Why this session matters to us
A former head of state used a UNESCO keynote to say something that is not usually said out loud in the building: that global science cooperation is now in a worse state than it was at the height of the Cold War, and that the institutions hosting him are not going to fix it. Then he proposed a remedy, a half-page set of guidelines called the Rules of Science Engagement, and he was explicit that it would have no author, no owner, no enforcement, and no treaty behind it.
That is a governance proposal with nothing to count. It has no indicator, no baseline, and by design nobody accountable for it. Ninety minutes later, on the same stage, the FAO’s chief scientist walked through exactly what separates a UN decade that delivers from one that stalls, and his list was the precise inverse: a named lead entity, a flagship mechanism, mandated reporting. Nobody in the room put those two halves next to each other.
The diagnosis: worse than the Cold War
Grímsson’s case rested on a comparison he is unusually placed to make. He recalled working as a young academic with American and Soviet scientists in the Reagan years, when Jerome Wiesner of MIT, President Kennedy’s science advisor, met the Soviet Academy’s Yevgeny Velikhov and colleagues several times a year for sustained scientific dialogue. Against that, he set the present.
His evidence was his own convening record rather than any dataset. He organised a conference in India last year with several key ministries, and although prominent Chinese scientists were on the invitation list, “they were not granted a visa.” The same thing happened, he said, at a similar conference he organised with India’s then Prime Minister more than a decade ago. He named the United States and China, the Middle East, and relations between some African countries as further instances, and made a point of saying this is not reducible to the war in Ukraine: “This is a deep-rooted global problem.”
The important move came next, and it was a judgement about institutions delivered inside one.
The proposal: five protected categories, and nobody in charge
The Rules of Science Engagement began at a science diplomacy conference in Copenhagen in December, were advanced at a meeting of scientists at the Royal Society in London, were read aloud at UNESCO here, and continue at the Arctic Circle Assembly in Iceland in October. Grímsson read the text from the lectern. It asks that, in times of conflict and military confrontation between states, science cooperation across national boundaries continue unhindered in five areas:
- research on potential natural disasters, a list he ran through as volcanic eruptions, glacial melting, tundra thawing, ocean currents, weather changes, sea level rise, desertification and the transformation of the polar regions;
- “potential health hazards with global pandemic potential”;
- “research on culture, history and languages, especially among indigenous populations”;
- research by students on master’s and PhD theses, and postdoctoral research, “in areas without military or defense applications”;
- “attendance at scientific conferences, seminars or other such gatherings” in service of the above.
Its form is the argument. He stressed that it is “less than half a page” and that the absence of machinery is deliberate.
The model is explicit. “There is nobody enforcing these Chatham House rules,” he said, “but they are being respected all over the world,” and the hope is that the Rules of Science Engagement spread the same way, meeting by meeting and country by country.
The analogy is worth checking, because it is carrying the entire proposal. Chatham House itself insists on the singular, the Chatham House Rule: there is only one. It dates to 1927 and was refined in 1992 and 2002, so it is closer to a century old than the “few decades” Grímsson gave it. And it does a narrower job than the framing implies. The Rule governs attribution: participants may use what they hear, but may not reveal the identity or affiliation of any speaker or participant. It does not govern who may attend, and it asks nothing of states.
That last difference is not pedantry. The Chatham House Rule asks people already in the room to give up something cheap, the right to name who said what, in exchange for something they want, candour. The Rules of Science Engagement ask states to stop doing something they are choosing to do deliberately, at political cost, with a domestic constituency behind it. A visa denial is not an oversight. It is policy. The analogy transfers the mechanism, spread by convention rather than enforcement, but not the incentive structure that made the mechanism work.
Is any of it measurable?
On the terms this site cares about, the proposal is unfalsifiable. There is no indicator attached to it. There is no baseline: nobody publishes how many scientists are refused visas for conferences in a given year, and Grímsson’s own evidence for the problem was two conferences he personally convened. There is no owner, which he says is a feature. There is no reporting line, no review point, and no date by which anyone has agreed to ask whether it worked. In 2033, when the Science Decade closes, there will be no way to say whether the Rules of Science Engagement changed a single visa decision.
The honest counter is that the Chatham House Rule is itself the refutation of the demand. It spread across the world without an enforcer, and Chatham House has never published a compliance rate, because there is nothing to publish. Norms of that kind work by being cheap to honour and embarrassing to break, and an indicator would add nothing. A proposal is not weak merely because it declines to be counted, and the alternative on offer, a negotiated instrument among the same governments Grímsson has just described as the blockage, is not obviously more likely to move a visa queue.
