The Vajont Dam Held. The System Failed. Around 2000 people died.

Vajont cut into me the way it cut into the landscape. The dam stands. The valley below it died. Ever since, I keep returning to one question — not the one most people ask. They ask who was to blame. I keep asking how is it possible that a system can know almost everything and still do almost nothing with what it knows?

That question is not really about 1963.

It is about the distance between knowledge and decision — a distance I spend my working life trying to close inside institutions and projects.

Vajont is the most extreme illustration of that distance I have ever found. It is also, I have come to think, widely misread. So this essay does two things at once: it tells the story, and it interrogates the way the story is usually told — including the way I was first tempted to tell it.

One of twentieth-century Europe's worst infrastructure catastrophes is today one of Friuli Venezia Giulia's most visited sites. Few visitors descend into the deeper layers of a story that is still moving. That deeper, multilayered dimension is what has always held me.

Vajont is one of the rare disasters in which the presumed culprit survived almost untouched. The dam did not fall. The concrete held. Around two thousand people died anyway.

On the night of 9 October 1963, a vast mass of rock — most current estimates put it near 270 million cubic metres, some higher — slid from Monte Toc into the artificial reservoir. It moved as a single body and reached the valley floor in well under a minute. Twenty seconds, the local accounts say; the geological literature is more cautious, but agrees the collapse was complete in less than forty-five seconds — fast almost beyond belief for a mass that size.

The displaced water rose more than a hundred metres up the opposite slope, vaulted the crest of the dam, and poured into the valley. Longarone, the small town below, was almost completely destroyed. And yet the dam remained.

In most disasters an object collapses: a building, a bridge, an aircraft.

At Vajont, the guilty object survived. What failed was harder to see — the system of judgement, communication and decision built around the concrete, and far weaker than it.

Longarone is unusual in its grief as well. Elsewhere, a wind carries the murmuring of mourners over the graves. In Longarone the water carried away whole families at once, so that there was no one left within a family to do the mourning.

The memory is carried instead by the municipality, the cemetery, an extraordinary church, and a long list of names.

The paradox of such disasters is that with each passing year they grow more relevant, not less.

New photographs, documents and testimonies keep surfacing; new analyses overturn old certainties. Vajont does not subside into a statistic — it keeps swelling inside it. Which is exactly why its questions are not questions of the past.

Information is not coherence

Most accounts of Vajont run on a single rail: the story of silenced warnings. The warnings were real.

But if we stop there, we miss the more uncomfortable thing — the thing that matters to anyone who works with decisions, projects and institutions.

Vajont was not a disaster of missing data. It was a disaster of what a system does with data once it has it.

The locals spoke.

A journalist wrote.

A young geologist warned.

A laboratory measured.

A court, before the catastrophe, formally recognised that the danger was well-founded.

In the final days the mountain was visibly moving.

At no level was information actually lacking.

But notice the pattern: at every level, the data had to pass through a system that knew how to make it bearable.

Local fear was reclassified as panic.

A journalistic warning became a threat to public order.

Geological insight was softened into an acceptable report.

A laboratory model was distilled into a single reassuring number.

Responsibility dissolved into hierarchy. At the end of this chain the truth was not denied — it was diluted, until it no longer demanded that anyone act.

This is the difference between information and coherence. Information is what a system knows. Coherence is its capacity to translate what it knows into a decision — especially when the knowledge is unwelcome, expensive, and points toward stopping.

Vajont was information-rich and coherence-dead. That is what makes it not a historical curiosity but a diagnostic model, recognisable in many a modern organisation that mistakes having a dashboard for being able to act on it.

I will trace it through five people. Not as a gallery of heroes and villains — I will argue, in fact, that the heroes-and-villains version is part of the problem — but because each of them stands at one point where knowledge failed to become action.

1. Tina Merlin: when a warning becomes a public-order offence

Tina Merlin did not predict the catastrophe. She wrote down what she heard and saw. In May 1959, as a local correspondent for the newspaper L'Unità, she reported the fear of the people of Erto and Casso. They did not speak the language of geotechnics.

They spoke of cracks, of rumbling inside the mountain, of ground that shifted, of trees no longer standing where they had stood.

They spoke of a mountain they lived with and knew. Merlin took them seriously — and that is precisely what made her dangerous.

History calls her the Cassandra of Vajont.

The name is misleading, and her son, also a journalist, rejected it.

Cassandra foretells the future.

Merlin did something less mystical and more threatening: she described the present, for a system that had already chosen its future and had no room in it for this scenario.

