons against India. Much of Hersh's analysis of the crisis, including his allegations about the evacuation of Kahuta, the movement of the truck convoy, and the F-16s on strip alert, come from a single, anonymous U.S. analyst whose interpretation of the intelligence data seems not to have been corroborated by anyone else. The evidence against Hersh's characterization of events comes from on-the-record discussions with multiple senior policymakers--U.S., Indian, and Pakistani--who were privy to the most highly classified intelligence information that their governments could generate.(77)
EXACTLY WHAT PRECIPITATED THE GATES MISSION? Based on the available evidence, the following explanation seems plausible: President Bush and his senior advisers were presented with information on the South Asian crisis that ranged from mildly alarming to extremely disturbing. At a minimum, Kashmir was in flames, the Indian and Pakistani armies were preparing for conflict, a shrill war of words had erupted between the two capitals, public passions were aroused, and, hovering over it all, the two sides had the capability to inflict enormous damage against each other. Clark remembers that the Pakistanis made "slightly veiled threats" to the effect that "'we have something that will make you very sorry'." The Indians replied that "'if something happens, we will respond in the appropriate manner'." Clark says, "I know how to read that: 'we've got one too'."(78) Beyond this, however, there was a distinct divergence of opinion within the U.S. government. On the more relaxed side of the ledger, the U.S. ambassadors in Islamabad and New Delhi saw little to indicate that war--either nuclear or conventional--was imminent. On the more alarmed side, Bush and his aides were hearing from the intelligence community that Pakistan had definitely resumed enriching uranium and that certain signs indicated possible Pakistani preparations for the delivery of nuclear weapons. Given these conflicting estimates, and the small but real possibility of a nuclear catastrophe in South Asia, Bush opted to err on the side of caution by sending a high-level U.S. delegation to help ease tensions in the region.
Hersh claims the Gates mission was a direct outcome of signs picked up by U.S. intelligence that Pakistan was making nuclear delivery preparations. In fact, the embassies' concern was different: they believed that without timely intervention, Indo-Pakistani tension would develop an inexorable momentum toward war. Oakley says: "Let me make this very clear. I tried to make it clear to Mr. Hersh and he diddled with it. But at least from Islamabad, we never believed, in part because of what we did see and in part because of the very good information which Bill [Clark] was getting from the Indian Army, that there was going to be an explosion in the spring of 1990." Rather, Oakley recounts, "we feared that if the momentum of this ratcheting up were not stopped by the fall, the prime fighting season, the two armies might be face-to-face again, as they had been at the time of Brasstacks, and the momentum would be so strong that it couldn't be stopped. So we wanted intervention in the spring in order to preempt something we feared might happen in the fall." On the nuclear dimension, Oakley says, "despite what Hersh says--at least in Islamabad, we were not worried about a conflict becoming nuclear. There's always that potential, but there was nothing at that time to indicate that this was the case." Clark agrees, adding that "my views were not as apocalyptic" as Hersh's. "My comments really didn't fit his thesis, and so you will not find me in the article anywhere.... [He] chose not to use what I said."(79)
U.S. intelligence estimated that there was a 50-50 chance of war.(80) As one former senior Bush administration official says, the U.S. view was that "there was a considerable chance of war. That said, no one could map out exactly what it meant. How serious it would go, how it might escalate, whether it would become a major conventional war, or something else; nobody knew exactly whether it would take place, much less how it might evolve." This official also recalls that Indian and Pakistani leaders "were not acting with sufficient sobriety. There was a little bit of recklessness in the air." Furthermore, he says.
I think for most of us who were involved, nuclear weapons formed the backdrop for the crisis.... the concern was not that a nuclear exchange was imminent; the concern was that this thing was beginning to spin out of control and that that would lead to clashes, potentially conventional warfare. Most of our analysis suggested that India would fare better than Pakistan, and that very early on, as a result, Pakistan might want to consider threatening . . . a nuclear action. Or, that India, thinking about that, would escalate conventionally very early on, to eradicate it.
According to this official, the Hersh article "exaggerated considerably the sense that there was kind of a situation where the nuclear trigger was cocked or something."(81) Summing up the nuclear dimension of the crisis, Clark remembers: "There was a little bit of nuclear tension: 'don't threaten me with yours because I've got mine.' I don't think it went beyond that. Nobody was loading weapons."(82)
WHAT WAS THE IMPACT OF THE GATES MISSION? The prevailing view among U.S. officials is that Islamabad and New Delhi quietly used the Gates intervention as an excuse to de-escalate the crisis. Clark says, "At the end of the day, I think you could say that both Delhi and Islamabad used Bob Gates and his mission as an excuse, if you will, to back off of positions they had been taking." This perspective is not limited to U.S. decision-makers, who recommended the Gates mission and thus had a stake in its success: South Asian officials also view the U.S. intervention in a positive light. Sattar says: "I think that what is important is not what was happening in the months of January and February, but the projection of what might happen if the trends in motion were not arrested. And I think it is here that the American diplomacy deserves credit.... What happened in the spring of 1990 is an illustration of good, useful preventive diplomacy." Clark reports that Indian officials, too, appreciated the chance to ease the tension: "I did have several senior people, including the Prime Minister, tell me afterwards that it had been a useful visit, it had allowed a way to back off for both sides, without one having to back down to the other."(83)
