This article is based on BASIC’s latest report, Addressing Future Nuclear Crisis Scenarios in South Asia through a Responsibility-Based Approach, and subsequent report launch event co-hosted by Center for Security, Strategy and Policy Research (CSSPR) at the University of Lahore and BASIC’s Responsibilities and Global Governance Programme.
The report uses two nuclear crisis simulations involving Indian and Pakistani participants to examine how deliberate and inadvertent escalation may unfold in future India-Pakistan crises, highlighting the role of fear, cognitive biases, and misperceptions in driving escalation, and proposes six responsibility-based centred on preventing escalation from occurring in a crisis as well as contributing to crisis prevention in the first place.
The views expressed belong solely to the author of the article and do not necessarily reflect their government’s position, any affiliated institutions, or that of BASIC. BASIC welcomes other contributions on this topic from a range of perspectives, including from Indian experts and authors. Please get in touch if you are interested in submitting an article.
What if the greatest danger in the next India-Pakistan crisis is not that either side deliberately chooses catastrophic escalation, but that both believe they are acting rationally, proportionately and with restraint while jointly producing an outcome neither intended?
This is a key question that emanates from BASIC’s recent report, Addressing Future Nuclear Crisis Scenarios in South Asia through a Responsibility-Based Approach. Based on two Track 2/1.5 crisis simulations involving Indian and Pakistani participants, the report examines how escalation and de-escalation might unfold in a future nuclear crisis. It challenges some of the assumptions about how much control decision-makers may actually exercise once a crisis begins.
It also speaks to the utility of Track 2 deliberations, which are sometimes treated as peripheral to strategy. However, in South Asia, where sustained Indian-Pakistani engagement has steadily contracted, they create a rare space in which assumptions can be exposed before they become decisions.
BASIC used this conduit by placing Indian and Pakistani participants inside simulated crises, asking them to make choices under uncertainty and time pressure. Importantly, these simulations debarred third-party intervention, putting the onus of escalation and de-escalation on them alone.
The report also challenges Herman Kahn’s ladder metaphor, which implies identifiable rungs, deliberate movement and some awareness of where one is standing. It introduces the concept of an escalation cascade: deliberate and inadvertent escalation can occur simultaneously, feeding into one another. What one side regards as proportionate retaliation appears anything but proportionate to the other, providing a rationale for another response.
Apropos of South Asia, this raises a number of questions: What if Islamabad believes it is restoring deterrence while New Delhi believes Pakistan is preparing for the next round? What if India believes it has offered an off-ramp while Pakistan interprets the same offer as coercion? What if each side measures proportionality differently? All this means that escalation is not simply something one state does to another. It becomes something they co-produce.
It identifies another familiar South Asian problem: uncertainty does not necessarily induce caution. Participants became more certain about the adversary’s hostile intentions, viewing their own actions through what the report calls a “peaceful/defensive self-image” (the concept was first developed in Ken Booth and Nicholas Wheeler’s 2008 book, The Security Dilemma and represents an elaboration of ideas developed by Herbert Butterfield and later Robert Jervis). This is a simple way of expressing this: my military action is defensive because I know why I ordered it. Yours is threatening because I cannot know with confidence why you ordered yours. Both India and Pakistan can use this refrain. These findings have implications for deterrence, too, as it operates between human beings who offer biased interpretations of capabilities, doctrines, actions and signals. Therefore, the problem is not simply about credibility but also legibility.
The next India-Pakistan crisis will occur in a completely different technological milieu. Emerging and disruptive technologies, such as cyber weapons, artificial intelligence, drones, long-range precision strike, and dual-capable delivery systems, will create more information, signals and targets, giving less time to determine what any of them means. The report rightly locates future escalation risks in this combination of emerging technologies, cross-domain interaction, and entrenched adversarial perceptions. Therefore, more time should be spent on thinking about escalation comprehension than escalation dominance.