Born in 1926 into a poor peasant family near Belluno, she went into domestic service in Milan at twelve and, during the war, served as a partisan courier. She was not an outside intellectual discovering a poor mountain community; she belonged to the world that the official language of progress treated as an obstacle. A provincial correspondent, not a Rome or Milan name — which is exactly why she saw what the central press did not: the expropriations, the cracks in the houses, the pressure of SADE, the company building the dam.

Her contested article of 5 May 1959 carried the headline SADE lords it as it pleases, but the mountain people defend themselves. She recorded the fear of an artificial basin that, she warned, might flood fertile fields and pull houses into the lake.

The response was swift.

Count Vittorio Cini, the last president of SADE, had her reported for spreading false and tendentious news liable to disturb public order.

The surviving complaint from the Erto carabinieri, dated 8 May 1959 — three days after the article — is, in its bureaucratic dryness, almost more eloquent than anything else: it forwards the article to the public prosecutor while calmly noting that it is not possible to say whether the lake's waters would in fact undermine the ground and swallow the houses of Erto.

The question of whether the danger was real was thus left open — but the proceedings against the woman who had named it began at once.

Here is the whole mechanism in miniature. The company did not prove that the mountain was stable. It proved that it was dangerous to talk about it.

The legal basis was Article 656 of the Penal Code, which punishes publishing news liable to disturb public order. The protected good is not truth but calm. She was not charged with lying. She was charged with disturbing. She later turned this into bitter irony, titling her account of the trial with words that belong above every debate on how institutions handle uncomfortable truth: if only I had managed to disturb public order — because had she truly disturbed it, the lake might have stopped filling.

And here is the part most people do not know.

On 30 November 1960 the court in Milan acquitted her — not on a technicality. The people of Erto and Casso testified in her favour. The ruling stated that the article contained neither false, nor exaggerated, nor tendentious news, since the author had merely exercised the right to report and had voiced a concern with a real basis.

To put it plainly: an Italian court formally acknowledged that the danger was well-founded almost three years before the catastrophe.

The data was not merely available.

It was inscribed in a binding judgement.

And the lake kept filling.

Merlin kept warning afterward, too.

She recorded that on 15 February 1961 — more than two years before the end — the regional council unanimously demanded the revocation of SADE's concessions and timely safety measures, explicitly naming the Vajont situation.

It went unanswered.

The authorities had replied to a warning with a prosecution, and the courts had indicted the messenger without once travelling to the site to check whether she was right.

That the system silenced not only the warning but the memory is told by two fragments. Her interview for French television immediately after the catastrophe was suppressed in Italy under political pressure.

Her book — published only in 1983, twenty years later — found a home long after the urgency had passed.

That same year she refused a medal the municipality of Longarone meant for all journalists who had written about Vajont. It troubled her that those who wrote only after the disaster were being honoured alongside those who wrote before.

The task of journalism, she believed, is to write before the tragedy, when the price of truth is high.

2. Knowledge without the right stamp

Behind Merlin stands something larger than one journalist: an entire layer of knowledge the system could not read.

Before Monte Toc was the subject of geological reports, it was part of the daily life of Erto and Casso.

People crossed it, mowed it, grazed it, named its meadows and ravines and summer huts.

They read the mountain's movement not from instruments but from a change in their own space — a crack that had not been there yesterday, a tilted tree, a rumble underfoot.

This knowledge was not unscientific.

It was scientific in another register: slow, bodily, local, accumulated over generations of living with the mountain.

The system discarded it — not because it was wrong, but because it came from the wrong source.

Here, perhaps, is the deepest lesson of Vajont: a system does not always ignore data because it lacks it. Often it ignores data because the data lacks status.

(Photo: Andrea Bussarlo) Longarone After and Before

The same fact is worth something from a report with the right stamp and nothing from a shepherd. Provincial fear was read as backwardness, not as information.

One of the two surviving firemen of Longarone said, three decades latter, something that holds the whole in a sentence: the danger was known to everyone, but no one could speak, because a kind of omertà prevailed.

Knowledge was not lacking; permission to utter it was.

The unease was already carried in the language of the place. The mountain was called Toc — a name local explanation ties to a piece, but also to something sodden, rotten, decaying. The etymology can be disputed. What cannot be disputed is that the decay of the mountain lived in the speech of the people long before anyone translated it into a technical report. And when the report finally came, it said the same thing. Only no one wanted to believe it.