The firmest conclusion that can be drawn about the Gates intervention is that it certainly could not have hurt, and might indeed have helped, the prospects for peace in South Asia. It is probably not a coincidence that India offered to withdraw its forces from Mahajan in the weeks immediately following the Gates visit. In all likelihood, the reason that Indian policymakers delayed these decisions for a week or two is that, for domestic political reasons, they did not wish to appear overly influenced by the United States. In addition, the fact that the tension abated so quickly in early June would seem to indicate that both sides were anxious to back away from the brink of war, and that Gates provided them with a mechanism for doing so without appearing weak. Few knowledgeable people would likely quibble with the views of a former senior Bush administration official:
At worst, you could say what we did was unnecessary.... At the risk of sounding self-serving, it was a success.... My instincts are we slowed it down, we forced people to face up to the consequences.... We may have ... affected the internal debates. What matters is sometimes that when you leave town, the internal debates that took place on either side were affected by what it was we said. We knew we'd given arguments to certain people. And my hunch is again we may have stabilized by simply what we said . . . we certainly didn't make the situation worse, and my guess is we made it better. The facts speak for themselves. If one looks at what South Asia was like, say June 15th, it looked a lot better than it looked May 15th.(84)
What was the influence of nuclear weapons on the resolution of the crisis? A strong case can be made that India and Pakistan were deterred from war in 1990 by the existence of mutual nuclear weapon capabilities and the chance that, no matter what Indian and Pakistani decision-makers said or did, any military clash could escalate to the nuclear level. The case for existential deterrence is admittedly circumstantial; as with all deterrence theory, tracing the causality of non-events is practically impossible. In this instance, one would have to get authoritative Indian or Pakistani officials to admit that they were planning to go to war, but were dissuaded from doing so by the possibility that conventional conflict might escalate to a nuclear exchange. No leaders would do this, for two reasons. First, such an admission would reveal that their country was actually planning to start a war, which would make it look bad internationally; second, backing down from such plans would imply national weakness. Having acknowledged the difficulty of "proving" that India and Pakistan were deterred from war in 1990 by the shadow of nuclear destruction hanging over them, what evidence suggests the plausibility of my thesis?
NUCLEAR DETERRENCE EMBRACED. Senior Pakistani officials began talking about nuclear deterrence in South Asia as early as 1988. Prior to 1990, Indian commentary on the regional nuclear balance was more restrained: few Indian analysts publicly accepted the Pakistani thesis that nuclear weapon capabilities deter war between India and Pakistan. The Kashmir crisis brought about a decisive shift in the Indian nuclear discourse. Although serving officials still do not discuss nuclear deterrence in South Asia, other members of India's strategic elite now do. A quasi-official community, composed mainly of retired civil servants and military officers, has in the last few years closely examined the implications of regional nuclear capabilities for Indian security. For the most part, India's nuclear-strategic thinkers now embrace the idea that nuclear deterrence dampens tendencies toward war between India and Pakistan, and that this phenomenon was especially apparent in 1990.
India's two most prominent nuclear strategists (along with two co-authors) observe that, unlike its predecessor nuclear weapon states, "India has been content to demonstrate capability, put basic infrastructure in place, and leave deterrence implicit and somewhat ambiguous.... It appears that atomic capabilities on both sides in the Indo-Pakistani conflict have so far led to a moderation in actions between the two states."(85) K. Subrahmanyam writes elsewhere:
The awareness on both sides of a nuclear capability that can enable either country to assemble nuclear weapons at short notice induces mutual caution.
This caution is already evident on the part of India. In 1965 when Pakistan carried out its "Operation Gibraltar" and sent in infiltrators, India sent its army across the cease-fire line to destroy the assembly points of the infiltrators. That escalated into a full-scale war. In 1990 when Pakistan once again carried out a massive infiltration of terrorists trained in Pakistan, India tried to deal with the problem on Indian territory and did not send its army into Pakistan-occupied Kashmir.(86)
K. Sundarji concurs: "The chances of a conventional war between India and Pakistan have gone down considerably." He says that while leaders on both sides once viewed war as a means to achieve certain policy objectives, "today I don't think the same calculus can apply." Sundarji adds that because of nuclear deterrence, the menu of Indian responses to Pakistani provocation in Indian-held Kashmir no longer includes launching a bold offensive across the Punjab border. Of Indian leaders, he says: "The reason why they've hesitated to take recourse to their stated, avowed strategy of reacting in the plains conventionally is because of the nuclear equation.... I've got no doubt in my mind at all."(87)
If anything, faith in nuclear deterrence is even more pronounced in Pakistan. Many Pakistanis believe that Gates transmitted a nuclear deterrent threat from Islamabad to New Delhi in May 1990, and that this dissuaded Indian leaders from attacking Pakistan.(88) Sattar told me that he also heard this version of events from "knowledgeable Indians."(89) U.S. officials deny the story,(90) but the more important point is that it is now part of Pakistan's national nuclear lore. Sattar notes the "indispensable contribution" that Pakistan's "nascent nuclear capability has made to deterrence of aggression and maintenance of peace."(91) Now retired, Ghulam Ishaq and beg have publicly stated their belief that Islamabad's nuclear posture has prevented India from attacking Pakistan. Beg says, "Far from talk of nuclear war, there is no danger of even a conventional war between India and Pakistan. . . . As compared to previous years, there is no possibility of an India-Pakistan war now."(92) During her spring 1995 visit to Washington, Prime Minister Benazir Bhutto enunciated Islamabad's nuclear posture: "Our nuclear program is peaceful. But if the existence of our technology and perceived capability has served as a deterrent to India--as a deterrent to a proven nuclear power that has gone to war against us three times in the last 48 years--I certainly have no apologies to make, not in Islamabad, not in New Delhi and not in Washington."(93) In sum, for many South Asian leaders, it has become an article of faith that nuclear weapon capabilities deter war between India and Pakistan.
