Home Arm Control and DisarmamentThe Deterrence That Did Not Fail

The Deterrence That Did Not Fail

by Dr. Zahir Kazmi
0 comments

Context

Nuclear deterrence, we are now told, is not working. Ukraine’s drones have burned Russian strategic bombers on their airfields. Missiles have flown between Iran and Israel. Pakistan and Bharat fought across their frontiers in May 2025. From these episodes Rose Gottemoeller, one of the respected arms controllers of her generation, drew a stark lesson in “The Strange Defeat of Nuclear Deterrence”: “The Ukrainian operation was a spectacular example of a wider trend: nuclear deterrence is not working.” State and non-state actors, she argued, “are in effect calling the bluff of nuclear-armed powers.” The idea has travelled fast. Gideon Rachman wrote that “Putin’s warnings about nuclear weapons are increasingly shrugged off.”

The diagnosis is wrong, and the error is not academic. It mistakes the purpose of nuclear weapons, misreads the evidence its authors cite and, in South Asia above all, risks becoming self-fulfilling. If influential voices tell decision-makers that the bomb no longer restrains war, some will be tempted to test the claim. The most dangerous thing about the “deterrence is failing” thesis is not that it is false. It is that someone might act on it.

This essay makes two propositions. First, the May 2025 conflict demonstrated that unilateral aggression cannot produce a fait accompli against Pakistan. Second, the principal danger now is that Bharat mistakes Pakistan’s restraint for a precedent, to be repeated or enlarged.

A Debate That Answers Itself

Read closely, the debate exposes a crucial distinction. When Gottemoeller’s essay drew formal responses, Eric Edelman and Franklin Miller, veterans of US nuclear policy, wrote that “strategic nuclear deterrence remains alive and well. Nuclear weapons are not and never have been all-purpose deterrents, and few people have ever argued that they can deter conflict at all levels.” On South Asia they were explicit: “it can be reasonably inferred that the prospect of nuclear use moderated strife not only in 2025 but also in 1999, 2001, and 2008.”

Gottemoeller’s reply specified the narrower claim. “My contention,” she wrote, “is not that nuclear deterrence is failing altogether, but that nuclear weapons are growing more unreliable in deterring conventional attacks among neighbours and within regions.” Her original essay also acknowledged that nuclear fear helped bring the May fighting to an end. Janice Gross Stein made the point bluntly: “Pakistan could not deter India.” In her judgement, “Nuclear deterrence is indeed failing when it comes to major powers in the midst of strife with neighbours or regional competitors.”

South Asia does not sustain the broad failure claim. The failure thesis judges deterrence by the first shot. If any conventional or hybrid attack occurs against a nuclear power, deterrence is said to have “failed”. Nuclear weapons cannot be judged solely by that standard. Deterrence was never a promise that a nuclear power would never be attacked, harassed or provoked. Their most consequential effect is to make aggression and escalation carry unacceptable risks. Pakistan’s Full Spectrum Deterrence also seeks to deter aggression across the spectrum. An assessment must therefore examine whether an attack began, what it achieved, how Pakistan answered it and why the fighting stopped.

What the Global Cases Actually Show

Take Ukraine. It is a non-nuclear weapon state outside NATO’s extended deterrence umbrella, and drawing a simile between its conflict with Russia and a nuclear dyad is incorrect. A non-nuclear state fighting a nuclear power, with outside supporters calibrating every transfer, faces a different risk structure from two nuclear powers across a shared frontier. Kyiv’s June 2025 operation against Russian long-range aviation was daring and damaging; Russia did not respond with nuclear weapons. To the failure school, this proves Moscow’s threats are hollow. The Middle East case is likewise irrelevant: Iran is a non-nuclear weapon state at war with two nuclear powers.

Consider, however, what did not happen across more than four years of the longest conflict in Europe since 1945. No NATO state has joined the fighting directly. Western weapons were supplied in increments, debated range by range, precisely because capitals feared escalation. Russia has not struck the logistics hubs on NATO territory through which those weapons flow. The war’s geography, tempo and participants have been shaped by the nuclear shadow: evidence of deterrent effects.

Nor does Russian non-use certify anything for the future. A nuclear power’s restraint in one encounter is a record of that encounter’s circumstances; it is not a licence for the next. Even Rachman conceded that “with the passage of time, it has become apparent that the taboo against the use of nuclear weapons is very powerful.” To call restraint a “bluff called” is to confuse prudence with impotence.

Gottemoeller also speculates that Saudi Arabia’s security agreement with Pakistan “may involve the provision of fissile material or even warhead technology to Riyadh.” Pakistan’s Foreign Office has already addressed such speculation: the Strategic Mutual Defence Agreement “is defensive in nature and not directed against any third country”, and “Any misplaced attribution or speculation of hypothetical scenarios is unwarranted.”

