They’d ridden ponies up to Baisaran because there’s no road to the meadow itself, just a green shelf hanging above Pahalgam that Kashmiris have called ‘mini Switzerland’ for so long the nickname stopped sounding like tourism copy years ago. On the afternoon of 22 April 2025, that isolation was the whole point.
Pakistan sponsored terrorists stepped out from the treeline into a meadow full of families on lunch breaks and pony rides, and witnesses would later tell investigators the attackers took their time sorting targets, asking questions, checking for the physical markers of religion, before opening fire. By the time it was over, 26 civilians lay dead, all but one of them tourists, and Kashmir’s tourism season, its fragile peace, and a 65 year old water treaty were all casualties nobody had budgeted for that morning.
The Resistance Front, a front for the Pakistan-based Lashkar-e‑Taiba, claimed responsibility within hours and retracted the claim within days, a pattern of deniability investigators would later say was itself the point.
Within thirty six hours, the Cabinet Committee on Security met in Delhi under Prime Minister Narendra Modi, and Foreign Secretary Vikram Misri read out a five point response. The Attari land border checkpoint would close immediately, Pakistani nationals travelling under the SAARC visa exemption scheme had forty eight hours to leave the country, and the defence, naval and air advisers at Pakistan’s High Commission were declared persona non grata.
Buried in the middle of that list, between the closure of the Attari checkpoint and the expulsion of Pakistani defence attachés, was a sentence that would occupy international lawyers for the next sixteen months. ‘The Indus Waters Treaty of 1960 will be held in abeyance with immediate effect, until Pakistan credibly and irrevocably abjures its support for cross‑border terrorism.’
To most people, it read like a new and drastic escalation. It wasn’t new. Nine years earlier, after Pakistan-sponsored terrorists killed 18 soldiers at the Uri army base, Prime Minister Modi had chaired a nearly identical review meeting and delivered a line that has followed the treaty ever since, “Blood and water cannot flow together.”
In 2016 it was rhetoric, India reviewed the treaty, set up a task force, and ultimately changed nothing. In 2025, Delhi finally did something with it.
The most common misreading of what happened next, in Indian and international coverage alike, is that India cut-off Pakistan’s water. It didn’t. And in many ways, it couldn’t. Under the treaty, India is allowed only limited, run of the river use of the western rivers, the Indus, Jhelum and Chenab, which together carry roughly 80 percent of the system’s flow to Pakistan. Even if Delhi wanted to, it simply doesn’t have the storage capacity to physically hold back that volume of water.
What actually stopped was procedural. India suspended the sharing of hydrological data, halted meetings of the Permanent Indus Commission, and put on ice the treaty mandated consultation process for its own hydropower projects on the western rivers, including Ratle and Pakal Dul.
Pakistan had already formally objected to both projects on design grounds. Now, with the treaty in abeyance, neither side is using the very channel that would normally resolve such objections, so the projects sit frozen in the same limbo as the treaty itself. One Indian foreign policy commentary put it precisely: this was using law, not hydrology, as the weapon.
A calibrated administrative squeeze, not the kind of crude, unpredictable leverage Arunachal Pradesh’s chief minister accused China of wielding upstream on the Brahmaputra in the previous article of this series. That distinction matters, because it underpins the entire legal argument that follows, India did not commit an act of hydrological warfare. It withdrew cooperation from a legal architecture, which is a different, and far more defensible, move.
It also matters because the treaty was deliberately designed to make hydrological warfare by India close to impossible in the first place. Annexures D and E of the 1960 agreement cap how much general storage India may build on the western rivers at just a few million acre feet, against a system that carries well over a hundred million acre feet of flow every year, a design choice the treaty’s own drafters made precisely so that neither government could later turn the rivers into a weapon.
Whatever India intended by holding the treaty in abeyance, physically choking Pakistan’s water supply was never actually on the table; the infrastructure for it does not exist, and cannot be built quickly even if Delhi wanted to try. What India suspended was information and process, not flow.
The usual pitch for the Indus Waters Treaty is that it’s uniquely tough. It kept working through the wars of 1965 and 1971 and the Kargil conflict of 1999, a record water diplomats point to as proof that technical cooperation can outlast almost any crisis. That record is real but it’s also the wrong lesson.
The treaty took nine years to negotiate, under World Bank mediation led by its president Eugene Black, at a Cold War moment when the core problem between the two governments was seen as essentially engineering, how to split six rivers between two new states without one side flooding or starving the other’s farmland. What that history actually shows is that the treaty was designed to survive declared, conventional war, set piece conflicts with clear fronts, casualties acknowledged by both sides, and definite beginnings and endings, because that was the only kind of India‑Pakistan conflict its 1960 negotiators, working in the early Cold War with World Bank mediation, had any reason to anticipate.
No one in that room was drafting a treaty meant to endure a state using deniable proxies to kill civilians on a hillside every few years while officially denying involvement each time.