THE LOGIC OF NONPROLIFERATION UNREALIZED. A second piece of circumstantial evidence that existential deterrence "worked" in 1990 is the fact that war did not break out, despite nearly every element of the logic of nonproliferation being in place. The crisis was a severe political conflict between two opaque proliferants. The struggle over Kashmir directly threatened each state's raison d'etre; moreover, India and Pakistan were plagued by weak readerships and an uncertain international environment. Their nuclear-technical sophistication was limited, which the logic of nonproliferation predicts should have led to intense pressures for preemptive nuclear strikes. A shrill war of words created an overriding impression of impending conflict. Yet war did not erupt. Why? In an opaque nuclear competition, there is simply no way that Indian or Pakistani planners could have confidence in launching an entirely successful nuclear first strike. Even at the height of the crisis, U.S. analysts were split over the precise nature of Pakistan's nuclear capabilities. As one said: "The intelligence community is divided as to whether this is a real threat or just bluff. Some people in the CIA believe Pakistan has nuclear weapons already, others believe they could put a bomb together in two weeks. A third faction thinks they may be as much as six months away from going nuclear."(94) If anything, Islamabad knew even less about India's nuclear weapon capabilities, which were embedded in a sprawling nuclear infrastructure. Given these constraints, a perfectly successful nuclear first strike is a pipe dream.
THE WEAKNESS OF ALTERNATIVE EXPLANATIONS. The existential deterrence thesis is also supported by the relative weakness of alternative explanations for South Asia's "non-war" of 1990. Which of these are most compelling? First, Indian and Pakistani leaders may have been deterred from war in the spring of 1990 by conventional, rather than nuclear, capabilities. India and Pakistan had engaged in a military buildup throughout the 1980s. In particular, the two sides had acquired vastly improved conventional bombing capabilities. That these capabilities have a deterrent effect on Indian and Pakistani leaders is undeniable: even absent nuclear weapons, each side could have launched devastating attacks on the other's nuclear facilities, military installations, industrial assets, transportation infrastructures, and population centers. It is not hard, however, to envision war having erupted in a non-nuclear South Asia in 1990. The 1987 Brasstacks episode illustrated just how close the two sides could come to blows, even after years of conventional arms racing. The main difference after Brasstacks was that Pakistan had emerged in the minds of Indian planners as a real nuclear weapon state. Any lingering ambiguity about Pakistan's status had been dispelled by both the available objective evidence, provided by the nonproliferation community, and the public statements of Pakistani leaders, who signaled at every turn their resolve to deploy nuclear weapons in a future conflict. Elements of both conventional and nuclear deterrence operated in 1990, but the sine qua non of conflict resolution was the nuclear factor.
A second alternative explanation for South Asia's "non-war" of 1990 is that the two sides learned from past experience that war between them does not pay. There is clear evidence that South Asia's revisionist power, Pakistan, had grown more realistic in the 1980s about the costs of war with India. The same explanation cannot be applied to India; if anything, Indian leaders may have learned that in some circumstances, war with Pakistan does pay. India won both the 1965 and 1971 wars with bold offensive operations. Why should Indian leaders have learned that war does not pay? Their lesson may have been that there are certain intolerable circumstances under which the forceful application of offensive military doctrines can ease the security threat from a smaller but determined neighbor. Thus, existential nuclear deterrence provides the most persuasive explanation of why New Delhi did not go on the offensive in 1990.
Third, it might be argued that India and Pakistan were democracies in 1990, and that democracies rarely fight one another. This is the most easily refuted explanation of South Asia's 1990 "non-war." In the first place, Pakistan was nominally a democracy in 1990, but the country's most sensitive national security decisions were made not by Prime Minister Bhutto, but by the army and by a civilian president with close ties to the army. Ghulam Ishaq and Beg were not subject to the pressures of public opinion, parliamentary oversight, and the like, which some scholars suggest are the mechanisms by which democracies desist from fighting one another. Second, democratic practice in both countries exacerbated rather than dampened tendencies toward conflict. Bhutto's initial instincts on the Kashmir crisis were dovish. It was only after sustained pressure from her political opponents that she threatened war in Kashmir. In India, the V.P. Singh government faced relentless pressure from the newly influential BJP and the recently ousted Congress to take a tough stand against Pakistan on Kashmir. As a result, Singh grew increasingly hawkish and began warning Indians to prepare for war. Clearly, then, this variable also fails to capture the essence of conflict resolution in 1990.
In sum, although there can be no mono-causal explanations for complex decisions involving war and peace, existential nuclear deterrence was the most important factor in South Asia's "non-war" of 1990. Indeed, the existence of Indian and Pakistani nuclear weapon capabilities now constitutes the chief foundation for peace in the region.
What are the implications of this analysis for regional stability and the broader phenomena of nuclear proliferation and nuclear deterrence? In South Asia, three futures are conceivable: a reversion to pre-nuclear status; something closely resembling the prevailing situation; or a more overt transition into what might be termed "minimum deterrence" capabilities. The first of these alternatives is politically infeasible: there is little short-term prospect of New Delhi and Islamabad reversing their nuclear strides. Of the remaining possibilities, traditional Western nuclear deterrence theory would suggest that, having arrived as de facto nuclear weapon states, India and Pakistan should now discard opacity and instead strive for assured second-strike capabilities. According to this logic, each should field operational nuclear weapons and delivery systems, and articulate an unambiguous deterrent doctrine. In particular, this reasoning suggests that New Delhi and Islamabad should deploy nuclear-capable ballistic missiles, which would enhance the diversity, and thus the survivability, of their nuclear forces.