In each case the critics conflate the presence of violence below the strategic level with the absence of deterrence. Gottemoeller herself concedes that in South Asia “the fear of their use has helped bring fighting to a rapid end.” That concession leaves out Pakistan’s agency. Islamabad’s conventional response denied Bharat its objectives; the ceasefire followed that response amid diplomatic engagement.

South Asia: The Case That Proves the Rule

Four Decades of Contraction

The South Asian record is best read over the long run. Lieutenant General Kidwai (Retd) set out that record in his keynote at the Islamabad Non-Proliferation Conference in October 2026. He traced Bharat’s military planning against Pakistan from the “mechanize, mobilize and hit” doctrine behind Exercise Brasstacks in 1986–87, through “limited war while remaining within Pakistan’s nuclear thresholds” in 2001–02, a lukewarm mobilization after Mumbai in 2008, a Cold Start doctrine that he called “still-born”, and the strikes of 2016 and 2019, to the stand-off strikes of May 2025. His conclusion was that this was “a sequential and progressive climb down for India in politico-military choices, clearly indicative of the strategic deterrence effects of Pakistan’s Full Spectrum Deterrence.”

General Kidwai’s account shows repeated Indian efforts to devise options under the nuclear overhang. He observed that even Bharat’s newer Dynamic Response Strategy “is in itself a clear reflection of the limits and constraints imposed by Pakistan’s nuclear capability on India’s strategic and operational options.” Doctrine written to find room for war under the nuclear overhang is itself evidence that the overhang binds.

May 2025: What Actually Happened

The facts of the May 2025 crisis matter, and they are routinely flattened. Gottemoeller writes that Pakistan and Bharat were “attacking far across each other’s borders,” as if the two acted in mirror image. They did not. After the Pahalgam incident, Pakistan denied any involvement. Bharat attributed responsibility to Pakistan without presenting publicly verifiable evidence. Four days after the incident, Prime Minister Sharif offered Pakistan’s participation in “any neutral, transparent and credible investigation,” as Dawn reported. New Delhi neither took the offer nor presented verifiable evidence; on 7 May it used force instead. A year later, President Zardari reiterated that Pahalgam had been a false flag used as a pretext for unprovoked strikes. Even analysts acknowledge the sequence. Two scholars writing for the Carnegie Endowment note that “India’s initiation of hostilities on 7 May confirms that its leaders may have felt they had enough diplomatic cushion to use force.”

The legal point is simple. No unverified allegation creates a right to war or permits a state to become accuser, judge and executioner. International law separates the incident, attribution, state responsibility and lawful response; New Delhi collapsed all four into a single political judgement. Pakistan’s response was taken in self-defence under Article 51 of the United Nations Charter.

That response was a calibrated conventional operation under political authority. It denied Bharat its objectives and imposed a cost greater than the original act, in keeping with Pakistan’s policy of Quid Pro Quo Plus. Within four days, a ceasefire was in place. Islamabad issued no nuclear warning or red-line statement during the crisis. Claims of “nuclear blackmail” or “threshold manipulation” therefore lack support in the public record. As General Kidwai has noted, equating Pakistan’s deterrence posture with nuclear blackmail “reflects a perversion of deterrence theory.”

Applying the Right Test

Now apply the right test. Did a nuclear dyad fight its way to a major war? It did not. The crisis ended short of major war, but its initial attack and Pakistan’s response must both be explained. The sequence matters: Pakistan responded notch-up, Bharat’s objectives were denied, and only then did New Delhi accept the ceasefire. Outside capitals engaged with urgency, but their efforts did not substitute for Pakistan’s response and political decisions.

General Kidwai describes this as a deterrent effect working in two directions at once. Towards Bharat, Pakistan’s capability “becomes a limiting factor on the possible scale, scope, and character of military options that India might consider.” Towards Washington and other capitals, “alarm bells start to ring, compelling diplomatic interventions to cool the emerging crisis.” In his assessment, “India’s desire for seeking ‘space for limited war despite a nuclear overhang’ remained shrunk.”

Others reach compatible conclusions. The same scholars note that “these neighbours have escaped escalation to a major war since they acquired nuclear weapons capability.” Former Indian diplomat, Rakesh Sood, accepts that “when general deterrence fails and conflict erupts, as it often does, nuclear deterrence can help contain the violence,” yet draws a dangerous inference: that Bharat’s unilateral aggression has normalised space for limited war. A IISS study concluded that “The presence of nuclear weapons prevented the crisis from spiralling into all-out war.”

What, then, did “Pakistan could not deter India” actually measure? Only the opening strike. Stein’s sentence identifies the opening strike but omits its outcome. Bharat initiated force; Pakistan’s response denied the result it sought. Any further aggression would have faced escalating costs and risks. There was no safe band of war. What Bharat attempted was not deterrence but compellence, and it failed against a nuclear power.