But there is a tension inside the keynote that neither reading dissolves. Grímsson diagnosed the problem as deep-rooted, political and invested, immune to governmental channels for “years if not decades.” His remedy is a voluntary norm that must, in the end, be honoured by governments, because governments are the only bodies that issue visas. His theory of change is that the norm spreads through the scientific community first and binds states later by convention. That is a real theory, and it is untested at the level of state security policy, which is where visa denial actually lives.
One panellist engaged the proposal directly, and it was the historian. Maria Fernanda Rollo called the draft principles important and singled out the fourth: “I like fourth one, protecting the research of young scholars in fields without military application.” She had reason to. She was raised in a country under dictatorship for almost fifty years, where colleagues were dismissed from universities or left because of their politics, and she credited Portugal’s reintegration into international scientific cooperation after 1974 with rebuilding its institutions and restoring academic freedom. “Without freedom there is no criticism,” she said, “and without criticism there is no science worthy of the name.” Nobody else on the panel took the proposal up, and nobody argued against it.
Who brought numbers to the commons
The rest of the panel answered a question about governing the global commons, and most of it was declaratory. Science is “a bridge builder. It’s maybe the last one that is still here” (Brito). “The science is not the truth, the science is what evidence tells us” (Andersen). These are good lines and they are not measurements. Three interventions did carry real counting, and they are worth separating out.
Anil Mishra gave the session its one externally verifiable number. Between 2000 and 2023, he said, humanity lost roughly 6,500 gigatonnes of glacier mass from the polar and high mountain regions. That figure tracks the published community estimate almost exactly, and it is the kind of claim that can be checked against a source rather than a podium. He put 700 million people in coastal regions in the path of the resulting sea level rise, and closed on the line that did the most work in ninety minutes: the world must commit itself to the position that “might is not the right but science is right.” He also quoted Darwin, from 1836: “If the misery of the poor be caused not by laws of nature, but by our institutions, great is our sin.”
Charles Spillane brought the accountability machinery. Asked how the decades might work together, the FAO chief scientist did not answer with a principle. He answered with a mechanism, drawing on the three decades FAO leads or co-leads: family farming (with IFAD), ecosystem restoration (with UNEP), and action on nutrition (with WHO). Some have “measurable results,” he said, while others “have been held back by things like weak funding, unclear targets, and again in some cases no single entity driving accountability.”
He then gave the list, and it reads as a specification for countability: “A clear named lead entity with a joint secretariat. A flagship mechanism to make progress visible. Mandated reporting that forces accountability.” National-level action plans and committees to carry the global into the national, and multi-stakeholder engagement to spread ownership. The evidence he offered was of the same type: the family farming decade credited with supporting 385 policies and legal frameworks across 119 countries and national frameworks in 50 more; the restoration decade with World Restoration Flagships covering 24 initiatives in 18 countries, feeding a global goal of restoring one billion hectares by 2030. The nutrition decade, on his own account, is “a more mixed picture,” with progress on stunting but slower movement on financing and diet quality, and the General Assembly has extended it to 2030.
Spillane also produced the sharpest accountability question of the afternoon, and aimed it at the room. With close to nine million active researchers in the world, he asked, “are they all fully aware that most SDG targets are off track, and that each of them and their networks and communities could play a key role in closing that gap?” He warned that “the commons of science is under pressure,” naming restrictions on data driven by the rise of generative AI, with “shutters coming down in terms of access to data” risking starvation of the very cooperation that produced the knowledge.
Robert Blasiak brought the ocean decade’s working fixes, which are also measurement infrastructure: a dedicated working group on corporate data, on the argument that the private sector holds data useful to the decade’s goals but “it’s sitting on hard drives somewhere”; an intentional push to stop the ocean decade being seen as a UNESCO decade (“it was never that. It was always our decade, everyone’s decade”) by bringing the twenty-plus UN bodies with ocean mandates to the table; and an effort to align ocean philanthropy with the decade’s priorities rather than leaving it ad hoc. On the Cold War point he was with Grímsson and with the history: US and Soviet scientists worked on ocean currents together for over a decade in the 1970s, and US and Cuban scientists shared turtle nesting data across the Gulf of Mexico years before diplomatic ties resumed. “You find example after example of it actually working even in places where nothing should be working.”
Rollo supplied the historical version of the same claim, and the best-evidenced one. The International Geophysical Year of 1957 to 1958 brought 67 nations, the United States and Soviet Union included, into coordinated observation of the atmosphere, the oceans and the poles at the height of the Cold War. “It did not require trust to begin,” she said, and it produced enough of it to make the Antarctic Treaty possible. Her conclusion inverts the usual sequencing: cooperation is not the reward for trust, it is the instrument that manufactures it. She also drew the line that best explains why so little of this panel was countable.