Such advice would be misguided. As matters stand today, India and Pakistan derive deterrent security from their nuclear capabilities, while the fact that their weapons are unassembled minimizes the likelihood of nuclear accidents or unauthorized nuclear use. The existing deterrent standoff is a pure, aircraftborne, countervalue balance of terror. Each side implicitly threatens to inflict massive retaliatory punishment on the other side's cities. Absent deployed nuclear weapons and ballistic missiles to deliver them, there is little prospect of either side carrying out counterforce preemptive attacks during moments of crisis. Bombers are vulnerable to interception and too inaccurate to ensure first-strike success. Deploying missiles would increase the likelihood of India and Pakistan adopting counterforce nuclear doctrines, because missiles are invulnerable to interception and may be accurate enough to disable airfields and the opponent's own missiles. First-strike uncertainty would still be a daunting deterrent, but counterforce temptations would be stronger in a strategic "dyad" than in the extant equation.
Although New Delhi and Islamabad have acquired ballistic missile capabilities, neither has yet deployed operational missile systems. However, senior U.S. intelligence analysts believe that such deployments are an immediate possibility.(95) In the interests of South Asian crisis stability, Indian and Pakistani leaders would be wise not to cross this threshold. One solution they might find attractive is a progressive moratorium on missile deployments. Each side would agree to an initial, temporary missile deployment freeze of perhaps three years, thereby formalizing policies they have already chosen tacitly. Missile testing might still be allowed during this initial period. The United States could verify a missile moratorium via National Technical Means, which would temper each side's fear of being surprised by the other's "breakout." Assuming that both countries find this arrangement reassuring, it could then be extended for three or more years, perhaps with an additional agreement not to test ballistic missiles. Eventually, New Delhi and Islamabad may come to realize that their nuclear arms competition is more stable without missile deployments.(96)
Turning to the broader issues of nuclear proliferation and nuclear deterrence, my theoretical analysis and the South Asian case study call into question the utility of the concept of the "reciprocal fear of surprise attack." The notion that nuclear weapon states embroiled in crises will inevitably face strong, perhaps irresistible, pressures to decapitate their opponent's nuclear forces preemptively is deductively appealing but empirically unsupported. Five decades of the nuclear age have now seen nuclear powers weather several serious crises without succumbing to the supposedly inherent logic of preemption. The universe of cases is admittedly small, but my argument is supported by recent research indicating that preemptive attacks of any kind have been historically rarer than conventionally believed.(97) The nuclear era has seen two instances of preventive attacks against nuclear facilities--the 1981 Israeli bombing of Iraq's Osirak nuclear facility and the allied coalition's 1991 air war against Iraq--but both of these actions were taken without fear of nuclear reprisal. In situations where nuclear retaliation has been a possibility, no leader of a nuclear weapon state has chosen to launch a preemptive first strike.
The South Asia case also suggests the need to separate the universal effects of nuclear weapons on world politics from the particular consequences of proliferation in individual regions. The 1990 Indo-Pakistani crisis lends further support to the already impressive evidence that the chief impact of nuclear weapons is to deter war between their possessors. Having said that, it is important to recognize that patterns of proliferation and modes of deterrence will vary across regions. For too long, consideration of these issues has stalled in a quicksand of irresolvable deductive debates that neglect the distinctive historical, political, cultural, and geographical circumstances that shape nuclear behavior in specific regions. Even more troubling, many U.S. analysts continue to view the rest of the world through outdated Cold War lenses, which raises the possibility that the dynamics of regional nuclear competitions may be profoundly misunderstood. U.S. analysts should be prepared to question, modify, or even jettison the models they inherited from their Cold War predecessors. The East-West nuclear deterrence paradigm was the product of a unique historical milieu. Only a careful combination of sound strategic concepts and intensive area studies will enhance cumulative understanding of nuclear weapons' evolving influence on international politics.
(1.) Seymour M. Hersh, "On the Nuclear Edge," New Yorker, March 29, 1993, pp. 56-73. The quotations are on pp. 56-57.
(2.) William E. Burrows and Robert Windrem, Critical Mass: The Dangerous Race for Superweapons in a Fragmenting World (New York: Simon and Schuster, 1994), p. 506.
(3.) Scott D. Sagan, "The Perils of Proliferation: Organization Theory, Deterrence Theory, and the Spread of Nuclear Weapons," International Security, Vol. 18, No. 4 (Spring 1994), pp. 83, 99.
(4.) Hersh, "On the Nuclear Edge," pp. 64 66.
(5.) Major works addressing this debate have included: Leonard Beaton and John Maddox, The Spread of Nuclear Weapons (New York: Praeger, 1962); Richard N. Rosecrance, ed., The Dispersion of Nuclear Weapons: Strategy and Politics (New York: Columbia University Press, 1964); Alastair Buchan, ed., A World of Nuclear Powers? (Englewood Cliffs, N.J.: Prentice-Hall, 1966); William B. Bader, The United States and the Spread of Nuclear Weapons (New York: Pegasus, 1968); George Quester, The Politics of Nuclear Proliferation (Baltimore: Johns Hopkins University Press, 1973); Joseph I. Coffey, ed., "Nuclear Proliferation: Prospects, Problems, and Proposals," special issue, Annals of the American Academy of Political and Social Science, No. 430 (March 1977); Albert Wohlstetter et al., Swords from Plowshares: The Military Potential of Civilian Nuclear Energy (Chicago: University of Chicago Press, 1979); George H. Quester, ed., "Nuclear Proliferation: Breaking the Chain," special issue, International Organization, Vol. 35, No. 1 (Winter 1981); Kenneth N. Waltz, The Spread of Nuclear Weapons: More May Be Better, Adelphi Paper No. 171 (London: International Institute for Strategic Studies [IISS], 1981); Lewis A. Dunn, Controlling the Bomb (New Haven, Conn.: Yale University Press, 1982); Benjamin Frankel, ed., "Opaque Nuclear Proliferation: Methodological and Policy Implications," special issue, Journal of Strategic Studies, Vol. 13, No. 3 (September 1990); Lewis A. Dunn, Containing Nuclear Proliferation, Adelphi Paper No. 263 (London: IISS, 1991); Zachary S. Davis and Benjamin Frankel, eds., "The Proliferation Puzzle: Why Nuclear Weapons Spread (and What Results)," special issue, Security Studies, Vol. 2, Nos. 3/4 (Spring/Summer 1993).
