Indian Prime Minister Narendra Modi’s recent public references to Balochistan and Gilgit-Baltistan confirm for many Pakistanis what they have long suspected: that India is employing covert means to destabilise and foment violence in Pakistan. If India is pursuing covert operations to punish Pakistan, it would be a disturbing development in the nearly seventy-year security competition between the two states; but it should not come as a great surprise.

The prospect of covert Indian operations and rhetoric regarding Balochistan and Kashmir widens the front that Pakistani leaders must defend. While alarming to Pakistan, these tactics reflect the Indian leaders’ attempts to find ways to motivate the Pakistani establishment to demonstrably renounce anti-India terrorism and to neutralise actors that threaten to conduct it.

“It’s a Rubik’s cube — dealing with Pakistan,” a former Indian national security adviser told us in late 2014. “You keep fiddling with squares. As you move one set, others are affected or become problems.” Since the terrorist attacks in New Delhi and Jammu in late 2001 and early 2002, in Mumbai in November 2008 and, most recently, on the Pathankot air base, Indians have increasingly concluded that eschewing forceful responses does not work. As one retired senior Indian military officer lamented to us in 2014: “How do you prove deterrence if you don’t use force at any time?” Modi won the 2014 election in part because he displayed his resolve to use force to fight threats against India.

It is notable that no theories in the existing international relations literature, or in other states’ practices, offer guidance as to how India and Pakistan could most effectively proceed here.

Pakistan’s conventional and nuclear forces make Indian conventional military operations against Pakistan exceedingly risky. Indian leaders are trying to find alternatives that could simultaneously satisfy domestic demands to punish Pakistan, deter Pakistan from escalating conflict in reaction to Indian punitive actions, and bring conflict to a close in ways that do not leave India worse off — in terms of casualties, costs and overall power. Pakistani officials and strategists, too, should have a keen interest in understanding these alternatives.

Our focus on possible Indian options for changing Pakistan’s behaviour regarding terrorism does not ignore Pakistan’s legitimate interest in motivating India to redress the grievances of Kashmiri Muslims and create conditions in and around Kashmir that are acceptable to Kashmir, Pakistan and India. The best solution would be for both states to eschew violence against each other and to take reciprocal, verifiable steps to demonstrate to each other that they are doing so. Indeed, Pakistani experts as diverse as Munir Akram and Pervez Hoodbhoy, in recent contributions to the daily Dawn, have sketched non-violent steps that Pakistan and Kashmiris, on both sides of the Line of Control (LoC), could take to increase pressure on India to pacify the situation.

Akram urged Pakistan to “launch a major diplomatic offensive in international forums and the world’s capitals to halt India’s massive human rights violations in occupied Kashmir,” and to revitalise “the legitimacy of the Kashmiri struggle” as something distinct from “terrorism”. Hoodbhoy noted that Pakistan could increase international support by highlighting the indigenisation of the Kashmir movement and “cracking down upon Kashmir-oriented militant groups still operating from its soil.” Combined, such developments would strengthen Pakistan’s position in talks with India and with outside powers.

Yet, as long as Pakistan and groups such as Lashkar-e-Taiba (LT) do not clearly demonstrate their renunciation of cross-border violence, and India does not demonstrate that it will reciprocate by accommodating the interests of reasonable Kashmiri stakeholders in a peace process, more violence with the potential to escalate the conflict remains all too possible. This is why we have written Not War, Not Peace? — to analyse the implications of possible Indian policies and capabilities to deter and to respond to another major terrorist attack on India. At stake is the potential for war that could escalate to nuclear devastation of Pakistan and India. This would be the most destabilising and catastrophic event in the international system since World War II.

For a problem this profound, it is notable that no theories in the existing international relations literature, or in other states’ practices, offer guidance as to how India and Pakistan could most effectively proceed here. Unlike any other nuclear-armed antagonists, India and Pakistan directly border each other, have unresolved territorial disputes (Kashmir and Sir Creek), and have engaged in armed conflict four times, not to mention multiple other militarised crises in places such as Siachen and across the LoC in Kashmir. Furthermore, terrorism poses a threat that could instigate future conflict. Studies on deterring and defeating terrorism have not addressed situations in which the major antagonists possess nuclear weapons. Theories and case studies of nuclear deterrence and escalation management have not involved cases in which terrorists are the instigators of aggression and may not directly be under the control of state leaders.

