The theory and practice of deterrence between the United States and the Soviet Union were predicated on the unitary, rational-actor model. Life-or-death decisions in deep crises would be made by leaders driven by rational calculations of national interest and executed by coherent chains of command. The unitary, rational-actor model, from which the constructs of escalation control and escalation dominance were built, was perhaps the greatest intellectual conceit of the Cold War.

Rational decision-making requires sufficient, real-time information, awareness of risks, and control over all the pieces on the chessboard. The stolid line-up of leaders atop the Kremlin Wall would act as one, even if they didn’t think as one. When core members of the Politburo met, no matter how paranoid, they wouldn’t make foolhardy decisions. Nor would the White House and the National Security Council. Both Superpowers would be able to decipher messages and actions rationally in extremis, avoiding catastrophic choices.

The 50th anniversary of the Cuban missile crisis and the rehashing of its particulars remind us that these assumptions were heroic. They are even more tenuous in South Asia, where the chain of escalation begins with mass-casualty attacks on iconic Indian targets planned by extremist organizations that enjoy safe havens on Pakistan’s soil. The number one offender so far has been Lashkar-e-Taiba. Pakistan’s military and intelligence services acknowledge continuing linkages to the LeT and other shadowy, mutating organizations, in order to keep tabs on their activities. This makes sense.

Assistance provided to these groups for operations carried out against Indian targets is another matter entirely. Intercepted transcripts of conversations during the 2008 Mumbai attacks clarify direct involvement by ISI operatives. These transcripts may not be acceptable for prosecution in Pakistani courts, but they are more than sufficient in the court of international opinion. Whenever spectacular acts of terrorism have a “made in Pakistan” label, Pakistan loses without India’s firing a shot. It doesn’t matter whether active-duty or retired ISI operatives are involved. It doesn’t matter that determining prior knowledge of the operation up the military chain of command is guesswork. Either way, Pakistan’s military and intelligence communities are culpable or incapable of preventing these attacks. Either way, Pakistan loses international standing and its prospects for economic recovery recede.

Pakistan is a non-unitary state. The civil-military divide runs deep. Leadership and the procedures for its safekeeping and transfer are not settled. Governance is poor at multiple levels. Swaths of the country are beyond the writ of the state. Regional and sectarian divisions are worsening. Horrific Muslim-on-Muslim violence is a daily occurrence.

How, under these circumstances, can deterrence stability be strengthened in South Asia? Stimson has just published the first in a series of analytical and prescriptive essays around this topic. The author is George Perkovich of the Carnegie Endowment. Readers will not be disappointed in George’s essay, “The Non-Unitary Model and Deterrence Stability in South Asia,” which can be found on Stimson’s website. Here’s a sampler:

Within the large frame of deterrence stability, a vital subsidiary concept is that a state cannot be a responsible possessor of nuclear weapons if it does not have sovereign control over organized perpetrators of international violence operating from its territory. The absence of such sovereign control impedes efforts by state authorities to ensure national preservation and minimize risks of escalatory conflict that risk annihilation. To put it colloquially, US officials could say to Pakistanis, “We do not challenge your possession of nuclear weapons. Our objective is to promote in any way we can the responsible management of nuclear forces. First and foremost, this means sovereign control over all organizations that can project violence from your territory which is also an obligation under international law. Second, and relatedly, it means you should not tolerate acts that could start wars with other nuclear-armed states, because that would be suicidal and therefore irrational. Given the global implications of nuclear war and the breaking of the nuclear taboo, all states have a shared interest in Pakistan’s coherence, sovereignty, and responsible nuclear stewardship.”

…The challenge is enormous, obviously, but it is not impossible due to the vital fact that India does not harbor offensive intentions toward Pakistan. India does not covet territory that Pakistan controls. India does not wish for Pakistan to be dismembered. Indian leaders recognize that it is in their country’s interest for Pakistan to develop economically, to democratize politically, and to live in peace. India does not want Pakistan’s problems to spill over into its territory or restive Muslim populations. The two countries diverge in their visions for an ideal political outcome in Afghanistan, but could settle for an Afghan state that does not allow itself to be a base for hostile actions against Pakistan and India. The fundamental point is that India will not be a military or security threat to Pakistan if Pakistan will cease pursuing offensive strategies (albeit of a low-intensity nature) against it.