The potential for confrontation between India and Pakistan continues to worry many around the world. The two nuclear powers are highly crisis prone; they have been embroiled in at least three major crises since they declared their nuclear weapons capabilities to the world in 1998. Over the past decade, terrorism on Indian soil has become the number one trigger for Indian-Pakistani crises. The threat still remains clear and present. Prior crises were initiated due to provocative posturing (1987) and even confusion and misperception (1990). These also remain plausible drivers of the next crisis.1
Each Indian-Pakistani crisis implies increased tensions, tit-for-tat brinkmanship, and an inherent risk of escalation. This bodes ill for peace in the region because the most likely scenario leading to a nuclear war is an Indian-Pakistani military escalation caused by a crisis-triggering event. Crisis behavior in the past has tended to bring out the most dangerous elements of the Indian-Pakistani nuclear equation. Concerns emanate from the lack of transparency in nuclear postures and strategies, ambiguous red lines, lack of early-warning capabilities, concerns about the safety and security of the arsenals, structural realities such as geographical proximity, and, not least, a tendency to look to a third party - the United States - to avoid uncontrolled escalation.
This last aspect, namely the expected role of the third party as the principal agent for de-escalation in a nuclear environment, is destabilizing in that it attaches expectations that this outsider may be unable to fulfill. Moreover, it leads the principal parties to avoid institutionalizing bilateral mechanisms for escalation control. An absence of these mechanisms, in turn, makes the third party, in this case the United States, eager to mediate, not only because of its fears of uncontrolled escalation but also because of its important interests in the region. During the Kargil conflict of 1999 and the 2001-2002 military standoff - the two most serious crises since the 1998 nuclear tests - the third-party role was prominent. A significant proportion of the signaling also was routed through third parties. This was most obvious in the 2001-2002 crisis when both sides actively looked to the United States to weigh in on their side and force the other to pull back. In essence, escalation control was "contracted out."
This model of escalation control is especially dangerous in the South Asian context, not only because future crises are believed to be highly likely but also because there is a strong belief among decision-makers on both sides that U.S. diplomatic intervention will be able to keep a lid on escalation. Even more worrisome, they wish for the United States to intervene in support of their position; they want cessation of hostilities, but on their terms. Such a belief has implications, both for escalation control in South Asia and for U.S. policy and actions during future crises. Will the United States be in a position to play the role each side envisages? Even more fundamentally, is its intervention likely to be stabilizing, or will it end up inducing further instability into the equation?
Behavior in Past Crises
The only two nuclear crises since the end of the Cold War that saw active mobilization of military forces on both sides were between India and Pakistan. In 1999, just a year after Indian and Pakistani nuclear tests, Pakistan-backed operatives infiltrated Indian Kashmir and captured strategic heights at Kargil. Fears of escalation to the nuclear level were raised in many global capitals almost instantly. Even though the scope of this confrontation and the use of nuclear signals were limited,2 two aspects of the crisis did have implications for future behavior.
The first was the third party's role: rather than looking to resolve the conflict bilaterally, Pakistan sought external help. Pakistani Prime Minister Nawaz Sharif reached out to China directly, and subsequently to U. …