That kind of conflict simply didn’t exist as a recognised category of statecraft in 1960. It has since become Pakistan’s default, from the 2001 Parliament attack to the 2008 Mumbai siege, the 2016 Uri raid, the 2019 Pulwama bombing, and now Pahalgam, each traced to a Pakistan based or Pakistan linked terrorist group, and each met with an official denial from Islamabad.
The treaty’s ability to ride out three wars was never proof that it could handle this pattern, only that it had never been tested by it. India isn’t breaking the treaty’s internal logic by treating Pahalgam differently from a border skirmish. It’s revealing that the logic was never built to cover the case now in front of it.
This is where the argument has to leave rhetoric behind and get into actual legal work, because ‘it’s unfair to expect a 1960 treaty to anticipate 2025’ is a historical point, not a legal justification. Three separate doctrines carry that burden, listed here in ascending order of how far India has to reach to lean on them.
The first is Article 60 of the Vienna Convention on the Law of Treaties, which lets a party suspend or terminate a treaty if the other side commits a ‘material breach,’ that is, violates a provision essential to the treaty’s object and purpose.
The Indus Waters Treaty’s text says nothing about terrorism, because in 1960 nothing required it to, but treaty interpretation has never been confined to the four corners of a document when the question is what the parties’ obligations actually assume.
Negotiated, in the words of its own preamble, in ‘a spirit of goodwill and friendship,’ the treaty’s entire operating model, joint commissioners meeting twice a year, routine data sharing, disputes resolved through technical channels, rests on a baseline of good faith cooperation between the two states. A state that repeatedly sponsors lethal attacks on the other party’s civilians can plausibly be said to be in continuous material breach of that underlying condition, treaty text or not.
The second and the harder one to satisfy, is Article 62. Fundamental change of circumstances, still sometimes referred to by international lawyers under its old Latin name, rebus sic stantibus.
It kicks in only when a change was unforeseen at the time of signing and radically transforms the extent of the obligations still to be performed. Looked at one by one, none of the changes since 1960 clears that high bar. Taken together, though, a nuclearised rivalry that didn’t exist in 1960, decades of Pakistan based terrorism targeting India, and a security environment in which the Financial Action Task Force has repeatedly scrutinised Pakistan’s record on terror financing, with India pushing to keep it under renewed watch, the case becomes considerably harder to dismiss.
Even then, as the treaty’s own defenders note, Article 62 has succeeded in international practice only rarely, and tribunals have historically interpreted the radical transformation requirement narrowly, precisely to prevent states from treating changed circumstances as a catch all exit clause.
The third line of argument is necessity, codified in Article 25 of the International Law Commission’s articles on State Responsibility, a state may set aside an obligation if it’s the only way to protect an essential interest from a grave and imminent peril, and if the response doesn’t itself seriously impair another state’s essential interest. Both prongs of that test deserve serious attention, not a quick pass.
Is a security establishment openly debating whether to revisit a six‑decade old water treaty really a ‘grave and imminent peril,’ or an overreaction dressed up in legal language? And does suspending data‑sharing and commission meetings, rather than the water itself, truly avoid seriously impairing Pakistan’s interests, or does it just postpone the harm while making planning harder? These are precisely the kinds of questions a neutral tribunal is meant to weigh, which is exactly why India’s refusal to allow one to do so drains the necessity argument of real credibility, whatever its underlying merits.
This is where readers of Article 1 should feel a jolt of recognition, because the ICJ’s 1997 Gabčíkovo‑Nagymaros ruling already tested this very doctrine, and rejected it. Hungary tried to invoke necessity to walk away from a joint dam project on ecological grounds, and the Court held the threshold too high. Hungary’s own conduct had helped create the circumstances it labelled an emergency, and a state cannot use necessity as a licence to simply ditch treaty obligations it finds inconvenient.
India’s position is narrower in a way that matters, it isn’t diverting a single drop of water or abandoning a project outright, as Hungary did. It has suspended a set of procedural obligations, data, meetings, sign offs, while leaving the physical treaty architecture, and Pakistan’s actual water allocation, untouched. Necessity is a strict test. India’s abeyance is, at the very least, the narrowest possible version of the argument that failed at Gabčíkovo‑Nagymaros.
Here’s the argument no one in this debate has really made, and it doesn’t come from water law at all. It comes from trade law, which cracked this exact problem two generations ago and never told the water world about it.
GATT Article XXI lets any WTO member set aside its trade obligations for measures which it ‘considers necessary for the protection of its essential security interests’ during an emergency in international relations. The key phrase is ‘which it considers’. In principle, that makes the test self judging, though WTO panels have since said they still retain some power to check that a state’s claim of emergency isn’t just made in bad faith.
States have leaned on this clause more than once in precisely this kind of situation. Russia invoked it against Ukraine over transit restrictions, in a case that produced the first real WTO ruling on the provision. The United States used it to justify steel and aluminium tariffs. Most tellingly for this argument, Saudi Arabia and Bahrain invoked it against Qatar in 2017, cutting off commercial and diplomatic ties and citing, not coincidentally, Qatar’s alleged support for terrorism, in a WTO dispute that dragged on for three years before a panel actually got into the merits.