Her diagnosis of the Decade’s own problem was blunt and, for a stocktaking conference, useful: the ocean decade works because “people recognise themselves in it immediately,” a dying reef or a flooded coastline needs no explaining, while science for sustainable development does not land that way. Her prescriptions were concrete: name three to five missions people can actually join; publish the endorsed projects rather than merely launching them, with the global report as “first entry in a continuously updated observatory”; put a recurring, trusted question to universities about what they are doing on the SDGs and get the answers onto the tables where governments decide. On the architecture, she gave the Decade its fairest verdict: “the architecture succeeded, delivery of course lagging.”
Wei Zhang described the China Space Station as shared research infrastructure, with seven projects from twelve countries running through cooperation with the United Nations, a Swiss-developed astronomy facility due to be accommodated on the station, and long-running collaboration with the European Space Agency. Andersen made the anticipation case for GESDA, working with a community she put at about 2,500 scientists from 90 countries, and offered the Open Quantum Institute as a worked example of aligning a frontier technology to six of the SDGs and to use cases in quantum under-served economies before commercial pressure makes the conversation hard. Her framing was the panel’s most honest about why governance keeps losing: science and technology move faster than policy and institutions, so “we cannot go from crisis to crisis, we cannot be in catch-up mode, we have to anticipate indeed.”
Sevestre closed by taking off the moderator’s hat for a testimonial on her 80-day, 4,000km kite and ski traverse of Antarctica, followed by 300,000 children in 42 countries. Their recurring question, she said, was whether the ice would still be there when they grew up. Her closing line was the conference’s best rhetoric and, read strictly, its sharpest accountability claim: children “will probably forgive us for what we did not know, but they will never forgive us for what we knew and did too slowly.”
Our read
Grímsson is probably right about the diagnosis and has the standing to make it, and the panel’s historians and ocean scientists independently confirmed his central historical claim: cooperation across the Cold War’s worst years was real, and today’s is thinner. He may also be right that a short, ownerless norm is the only thing that moves when the formal channels are seized. The Chatham House Rule is a genuine existence proof that unenforced conventions can travel.
The difficulty is that the same properties that let a norm travel also make it impossible to know whether it arrived. An unenforced rule can still work; an uncounted one cannot tell you that it is working. And the strongest evidence against treating that as acceptable came from the same stage, from Spillane, who has run the experiment three times: the decades with a named lead, a flagship and mandated reporting move, and the diffuse ones stall. He was describing UN decades, not norms, and the comparison is not exact. But the Rules of Science Engagement are, on Grímsson’s own description, maximally diffuse: no lead, no flagship, no reporting, by choice.
The cheapest fix is not a treaty and would not compromise the design. Somebody could simply start counting the thing the Rules exist to prevent. A published annual figure for scientists refused visas to international conferences, disaggregated by field and destination, would cost a fraction of what this conference cost, would give the norm a baseline to move against, and would require no government’s consent to compile. UNESCO already produces exactly this class of statistic. The proposal will be back at the Arctic Circle Assembly in October, and the test worth applying then is not whether more people have signed it. It is whether anyone has started counting.
Why it matters for the SDGs
This plenary sat on SDG 17, the partnership goal, and specifically on the science, technology and capacity-building targets (17.6 on international cooperation and access to science, technology and innovation, and 17.16 on the global partnership). The commons the panel named map onto SDG 6 (water), SDG 13 (climate), SDG 14 (oceans) and SDG 15 (land), and the whole panel agreed that none of them can be governed by a country acting alone. That is the easy part, and it was said many times.
The harder part is the one Spillane put to the room, and it belongs on the record: with the world four years from 2030 and most SDG targets off track, close to nine million researchers exist and it is not clear how many of them know either fact. A Science Decade that cannot answer that question about its own constituency is in the same position as the norm proposed from its keynote lectern. It is doing something plausible, at scale, without an instrument that could ever tell it whether the thing is working.
Watch & read
- The Decade’s own stocktake, and the numbers this panel did not use: the launch of the first IDSSD Global Report, where 40% of endorsed initiatives named coordination rather than funding as their main barrier.
- The Chatham House Rule, Chatham House, the text and history behind the keynote’s model (singular, 1927, refined 1992 and 2002).
- The Arctic Circle Assembly, Reykjavík, 8 to 10 October 2026, where Grímsson said the Rules of Science Engagement continue.
- Community estimate of global glacier mass changes from 2000 to 2023, Nature, the published source matching the glacier loss figure cited from the panel.
- The UN Ocean Decade (UNESCO) and the UN Decade on Ecosystem Restoration, two of the decades discussed on the panel as models.
- Full IDSSD Global Conference 2026 coverage.
Quotations are lightly edited from an automated transcript of the UNESCO webcast and should be read as close paraphrase. Speaker labels are not reliable across the recording, so speakers are named here only where the content makes the attribution unambiguous, and by role otherwise. Figures are as speakers stated them and, except where this page says otherwise, were not independently verified. Non-English passages are summarised, not quoted.