The report’s treatment of fear is also instructive. In the first simulation, fear remained predominantly focused on the adversary. In the second, as participants moved closer to the nuclear threshold, the object of fear shifted. They increasingly feared that the situation itself was becoming uncontrollable, reflecting James Blight’s concept of “nuclear inadvertence” (Blight, J (1990), The Shattered Crystal Ball: Fear and Learning in the Cuban Missile Crisis, Savage, MD: Rowman & Littlefield). That shift contributed to a partial recognition of shared vulnerability and eventually to a tentative ceasefire.
This means that the crisis itself can become a common threat. Simply put, India and Pakistan could move from asking how do I prevail over you to exploring how do we avoid producing an outcome neither of us wants. The report describes it as “cooperative de-escalation”, which should not be conflated with reconciliation. It requires neither friendship nor the resolution of outstanding disputes but a recognition that neither India nor Pakistan can unilaterally control the consequences of an escalating nuclear interaction, making shared vulnerability a strategic resource.
The report challenges another assumption: communication produces de-escalation. In the second simulation, the hotline was used frequently. Yet the crisis escalated rapidly. Participants filtered messages through existing adversarial lenses, hearing the threatening component more clearly than the accompanying reassurance. This demonstrates that a hotline does not necessarily connect two understandings. The quality and timing of communication matter. This is why the report focuses on Security Dilemma Sensibility: decision-makers need not only communicate their intentions, but also consider how their actions are likely to be perceived by an adversary already operating under fear.
Furthermore, the report’s focus on building a common strategic lexicon deserves attention. This is primarily because decision-makers must understand what the other side means by terms such as restraint, proportionate response and strategic stability before missiles are in the air. That said, such a lexicon need not eliminate the ambiguity that inheres in deterrence.
The report also challenges the idea that reversing roles would create empathy. The simulations showed that participants carried their existing assumptions about the adversary into the role they were asked to play. Rather than dissolving the “bad faith” model of the other, altercasting sometimes reinforced it. However, the post-simulation reflections were insightful, as they underlined the difference between the messages communicated and their interpretations. Participants began to recognise how their own actions had contributed to the fear and insecurity of the other. This shows how Track 2 forums allow participants to discover that the image the adversary constructed does not always represent the decision-maker sitting across the table.
Moreover, the report highlights how every India-Pakistan crisis produces two battles. One pertains to the crisis itself while the other is centred on the meaning attached to it. Narratives surrounding deterrence, escalation, and who prevailed are not simply retrospective accounts; they permeate the next crisis. If one side concludes that escalation produced success, it may enter the next confrontation prepared to escalate earlier. If the other side believes its adversary has drawn precisely that lesson, it may feel compelled to deny it that advantage. These divergent narratives of victory could cause future escalation.
Here, the May 2025 crisis becomes relevant. While both sides will not have a shared narrative about the crisis, they need some kind of a shared understanding of risks. If both sides can agree when escalation became difficult to control, what signals were misunderstood, and where restraint occurred, the environment might become less risky. However, if India and Pakistan learn opposite lessons, victory narratives could become escalation scripts.
This makes BASIC’s responsibility-based approach important. It shows that nuclear adversaries have responsibilities that they cannot shirk. This puts a premium on understanding how one’s actions may be perceived, preserving channels to clarify intentions, and protecting systems such as nuclear command and control from actions that generate catastrophic uncertainty.
Considerable emphasis has also been placed on younger Indian and Pakistani scholars and practitioners. The next generation has fewer opportunities for substantive interaction across the border, yet these are the people who will populate the institutions that will manage future crises. These people could answer a harder question: Are India and Pakistan becoming better at deterring one another while becoming worse at understanding one another?
The report identifies how decision-makers cause and manage fast-paced, technologically-denser nuclear crises, highlighting how the process of escalation is becoming more convoluted. Its closer reading brings to the fore one thing: the institutional, military and cognitive capacity to stop an escalation process once it has begun needs to be built before the next crisis, not in the middle of it.
Image credit: MartinVMtl, CC BY-SA 3.0 <https://creativecommons.org/licenses/by-sa/3.0>, via Wikimedia Commons