3. Edoardo Semenza: the truth from the inside — and what we have since learned

The dam was designed by Carlo Semenza, one of the most eminent dam-builders of his time. In March 1959 a small landslide at the nearby Pontesei dam sent up a wave that killed an old caretaker, whose body was never found.

The event frightened Semenza enough that he commissioned a thorough geological study of the valley — from his own son, the young geologist Edoardo Semenza, working under the Austrian expert Leopold Müller.

(Foto: R. Urro) Eduardo Semenza

Edoardo read the slope as something far more serious than the surface instability described by the eminent older geologist Giorgio Dal Piaz. He proposed that beneath Monte Toc lay an enormous ancient slide mass — a paleo-landslide some two kilometres across — that could move as a single body once the reservoir filled.

The disagreement was not academic.

It was the difference between “occasionally a boulder breaks loose” and “the whole flank of the mountain can come down at once.”

His warning, in other words, pointed in exactly the right direction.

Here I have to be honest with my reader, because the science has moved and an essay that leans on the newest data cannot keep to the old story. For half a century Edoardo's reconstruction was the consensus: Vajont as the reactivation of an ancient slide gliding on a thin clay layer at residual strength.

In the last decade that consensus has been seriously challenged.

Re-examinations of the failure surface — most pointedly by Dykes and Bromhead in 2018 — argue that the pre-existing sliding surface may never have existed, and that 1963 was more likely a first-time failure of a largely intact, only marginally stable rock mass.

Others read the “ancient slide” instead as a slow, deep-seated gravitational creep of the slope rather than a single catastrophic ancestor.

The most recent survey of sixty years of research is candid: whether the slide was a first-time failure or a reactivation remains open, and several long-standing assumptions may have biased the conclusions built on them.

What does Semenza tell us, then?

The point of his story was never the precise mechanism. The point is what the system did with a warning it could not absorb.

Edoardo was pressed to soften his findings so as not to obstruct a colossal project.

The builder's own son saw a danger before it arrived and was nudged to file it down to an acceptable shape. SADE never forwarded his analyses to the supervisory authorities. There are strong indications the company knew of the report by August 1959 — a year before the great slide of November 1960, when some 700,000 cubic metres of rock came down into the lake. That slide, too, was not merely a warning.

It was a rehearsal: the mechanism that would destroy the valley three years later, played out in miniature before everyone's eyes.

The data existed. It circulated through internal corridors. It was systematically kept off the official decision path — and even where it circulated, it was minimised.

Müller, in one report, put the slide mass at “around 200 million cubic metres”; the real figure was nearer 270. The underestimation ran not only outward, toward the public and the supervisors, but inward, in confidential documents where there was no one left to reassure but the system itself.

So the mechanism survives the scientific revision intact, and is arguably sharper for it.

Semenza may have had the wrong ancestor and the right alarm.

A coherent system can absorb a correct warning even when the underlying model is imperfect.

An incoherent one cannot absorb even a correct warning — and reaches for the smaller number, the softer phrasing, the unforwarded file.

4. The model that reassured the system

The decisive turn happens in the laboratory, where science was assigned the task of measuring the danger — and instead made it acceptable.

To calculate the wave a possible slide would raise, SADE built a large physical model of the valley, with water, at Nove di Vittorio Veneto.

The experiments were led by Augusto Ghetti, an eminent professor of hydraulics at Padua. On paper, an entirely responsible procedure: instead of guessing, measure.

The scale model of the Vajont basin at the hydraulic model centre in Nove, under the supervision of its director, Augusto Ghetti.

The problem lay in how the model was built.

The slide mass was lowered too slowly, along a sliding surface of the wrong shape.

In July 1961 Edoardo Semenza came to Nove in person and demanded the geometry be changed. He was ignored.

More damning still, a fragment from the court record: when ministerial officials came to view the model, the worst experimental results were kept from them.

The consequence was devastating precisely because it looked rigorous.

In 1962 the experiments produced a single, reassuring number — the wave could overtop the dam only if the water stood less than twenty metres below the crest — and SADE concluded that a margin of twenty-five metres would remove all danger.

An entire strategy for managing a lethal risk was compressed into one parameter: the water level.

Ghetti's own report had named 700 metres above sea level a level of absolute safety. The slide came when the reservoir stood at almost exactly that level.

And here is the darkest irony.

By raising and lowering the level to control the risk, the system kept changing the water pressures inside the very mass it was trying to stabilise.

Modern geomechanics has confirmed: filling raised pore-water pressure at the foot of the slope and reduced its strength; the deformations accelerated precisely at the changes in level.