This is not a new pattern. In 2019, after Pulwama and the Balakot strike, Pakistan answered in kind and returned a captured Indian pilot within days; Pakistan itself was the key stabilising force in preventing a full-scale war. In 2001–02, in 2008, in 2019 and in 2025, the region’s crises have ended short of major war. A theory that calls this record “failure” has redefined success out of existence.

Three Myths

The failure thesis rests on, and feeds, three related myths.

Myth One: Deterrence Is Not Working

The first myth is the headline claim itself, and the evidence above disposes of it. Deterrence is judged by what does not happen, and what does not happen leaves no imagery. A drone strike produces footage; a war that was never escalated produces nothing. The analyst who counts only visible events will always find deterrence “failing”, because its successes are invisible by design. Edelman and Miller grasp this when they insist that nuclear weapons “are still successfully placing bounds on catastrophic aggression by states.”

There is also a selection problem. The failure school draws its cases from moments of crisis and ignores the long periods without major war. Bharat’s own planners have spent two decades designing doctrines for limited war precisely because the nuclear shadow forecloses unlimited war. General Kidwai’s conclusion that “Pakistan’s Full Spectrum Deterrence is alive and effective”, and that it remains in effect “on an hourly and daily basis in deterring a major war imposing strategic caution and operational limit on India”, describes exactly this invisible daily work.

Myth Two: There Is Safe Space for War Below the Nuclear Threshold

The second myth is more dangerous, because it invites action. It holds that a nuclear adversary can be attacked with conventional force, repeatedly, as long as the attacker stays beneath some identifiable line. In Bharat, this delusional view is acquiring the status of doctrine. Two Indian experts wrote in the Indian Express within days of the ceasefire that “the principal takeaway is India’s success in carving out space for conventional conflict below the nuclear threshold.” A study from an Indian think tank went further, claiming that air power and precision weapons offer “compelling options to expand the threshold under the nuclear overhang.”

The image of a “space” is seductive and false. A threshold is not a line on a map, fixed and knowable. It is a political judgement made by a leadership under pressure, in conditions that change with every strike. Morgan et al’s major study of escalation for the US Air Force put it plainly in 2008: “escalation thresholds are social constructs that exist in the minds of the actors rather than in objective reality.” The International Crisis Group has warned that the assumption that limited strikes “will trigger retaliation that will always remain below the nuclear threshold, and that any ensuing conflict will quickly subside,” is “hazardous.”

Nor was the space Bharat claims to have found “safe”. It was bounded by Pakistan’s resolute conventional response and closed within days, with urgent diplomatic intervention and Bharat taking the ceasefire off-ramp. A space that requires the attacker to seek an exit within four days is not controllable. Pakistan’s former National Security Advisor warned, “once the first shot is fired, it is impossible to tell how the crisis will progress.” Short is not safe. New Delhi may initiate a crisis, but it cannot reserve to itself the response, the escalation or the termination. The defender also has a vote.

The gamble is compounded by technology. In May 2025, for the first time, one nuclear power used dual-capable missiles against another. Pakistan told the United Nations that this was the “first-ever use of such capabilities by one nuclear-armed state against another.” An adversary cannot safely assume that a dual-capable cruise missile in flight is conventionally armed; a radar track cannot reveal the warhead, the command chain or the political purpose. The March 2022 BrahMos launch into Pakistan, which New Delhi called accidental while declining a joint inquiry, showed the peacetime version of this danger; May 2025 carried it into wartime. The same entanglement now extends to the sea, as argued in “The nuclear danger at sea”, where Bharat’s nuclear-armed submarines and dual-capable naval missiles share crowded waters. Entanglement in South Asia is built by Bharat’s dual-capable systems and targeting choices; it is not a lever that Pakistan exploits. The faster weapons fly and the more ambiguous their payloads, the less any “space below the threshold” can be measured in advance.

Pakistan’s posture of Full Spectrum Deterrence, within the framework of Credible Minimum Deterrence, exists precisely to deny that space. It is fundamentally defensive in nature and aims to deter aggression across the entire spectrum of conflict. Pakistan’s National Command Authority has consistently framed it as maintained “without entering into an arms race.” Field Marshal Asim Munir stated the governing reality most directly: “there is no space for war in a nuclearised environment.”

Myth Three: Restraint Creates a Repeatable Precedent

The third myth is the subtlest. It holds that because a nuclear power showed restraint once, it will show restraint again, so the experiment can be repeated and perhaps enlarged. After May 2025, this logic was stated openly: Sood observes that Bharat’s response has made “kinetic retaliation the new normal,” and that “the new normal outlined by Modi indicates an upward jump in escalation thresholds.” Dinakar Peri reports Bharat’s Chief of Defence Staff as suggesting that “further expansion of space in conventional operations is possible.”