(6.) For a comprehensive descriptive statement of this "logic," see Dunn, Controlling the Bomb, pp. 69-94. Sagan's "Perils of Proliferation" is a recent, more theoretical treatment. See also Steve Fetter, "Ballistic Missiles and Weapons of Mass Destruction: What is the Threat? What Should be Done?" International Security, Vol. 16, No. 1 (Summer 1991), pp. 5-42; and Steven E. Miller, "The Case Against a Ukrainian Nuclear Deterrent," Foreign Affairs, Vol. 72, No. 3 (Summer 1993), pp. 67 80.
(7.) The classic statement of this position is Waltz, More May Be Better. For other works in this vein, see John J. Weltman, "Nuclear Devolution and World Order," World Politics, Vol. 32, No. 2 (January 1980), pp. 169-193; John J. Weltman, "Managing Nuclear Multipolarity," International Security, Vol. 6, No. 3 (Winter 1981/82), pp. 182-194; and John J. Mearsheimer, "The Case for a Ukrainian Nuclear Deterrent," Foreign Affairs, Vol. 72, No. 3 (Summer 1993), pp. 50-66.
(8.) This term was first introduced by Benjamin Frankel in "Notes on the Nuclear Underworld," National Interest, No. 9 (Fall 1987), pp. 122-126. It was elaborated in Avner Cohen and benjamin Frankel, "Opaque Nuclear Proliferation," in Frankel, ed., "Opaque Nuclear Proliferation," pp. 14-44. Cohen and Frankel describe opacity as a "Weberian ideal-type theoretical construct." The specific features of individual opaque proliferants deviate in certain respects from the core concept, but these new nuclear states share most of the following characteristics: they conduct no nuclear tests; they deny possessing nuclear weapons; they make no direct nuclear threats; they conduct no open debate about weapon deployments and doctrine, and announce no nuclear doctrine; they refrain from deploying assembled nuclear weapons; and they insulate their nuclear activities from day-to-day security policymaking.
(9.) See Shlomo Aronson, The Politics and Strategy of Nuclear Weapons in the Middle East: Opacity, Theory, and Reality, 1960-1991 (Albany: State University of New York Press, 1992); and Peter D. Feaver, "Proliferation Optimism and Theories of Nuclear Operations," and Devin T. Hagerty, "The Power of Suggestion: Opaque Proliferation, Existential Deterrence, and the South Asian Nuclear Arms Competition," both in Davis and Frankel, "Proliferation Puzzle"
(10.) Space limitations preclude a discussion of the many other disagreements between these two schools of thought. For a comprehensive examination of these issues, see Devin T. Hagerty, "The Theory and Practice of Nuclear Deterrence in South Asia," Ph.D. dissertation, University of Pennsylvania, 1995, chap. 1.
(11.) Marc Trachtenberg's definition of existential deterrence is the most succinct: "The mere existence of nuclear forces means that, whatever we say or do, there is a certain irreducible risk that an armed conflict might escalate into a nuclear war. The fear of escalation is thus factored into political calculations: faced with this risk, states are more cautious and more prudent than they would otherwise be." See Trachtenberg, "The Influence of Nuclear Weapons in the Cuban Missile Crisis," International Security, Vol. 10, No. 1 (Summer 1985), p. 139.
(12.) Thomas C. Schelling, The Strategy of Conflict (New York: Oxford University Press, 1963), pp. 207-229.
(13.) Fetter, "Weapons of Mass Destruction," p. 29.
(14.) Brad Roberts, "From Nonproliferation to Antiproliferation," International Security, Vol. 18, No. 1 (Summer 1993), p. 159.
(15.) Bruce G. Blair, The Logic of Accidental Nuclear War (Washington, D.C.: Brookings, 1994), pp. 6-8, 19-20; 9-10.
(16.) Waltz, More May Be Better, pp. 16, 15.
(17.) Weltman, "Managing Nuclear Multipolarity," p. 190.
(18.) Thomas A. Keaney and Eliot A. Cohen, Gulf War Air Power Survey: A Summary Report (GWAPS) (Maxwell Air Force Base, Ala.: Air University, Air War College for the U.S. Air Force, 1992), p. 82, quoted in Barry R. Schneider, "Nuclear Proliferation and Counter-Proliferation: Policy Issues and Debates," Mershon International Studies Review, Vol. 38, Supplement 2 (October 1994), p. 226.
(19.) According to a study released by the U.S. House of Representatives Armed Services Committee, the performance of U.S. weapons against Iraqi SCUDs was grossly inflated. The report states that "a postwar review of photographs cannot produce even a single confirmed kill of a SCUD missile." Michael R. Gordon, "On Gulf War, a New Hat is a Headache for Aspin," New York Times, August 16, 1993.
(20.) McGeorge Bundy, Danger and Survival: Choices About the Bomb in the First Fifty Years (New York: Random House, 1988), pp. 401 402.
(21.) Dunn, Controlling the Bomb, p. 75.
(22.) For a similar argument, see Bradley A. Thayer, "The Risk of Nuclear Inadvertence: A Review Essay," Security Studies, Vol. 3, No. 3 (Spring 1994), pp. 448-451, 474.