In other words, trade law has already built and repeatedly stress-tested a legal home for exactly the situation India found itself in after Pahalgam, a state that believes it’s entitled to suspend cooperation with a neighbour it accuses of sponsoring terrorism against it. Water law has no equivalent clause, not in the Indus Waters Treaty, and not in the customary law around it.
India’s move to hold the treaty in abeyance is best read not as a treaty violation hunting for an excuse, but as a live demonstration of the exception every trans boundary water treaty should probably have baked in from the start, forty years before India actually needed one.
None of this is just theory for India, because the country has already had its conduct as an upper riparian tested at The Hague, and it passed. Pakistan’s objection to the Kishanganga project wasn’t about procedure, it was about substance. India’s design diverted water out of the Kishanganga/Neelum river entirely, sent it through a tunnel into a different river system, and used it to generate power, about as assertive a use of an upstream position as the treaty permits.
In 2013, a Permanent Court of Arbitration tribunal looked closely at that design on its merits and allowed India to go ahead with the diversion for power generation, subject only to maintaining a minimum environmental flow of 9 cubic metres per second below the plant. When a Hague panel actually examined how India behaves as the upstream state, not what India says, but what it builds, the record it found was of a country working within its rights, not breaking them. It’s worth noting that this is exactly the kind of case the treaty’s dispute‑resolution system was built to handle well. A technical disagreement, argued on engineering grounds, resolved by a tribunal both sides had accepted from the start.
The August 2026 ruling raises a different question altogether, not whether India’s water projects are lawful, but whether India can unilaterally declare a treaty in abeyance. On that point, the tribunal ruled unanimously against India, holding that the 1960 treaty remains fully in force regardless of Delhi’s April 2025 declaration.
It’s easy to caricature India’s rejection of that finding as sour grapes, the substantive objection beneath it deserves more than that. New Delhi has never recognised this particular Court of Arbitration, arguing that the treaty’s own dispute resolution architecture is sequential: a Neutral Expert handles technical questions first, and only if that fails does the matter escalate to full arbitration.
Pakistan effectively ran two tracks on the Kishanganga and Ratle designs at the same time, risking two forums reaching contradictory answers on the same engineering questions. Notably, India has continued to engage with the Neutral Expert process throughout this dispute even while boycotting the Court of Arbitration entirely. A distinction almost never mentioned in coverage that paints India as simply refusing to show up to international law. It’s closer to a state insisting on one forum while rejecting a second, parallel one it considers procedurally improper- a technical objection, not a blanket rejection of adjudication.
The tribunal had already brushed past a version of this argument once, issuing a Supplemental Award on Competence in June 2025 asserting jurisdiction regardless of India’s position. Whether that procedural objection would have persuaded a neutral observer is a fair question. At the very least, it’s a lawyer’s objection, not a slogan.
Start with this, a material breach, argued straight from the treaty’s own preamble. A fundamental change of circumstances, built up carefully across sixty-five years. Necessity, framed in its narrowest, most defensible form, and kept clearly separate from the version that failed in 1997. A trade law precedent showing that exactly this kind of security carve-out has been tested, survived scrutiny, and been invoked by everyone from Washington to Riyadh. And a personal record at The Hague, from Kishanganga, showing India as a state that builds within the rules rather than around them.
Every serious doctrine available was assembled and argued at full strength, and in August 2026, a unanimous tribunal still ruled against India. That outcome should trouble anyone inclined to believe international law reliably rewards the better argument. Because on the merits, India plausibly had one, drawn from four distinct legal traditions, none of them invented for the occasion, each tested seriously rather than merely asserted.
It should trouble Pakistan’s farmers on the Indus plains rather less than Islamabad’s rhetoric suggests, since not one cusec of water has actually stopped flowing. Abeyance, whatever a tribunal calls it, has so far cost Punjab and Sindh nothing but uncertainty about tomorrow, which is a strange thing to call an act of war. But uncertainty is not nothing. And a state that has just watched its most carefully constructed legal case fail against a neighbour it considers a serial aggressor is left with an uncomfortable question, the very question this series exists to answer.
If the doctrine doesn’t hold even when it’s argued correctly, across four separate legal traditions, against a neighbour with a demonstrable record of the conduct being complained about, what, other than a tribunal’s opinion, is actually going to keep a river, any river, on any border India shares, safe the next time blood and water end up in the same argument again?
The next article will address that.
In Chinese political culture, the ability to master a river has long served as a stand‑in for the legitimacy of the state, long before hydropower even existed as a technology.
From extreme orographic cloudbursts to thawing permafrost triggering massive rock-ice avalanches, the ‘Third Pole’ is facing an unprecedented ecological emergency. This deep dive explores the thermodynamic, geological, and human forces making the Himalayas increasingly dangerous.
Ask who owns a stretch of national highway, or an embankment, or the tunnel wall at Silkyara. The honest answer is often several departments, none of whom will admit it first. A road, a bridge, a drainage channel should each belong to one named office, the kind that gets credit when it holds and blame when it doesn’t.
Selected issues, clear explanations and thoughtful perspectives—delivered when we publish.