This is where the “deterrence is failing” thesis does real harm. When Western analysts declare that nuclear weapons no longer deter, they supply intellectual cover for the belief that restraint is a licence. It is not. Holding back from nuclear use lowers the danger of the moment, yet it can raise the danger of the next round if the other side treats a conditional outcome as a standing entitlement. A defender’s restraint reflects that crisis’s circumstances; the attacker cannot assume it fixes the terms of the next.

That Pakistan demonstrated calibrated response under political authority should therefore not be misread as a promise that every aggression will meet the same ceiling. Each new attempt would meet Pakistan’s resolve and prepared capabilities; the initiator cannot preselect the level or limits of Pakistan’s response. Restraint is responsible statecraft, not an endlessly renewable resource. General Kidwai has gone further, recommending that a further “plus” be added to QPQ+.

The record also refutes the idea that 2025 changed Pakistan’s calculus in Bharat’s favour. Sood concedes that Bharat “cannot claim that it has successfully changed Pakistan’s strategic behaviour.” Another analyst concluded in Foreign Affairs that Bharat’s “attempt to reestablish deterrence backfired.” Yusuf and Zeb find that Pakistan’s capabilities and diplomacy “played a role in deterring India from launching a major conventional offensive while leaving it unsettled in terms of its quest to establish the permissibility of limited conflict under the nuclear umbrella.” “Unsettled” is the right word: aggressor sought space for limited war but received a decisive notch-up response.

There is an outside dimension too. The same scholars note that US tolerance after an earlier crisis “would have possibly emboldened the Modi government and signalled new space to flex his country’s military muscle.” Third parties who reward the initiator, or who later explain the initiator’s adventure as proof that deterrence has failed, enlarge the incentive to try again. Gottemoeller’s narrower thesis, that nuclear weapons are becoming “more unreliable” among neighbours, risks doing exactly that. Telling the tester that the limit is gone is the surest way to bring the test closer.

Getting the Lesson Right

Debates about deterrence are not only descriptions of the world. They shape the world they describe. That imposes a responsibility on scholars and officials alike, and it suggests five conclusions.

First, judge deterrence across the entire crisis, from the initiating attack through the defender’s response and termination. May 2025 demonstrated that Pakistan could deny Bharat its objectives while preserving the strategic balance. The analysis must still ask why the fighting stopped by 10 May. Edelman and Miller’s assessment, that the prospect of nuclear use “moderated strife”, should be the starting point of analysis rather than an afterthought.

Second, retire the language of “safe space”. No threshold is a fixed line, and no attacker can know it in advance. Doctrines that seek room for limited war against a nuclear neighbour are bets placed with other people’s lives. Outside powers that wish to prevent the next crisis should say so plainly, rather than treating each limited strike as a new normal to be accommodated.

Third, do not reward the initiator. In May 2025, Bharat used force without publicly establishing Pakistan’s responsibility for Pahalgam. An international norm that unilateral force on unverified claims carries diplomatic costs would do more for stability than any number of essays on deterrence failure. Pakistan’s offer of a neutral investigation in 2025 deserved a serious hearing. As in the past, Islamabad will be resolute in acting in self-defence.

Fourth, take entanglement seriously. Dual-capable cruise missiles, sea-based forces and compressed flight times shrink the margin for error. Pakistan’s long-standing proposal for a Strategic Restraint Regime in South Asia rests on three interlocking elements: conflict resolution, nuclear and missile restraint, and a conventional force balance. The proposal has not stalled for want of merit. It has stalled because New Delhi has refused the sustained dialogue that any such regime requires.

Finally, restore dialogue. Neither bilateral contact nor outside diplomacy replaces credible deterrence or Pakistan’s capacity to respond. Diplomacy for the durable resolution of disputes, including Jammu and Kashmir, should now be allowed to prevent them. As Field Marshal Munir urged, core issues should be settled “as per international norms on the basis of equality and mutual respect.”

Conclusion

Nuclear deterrence has not failed. It has been dangerously tested by those who believe its limits can be found and exploited, and it has held. The danger lies not in its weakness but in the growing conviction, voiced in respectable places, that it is weak. In South Asia that conviction is not a theory to be debated. It is an invitation to try again. In a nuclear dyad, the true enemy is the belief that war can be started low, kept low and stopped on command. Political choice may begin an escalation; interaction decides where it goes. May 2025 left no licence for another experiment. The responsible answer is to state clearly what the record shows: Pakistan’s conventional response and credible strategic deterrent leave no safe space for aggression. The initiator cannot dictate the limits of the next confrontation.

Author: Dr Zahir Kazmi, Arms Control Advisor at the Strategic Plans Division. He can be followed at X @ArmsConAdvisor and reached on zahir.kazmi@gmail.com

You may also like

Leave a Comment