(23.) On Britain, China, and France, see Avery Goldstein, "Robust and Affordable Security: Some Lessons from the Second-Ranking Powers During the Cold War," Journal of Strategic Studies, Vol. 15, No. 4 (December 1992), pp. 476-527. On India, Israel, and Pakistan, see McGeorge Bundy, William J. Crowe, Jr., and Sidney D. Drell, Reducing Nuclear Danger: The Road Away from the Brink (New York: Council on Foreign Relations Press, 1993), pp. 67-72.
(24.) General Krishnaswami Sundarji "Changing Military Equations in Asia: The Relevance of Nuclear Weapons," paper prepared for the Indo-American Seminar on Non-Proliferation and Technology Transfer, Center for the Advanced Study of India, University of Pennsylvania, October 3-6, 1993, p. 7.
(25.) McGeorge Bundy, "Existential Deterrence and its Consequences," in Douglas MacLean, ed., The Security Gamble: Deterrence Dilemmas in the Nuclear Age (Totowa, N.J.: Rowman and Allanheld, 1984), pp. 8-9. Bundy's view of the nuclear condition is similar to those of Robert Jervis and Bernard Brodie. See Jervis, The Meaning of the Nuclear Revolution: Statecraft and the Prospect of Armageddon (Ithaca, N.Y.: Cornell University Press, 1989), pp. 74-106; and Brodie's War and Politics (New York: Macmillan 1973), pp. 375-32.
(26.) McGeorge Bundy, "The Bishops and the Bomb," New York Review of Books, June 16, 1983, p. 4.
(27.) Robert McNamara, Blundering Into Disaster: Surviving the First Century of the Nuclear Age (New York: Pantheon, 1986), pp. 44-45. On existential deterrence during the missile crisis, see also Trachtenberg, "Nuclear Weapons in the Cuban Missile Crisis," pp. 137-163; and Bundy, Danger and Survival, pp. 391-62.
(28.) Jervis, Nuclear Revolution, p. 47.
(29.) S. Rashid Naim, "Aadhi Raat Ke Baad (`After Midnight')," in Stephen Philip Cohen, ed., Nuclear Proliferation in South Asia: The Prospects for Arms Control (boulder, Colo.: Westview, 1991), pp. 2-32.
(30.) Bundy has expressed to me his agreement with the substance of this argument: "I generally agree with your view of the nuclear standoff in the subcontinent.... As I listen to officials from nuclear-weapon states I find an implicit acceptance of existential deterrence. I don't think it guarantees against all smaller conflict, but it does afford a thickening deterrent to acts of escalation. I find this line of thought consistent with my own assessments of existential deterrence between the superpowers." Correspondence with the author, December 1993.
(31.) Analysts disagree as to the relative weight that should be assigned to the weapons themselves (capabilities), or to the manipulation of the weapons (resolve) in calculations of deterrence. For Edward Rhodes, "the mere existence of an ability to inflict or withhold tremendous pain is logically not sufficient to result in coercive power: the idea of `existential nuclear deterrence'--the notion that simply because nuclear weapons exist, they deter--is logically false. For nuclear deterrence to operate, the opponent must also believe that the coercer is committed to a strategy that has some unacceptable probability of resulting in nuclear war if deterrence fails." See Rhodes, Power and MADness: The Logic of Nuclear Coercion (New York: Columbia University Press, 1989), p. 85. To the contrary, Bundy argues that "the uncertainties which make existential deterrence so powerful have the further consequence that what either government says it might do, or even believes it might do, in the event of open conflict cannot be relied on either by friends or by opponents as a certain predicter of what it would actually do." See Bundy, "Existential Deterrence and its Consequences," p. 9.
(32.) Schelling, Strategy of Conflict, p. 53.
(33.) Neil Joeck, "Tacit Bargaining and Stable Proliferation in South Asia," Journal of Strategic Studies, Vol. 13, No. 3 (September 1990), pp. 77-91.
(34.) Thomas C. Schelling, "Who Will Have the Bomb?" International Security, Vol. 1, No. I (Summer 1976), p. 85.
(35.) For examples, see the remarks by Bush administration CIA Director Robert Gates to the Nixon Library Conference, Washington, D.C., March 12, 1992; the testimony of former Clinton administration CIA Director R. James Woolsey in House Committee on Foreign Affairs, Subcommittee on International Security, International Organizations, and Human Rights, U.S. Security Policy vis-a-vis Rogue Regimes, 103d Cong., 1ts sess., July 28, 1993; and the comments of Clinton administration Under- Secretary of State for International Security Affairs Lynn E. Davis m Michael R. Gordon, "South Asian Lands Pressed on Arms," New York Times, March 23, 1994. Ironically, despite the CIA's unpopularity in South Asia, its judgments are treated by many Indians and Pakistanis as the ultimate "proof" of the adversary's nuclear weapon capabilities.
(36.) On the origins of the Kashmir dispute, see Robert G. Wirsing, India, Pakistan, and the Kashmir Dispute: On Regional Conflict and Its Resolution (New York: St. Martin's, 1994).
(37.) On the roots of insurgency in Kashmir during the 1980s, see W. H. Morris-Jones, "India After Indira: A Tale of Two Legacies," Third World Quarterly, Vol. 7, No. 2 (April 1985), pp. 242-255, James Manor, "Parties and the Party System," in Atul Kohli, ed., India's Democracy: An Analysis of Changing State-Society Relations (Princeton, N.J.: Princeton University Press, 1988), pp. 62-98; "The Enemy Within," India Today (editorial), March 31, 1990, p. 11; Sumit Ganguly, "Avoiding War in Kashmir," Foreign Affairs, Vol. 74, No. 1 (Winter 1990/91), pp. 57-73; M. J. Akbar, Kashmir: Behind the Vale (New Delhi: Viking, 1991); "Crossfire: Kashmir," India Today, August 31, 1991, pp. 77-87; George Fernandes, "India's Policies in Kashmir: An Assessment and Discourse," and Jagat Mehta, "Resolving Kashmir in the International Context of the 1990s," both in Raju G.C. Thomas, ed., Perspectives on Kashmir: The Roots of Conflict in South Asia (Boulder, Colo.: Westview, 1992), pp. 285-296, 388-409.