India cannot reasonably expect that Pakistani authorities will be willing and able to destroy groups, such as LT, and simultaneously eradicate the numerous militant groups that now threaten the internal security of Pakistan more directly. “They can’t stop everybody,” a senior Indian official acknowledged to us in 2014, “but we need to know they are trying … via signals we can mutually understand.” In other words, the reasonable objective is for Pakistan to make demonstrable, persistent efforts to pacify the tactics Pakistan-based actors use to pursue their political demands towards India, regarding Kashmir and other issues.

The more damage India might inflict on the Pakistani military, the greater the probability that Pakistan would resort to nuclear weapons.

India’s primary coercive options could centre on army incursions, or more limited airborne strikes, or covert operations. India’s development of operational military and intelligence capabilities to support these options aims to deter cross-border terrorism through the threat of future punishment. Depending on which of these options India pursues, nuclear strategy and capabilities would play a reinforcing role. For example, if Indian leaders decided to unleash major ground and air operations – as envisioned in ‘Cold Start’ – they would have to anticipate possible Pakistani nuclear responses and deploy more credible nuclear forces and plans to counter Pakistan than the current Indian doctrine of ‘massive retaliation’ implies. Since these are the options most discussed in India, they require deep analysis.

To optimise the potential of any strategy, Indian policymaking processes and military-diplomatic capabilities need to be improved. Meanwhile, our extensive analysis suggests that none of India’s most likely options – army-centric, air-centric, covert and nuclear – could confidently achieve the desired changes in Pakistani behaviour with acceptable risks to India. This suggests that India could channel more effort into developing capabilities and strategies to exert non-violent pressure on Pakistan to prevent cross-border terrorism.

Army-based operations, that would damage the Pakistani military enough to (theoretically) motivate leaders to curtail terrorist threats against India, would probably also reduce the capabilities of the Pakistani military and intelligence services to combat terrorist groups. This challenge could grow if Indian incursions drove more people to join anti-India militant groups in Pakistan. And the more damage India might inflict on the Pakistani military, the greater the probability that Pakistan would resort to nuclear weapons, leading to escalatory destruction that would cause much greater harm to India than the terrorist attack that instigated the conflict.

More limited, precise air strikes could entail lesser risk of escalation. However, strikes calibrated to mitigate escalation could signal to Pakistani leaders that India lacks resolve to actually force fundamental changes in Pakistani behaviour. Pakistan has the means to defend its airspace and to mobilise ground forces to widen a conflict in response to Indian air attacks. Finding a sufficient mix of destructiveness and restraint would confront India with challenges that, for example, the United States and Israel have not faced when they have used aircraft and missiles to attack their adversaries. This is not to ignore the potential value of air strikes against terrorist-related targets to satisfy Indian political necessities and mobilise international pressure on Pakistan. Still, whether such gains would durably alter Pakistani behaviour is highly uncertain.

As noted above, changes to the nuclear doctrine might provide a more credible buttress to punitive conventional military operations, especially those that involve major army campaigns in Pakistan. However, that alone would be unlikely to motivate Pakistani leaders to meet India’s counterterrorism demands. Moreover, creating capabilities and options for battlefield nuclear war would raise significant concerns about how such a war could be controlled and terminated. There is no history to draw upon here: no states have ever exchanged nuclear attacks.

Finally, covert or special forces operations might actually degrade the capability of terrorist groups to attack India and could harm Pakistani interests enough to motivate Pakistani authorities to do more to prevent cross-border terrorism against India. Of course, covert Indian operations also could invite Pakistani retaliation, which Indian policymakers acknowledge is a significant vulnerability. In any case, covertness necessitates restraint in claiming credit for such operations.

To the extent that if India’s covert activities in Pakistan became apparent to Pakistan and the wider world, India could lose reputation and political leverage over Pakistan, as many Indian commentators have pointed out in the wake of Modi’s Independence Day speech. An exceptionally experienced counsellor to several Indian prime ministers told us, “It is not in our interest to have people think we are little different from the ISI [Inter-Services Intelligence] ... if Pakistanis assert we are just like them … the international community will say ‘they both do it’, and the pressure falls off Pakistan.”

If the Indian government persists in the belief that it can manage Kashmir as an internal matter, without Pakistan’s negotiated cooperation, New Delhi will be unable to build an international coalition.

Contrary to military options, utilising diplomatic, economic and other means of international censure in a strategy of non-violent compellence may be a better way to motivate Pakistan. The punitive benefits of a non-violent strategy may be less direct than military action but it also comes with far lower risks of an escalating military conflict. With a clear comparative advantage over Pakistan in economic clout and soft power, India could utilise these tools to isolate Pakistan internationally in response to another major terrorist attack. However, in order to be successful with this strategy, India would have to develop greater deftness in international coalition-building. The Indian government’s own behaviour in Kashmir, and willingness to address grievances there, would need to be positive enough to make outside powers feel they will not be accused of hypocrisy if they side with India against Pakistan.