(38.) Leonard S. Spector, The Undeclared Bomb (Cambridge, Mass.: Ballinger, 1988), pp. 129-130.
(39.) "The Thinking Man's General," India Today, February 15, 1986, p. 78.
(40.) The best study of the Brasstacks crisis is Kanti P. Bajpai et al., Brasstacks and Beyond: Perception and Management of Crisis in South Asia (Urbana: Program in Arms Control, Disarmament, and International Security, University of Illinois, 1995).
(41.) For an analysis of the nuclear dimensions of the crisis, see Hagerty, "Theory and Practice of Nuclear Deterrence in South Asia," chap. 4.
(42.) Interview, 1994.
(43.) Interview with General K. Sundarji, 1993.
(44.) William R. Doerner, "Knocking at the Nuclear Door," Time, March 30, 1987, p. 42.
(45.) Spector, Undeclared Bomb, pp. 94-95.
(46.) Leonard S. Spector, Nuclear Ambitions: The Spread of Nuclear Weapons, 1989-90 (Boulder, Colo.: Westview, 1990), p. 100.
(47.) Spector, Nuclear Ambitions, p. 59.
(48.) S. Viswam and Salamat Ali "Vale of Tears," Far Eastern Economic Review, February 8, 1990, pp. 19-21.
(49.) Malcolm Davidson, "Bhutto Says Pakistan Does Not Want War With India Over Kashmir," Reuter Library Report, February 10, 1994.
(50.) Michael Krepon and Mishi Faruqee, eds., Conflict Prevention and Confidence-Building Measures in South Asia: The 1990 Crisis, Occasional Paper No. 17 (Washington, D.C.: Henry L. Stimson Center, April 1994), p. 6. This document is a transcript of a meeting convened by the Stimson Center to discuss the 1990 crisis. Participants included the U.S. ambassadors to New Delhi and Islamabad in 1990 and several South Asian diplomats and military officers.
(51.) "Kashmir: Echoes of War," Economist, January 27, 1990; "Indian Prime Minister on His Country's Nuclear Policy," Xinhua General Overseas News Service, February 21, 1990; "Iyengar, Ramanna Appointments Open bomb Speculation in India," Nucleonics Week, February 22, 1990.
(52.) "It's All Bluff and Bluster," Economic Times (Bombay), May 18, 1993.
(53.) Mushahid Hussain, "Pakistan 'Responding to Change'," lane's Defence Weekly, October 14, 1989, p. 779. U.S. scholar Stephen P. Cohen contradicts Beg: Pakistan's "strategic style" of offensive-defense has been "remarkably consistent over the years." During crises, Cohen says, "Pakistan has not hesitated to be the first to employ the heavy use of force in order to gain an initial advantage." Cohen, The Pakistan Army (Berkeley: University of California Press, 1984), p. 145.
(54.) "Bluff and Bluster."
(55.) Comments of the U.S. ambassador to New Delhi, William Clark, in Krepon and Faruqee, 1990 Crisis, p. 3.
(56.) Interview with Clark, 1995.
(57.) Krepon and Faruqee, 1990 Crisis, pp. 13-19.
(58.) Moses Manoharan, "Indian Leader Tells Pakistan to Stay Out of Kashmir Uprising," Reuter Library Report, March 13, 1990.
(59.) "Crush Pak Camps: BJP," Times of India, April 8, 1990.
(60.) David Housego, "India Urged to Attack Camps in Pakistan Over Strife in Kashmir," Financial Times, April 9, 1990.
(61.) David Housego and Zafar Meraj, "Indian Premier Warns of Danger of Kashmir War," Financial Times, April 11, 1990; Mark Fineman, "India's Leader Warns of an Attack by Pakistan," Los Angeles Times, April 15, 1990; "VP Urges Nation to Be Ready as Pak Troops Move to Border," Times of India, April 11, 1990.
(62.) James Clad and Salamat Ali, "Will Words Lead to War?" Far Eastern Economic Review, April 26, 1990, pp. 10-11; Malcolm Davidson, "Pakistan Condemns Indian Premier's Talk of War Over Kashmir," Reuter Library Report, April 11, 1990; Malcolm Davidson, "Kashmir Row Sparks Dangerous Period for India and Pakistan," Reuter Library Report, April 13, 1990.
(63.) "Indian Troops Reinforced Near Kashmir Border With Pakistan," Reuter Library Report, April 12, 1990.
(64.) Steve Coll, "Indian Troops, Separatist Violence Aggravate Kashmir Crisis," Washington Post, April 13, 1990.
(65.) Fineman, "India's Leader Warns of an Attack by Pakistan"; "The Makings of a Bloody, Old-Fashioned War," Economist, April 21, 1990; Steve Coll, "Assault on Pakistan Gains Favor in India," Washington Post, April 15, 1990; "Indian Forces Battle Moslems in Kashmir," Chicago Tribune, April 16, 1990.
(66.) Krepon and Faruqee, 1990 Crisis, p. v.
(67.) Hersh, "On the Nuclear Edge," pp. 67-68; comments of Oakley and Clark, in Krepon and Faruqee, 1990 Crisis, pp. 8-9, 4; John E Burns, "U.S. Urges Pakistan to Settle Feud With India Over Kashmir," New York Times, May 21, 1990.