Overall, India and Pakistan are approaching rough symmetry at three levels of competition: covert, conventional and nuclear. One of the countries may be more capable in one or more of these domains, but each has now demonstrated enough capability in all three to deny the other confidence that it can win more than it loses at any level of this violent competition. India drove Pakistani forces out of Kargil but Pakistani conventional and nuclear capabilities prevented India from escalating the war. India mobilised its forces massively after the 2001 attack on the parliament and Pakistan took some steps to curtail terrorist groups but the balance of power made neither of them want to fight.

Despite trying to develop the ‘Cold Start’ doctrine, India did not respond militarily to the 2008 attack on Mumbai; but Pakistani authorities have hinted to us and others that damage to Pakistan’s reputation and vulnerability to Indian destabilisation efforts made Pakistan take unspecified steps that have prevented further Mumbai-like attacks since 2008. Pakistan subsequently also tested a short-range nuclear missile with the stated intention to deter Indian army incursions that might follow another cross-border terror incident. This condition of rough balance and deterrence across the spectrum of conflict amounts to an unstable equilibrium. Any number of actions by leaders and non-officials, taken by mistake or on purpose, could destabilise it.

The basic balance in useable force creates an opportunity for leaders to take steps to stabilise and pacify the India-Pakistan competition. Diplomacy and dealmaking cannot shift balances of power and deterrence but they can solidify them through explicit agreements that clarify expectations and standards of behaviour. Two recent examples demonstrate that bargaining can result in stable outcomes that address the core concerns of contending parties. In August 2016, the government of Colombia and the Revolutionary Armed Forces of Colombia (FARC) signed a comprehensive peace agreement to end the civil war there. This agreement followed years of secret negotiations and a series of interim steps, including mutual paying of reparations to victims of the conflict and a commitment by FARC to end its relationship with the drug trade on which it had relied for funding. In July 2015, after similarly lengthy talks, the P-5 plus Germany reached a landmark nuclear deal with Iran to address concerns that Tehran was pursuing nuclear weapons.

Such agreements – essentially, negotiated accommodations – raise the costs for any authorities that would subsequently violate them. This is all the more relevant when major outside powers have a stake in the stabilisation that has been achieved. China and the United States both have great interests in stability between Pakistan and India. Both could be expected to press India and Pakistan to uphold any agreements, to contribute to fact-finding if there are disputes over compliance, and to reward both states by increasing investment and urging others to do so, when the security establishments in India and Pakistan demonstrate commitment to stabilisation.

Overall, India and Pakistan are approaching rough symmetry at three levels of competition: covert, conventional and nuclear.

Notwithstanding some intermittent high-level diplomatic engagements with Pakistan, including Modi’s own dramatic visit to Lahore in December 2015, the Indian government has toughened its position on Kashmir. This, too, should not be surprising given the doubts voiced by Indian officials about the intentions of the Pakistani security establishment. Yet, if the Indian government persists in the belief that it can manage Kashmir as an internal matter, without Pakistan’s negotiated cooperation, New Delhi will be unable to build an international coalition that would significantly raise the cost for Pakistan of future major attacks on India.

Indeed, by acting as if there is nothing to negotiate with Pakistan, Indian leaders would encourage proponents of violence in Pakistan and discourage international players who would like to fully embrace India, but are reluctant to do so if India insists that they reject Pakistan at the same time. India has the power, the habits of mind and institutions to sustain a war of attrition with Pakistan. But India cannot achieve its ambitions to be a global power if it remains bogged down in such a war.

The analysis presented in Not War, Not Peace? shows that there are no clear solutions that India can unilaterally pursue to end the threat of violence from Pakistan. Some are more likely to be effective, at greater or lesser risk and cost, for India and Pakistan. But only a combination of Indian coercive and non-violent capabilities, paired with a willingness to bargain, can motivate Pakistan to remove the threat of violence. And just as threat of force alone will not work for India, neither will support or tolerance of anti-India terrorism enable Pakistan to get what it wants from India. Both have to demonstrate willingness to compromise through bargaining, which is only possible if both reassure each other that they are eschewing violence. It is up to Indian and Pakistani leaders and societies, with encouragement from the international community, to find a combination that will work for them.

This article has been co-published online by the Herald in Pakistan and The Wire in India.

George Perkovich is the vice president for studies at the Carnegie Endowment for International Peace while Toby Dalton is co-director of the think-tank’s Nuclear Policy Program.