(68.) Interview with former senior Bush administration official, 1993.
(69.) Hersh, "On the Nuclear Edge," pp. 64 66, 57.
(70.) Interview, 1994.
(71.) Hamish McDonald, "Destroyer of Worlds," Far Eastern Economic Review, April 30, 1992, p. 24.
(72.) Krepon and Faruqee, 1990 Crisis, p. 7. The Pressler Amendment requires a cessation of U.S. military and economic assistance to Pakistan unless the president certifies annually that Islamabad does not "possess a nuclear explosive device." President Bush was unable to make this certification in late 1990, precipitating a cutoff of U.S. aid to Pakistan.
(73.) Ibid., p. 40.
(74.) Krepon and Faruqee, 1990 Crisis, pp. 44-46, 20-22. The moderator of the Stimson Center meeting, Michael Krepon, had the participants reiterate their recollections: "As I hear the answer to [the] question, at least with respect to the evacuation of Kahuta and this truck convoy with nuclear material in Baluchistan, and F-16s armed or equipped to carry nuclear weapons, we have categorical 'Nos'." both Jones and Oakley responded, "So far as I know." See p. 45.
(75.) Pervez Hoodbhoy, Nuclear Issues Between India and Pakistan Myths and Realities, Occasional Paper No. 18 (Washington, D.C.: Henry L. Stimson Center, July 1994), pp. 2-3.
(76.) "Bluff and Bluster."
(77.) One possibility that cannot be ruled out is that Islamabad undertook certain actions that were suggestive of nuclear delivery preparations, so that Washington would intervene to resolve the crisis. This would have been consistent with Pakistan's historical tactic of engaging outside powers in its disputes with India as a means of balancing against its more powerful neighbor. For a fuller discussion, see Hagerty, "Theory and Practice of Nuclear Deterrence in South Asia," pp. 258-261.
(78.) Interview, 1995.
(79.) Krepon and Faruqee, 1990 Crisis, pp. 8, 39, 4, 2.
(80.) Interview with U.S. intelligence analyst, 1994.
(81.) Interview, 1993.
(82.) Interview, 1995.
(83.) Krepon and Faruqee, 1990 Crisis, pp. 4, 30-33.
(84.) Interview, 1993.
(85.) M. Granger Morgan, K. Subrahmanyam, K. Sundarji, and Robert M. White, "India and the United States," Washington Quarterly, Vol. 18, No. 2 (Spring 1995), p. 164.
(86.) K. Subrahmanyam, "Capping, Managing, or Eliminating Nuclear Weapons?" in Kanti P. Bajpai and Stephen P. Cohen, eds., South Asia After the Cold War (Boulder, Colo.: Westview, 1993), p. 184.
(87.) Interview, 1993.
(88.) Charles Smith, "Atomic Absurdity," Far Eastern Economic Review, April 30, 1992, p. 25.
(89.) Interview, 1994.
(90.) Interview with Clark, 1995.
(91.) Abdul Sattar, "Reducing Nuclear Dangers in South Asia," Regional Studies (Islamabad), Vol. 8, No. 1 (Winter 1994-95), p. 3.
(92.) See "Ex-President Discusses Nuclear Program, Politics," Foreign Broadcast Information Service Daily Report, Near East and South Asia, July 26, 1993, pp. 69-71; and "General Beg Claims Country Conducted 'Cold' Nuclear Test," Foreign Broadcast Information Service, Daily Report, Near East and South Asia, August 3, 1993, p. 56.
(93.) Thomas W. Lippman and R. Jeffrey Smith, "Bhutto: Deliver F-16s or Return Payment," Washington Post, April 11, 1995.
(94.) James Adams, "Pakistan 'Nuclear War Threat'," Sunday Times (London), May 27,1990.
(95.) Interviews, 1994 and 1995.
(96.) In the long term, Indian and Pakistani leaders will have to decide between retaining nuclear weapon capabilities or giving them up. Each course has costs and benefits: staying nuclear would keep the likelihood of a fourth Indo-Pakistani war low, while running a risk of nuclear accidents or unauthorized nuclear weapon use; nuclear rollback would increase the chances of a future South Asian war, but reduce the likelihood of an inadvertent nuclear catastrophe to zero.
(97.) In a survey of all wars since 1816, Dan Reiter found that only three out of sixty-seven were preemptive in origin. Reiter, "Exploding the Powder Keg Myth: Preemptive Wars Almost Never Happen," International Security, Vol. 20, No. 2 (Fall 1995), pp. 5-34.
Devin T. Hagerty is a Postdoctoral Research Associate with the Program in Arms Control, Disarmament, and International Security and a Visiting Assistant Professor in the Department of Political Science at the University of Illinois, Urbana-Champaign. He received his Ph.D. in Political Science from the University of Pennsylvania in August 1995. This article was written while the author was a National Security Fellow at Harvard University's John M. Olin Institute for Strategic Studies. Additional research support was provided in 1993-94 by the United States Institute of Peace and the American Institute of Pakistan Studies.
The author would like to thank the following people for their helpful comments on earlier manifestations of this article: Stephen P Cohen, Daniel Deudney, Avery Goldstein, Herbert G. Hagerty, and Wendy Patriquin. Thanks also to participants in the Olin Institute's 1994-95 National Security Seminar; a colloquium series on "South Asian Security Issues After the Cold War" at the Watson Institute for International Studies, Brown University, Spring 1995; and a conference on "New Frontiers in Arms Control," at the Center for International and Security Studies, University of Maryland, March 30-31, 1995.
COPYRIGHT 1995 President and Fellows of Harvard College and of the Massachusetts Institute of Technology
COPYRIGHT 1997 Information Access Company