The water did not only weigh the mountain down — it lubricated the surface it slid on.

The control measure had become part of the problem. When they began lowering the level in the autumn of 1963, the mountain was already in accelerating motion.

Why was the model so wrong?

A full three-dimensional reconstruction of the disaster (Franci and colleagues, 2020) shows that Ghetti's model failed not through one mistake but through two compounding assumptions. First, the laboratory released the eastern and western lobes of the slide separately, never as the single simultaneous mass that actually came down. Second — and this proved the larger error — it used rounded gravel for the sliding material, a substance that shatters and brakes as it falls.

The real mass did not shatter.

It slid as a coherent block and struck the water at a speed no gravel model could reproduce; geologists call such an event a sturzstrom, in which, perversely, the greater the mass the lower the basal friction.

The numerical re-run is merciless on the point: model the right material and the right failure, and 700 metres is not a safe level — dam overtopping occurs even for a water level seventy-five metres lower than the one present on the night.

The wrong choice of material distorted the prediction more than the wrong volume did. This matters for how we read the laboratory's failure.

It would be too easy to say the true physics was simply known all along and wilfully ignored. The more unsettling truth is subtler: the model was not a lie. It was a faithful answer to the wrong question, built on inputs that happened to be commercially and politically comfortable.

Ghetti admitted at the trial that with the correct parameters his experiments would have produced exactly the catastrophe that buried Longarone.

This is where Vajont stops being a story from 1963.

The model was not the opposite of blindness; it was blindness in its most technological, most persuasive form.

When a system does not want to see a danger, it need not reject science. It need only feed science assumptions that are convenient, and let the rigour of the method lend authority to the result.

What comes out is not a lie that looks like a lie.

It is a number, a graph, a protocol — a reassuring form of false certainty.

Every organisation that today swears by dashboards, risk scores and predictive models should be able to read the warning: a model does not deceive only when it is crude. It deceives most when its inputs are comfortable.

5. Mario Pancini: the man who could not carry the system

Marco Pancini

The figure the accounts almost always omit is Mario Pancini, the site director — the man at the end of the chain, where the hierarchy's decisions turn into acts on the ground.

He maintained two things to the end: that an accident on this scale could not have been foreseen, and that he had merely followed his superiors.

Hold those two claims together and you can feel the whole machine in them: no one foresaw, and everyone above me decided.

He told the investigating judge he would kill himself if charged with multiple counts of negligent homicide.

On the morning of 24 November 1968, the day before the trial opened in L'Aquila, his two lawyers waited for him at Venice's Santa Lucia station to travel together.

He never came.

After five years under the weight of two thousand dead, he had taken his own life in his small apartment in Cannaregio — and even that he arranged so the gas would not endanger his neighbours.

On the sixtieth anniversary, the president of the Tina Merlin Association revealed that it was in all likelihood Pancini who had quietly passed internal project information to the journalist from inside SADE.

The man broken at the bottom of the chain may also have been the one who tried, in secret, to let the warning out.

He is the human cost of incoherence: not a villain, but a person asked to hold together with his body a structure that would not hold itself together with a decision.

The day before

Perhaps everything one needs to know about Vajont is in a single sheet of paper: the notice issued by the municipality of Erto on 8 October 1963 — the day before — at the prompting of SADE's own technicians.

It informed the population that the company's engineers were warning of slope instability on Monte Toc, and that it was therefore prudent to leave the area around the lake.

Because slides from Toc could raise terrible waves, people, especially in Casso, were advised to withdraw without delay, with their animals and belongings.

Read it again.

The system acknowledged the danger.

It put it in writing.

With the municipal stamp, at the company's prompting, in official language.

It knew of the lethal waves and named them — the day before.

And out of all that knowledge it managed to produce exactly one more thing: a sheet of paper, fixed to a door, advising people to be prudent.

It could not translate the knowledge into the single act that would have meant anything — stop the project, empty the lake, admit it should never have been built.

Between what the system knew and what it did, an abyss opened. In that abyss, two thousand people died.

One more fragment from the final hours holds the nature of that abyss. On the afternoon of 9 October, hours before the wave, the government supervisor at the dam drew up a report stressing the seriousness of the situation, for which instructions are awaited.

The report then set off for Rome — by ordinary post.

A document about a danger that would kill two thousand people that very night was handled as routine administrative correspondence.

And when the technicians at the dam watched the slope move that evening, the trees tilting like toppled bowling pins, and asked the works manager what to do, he could only whisper: may God help us.

A man who had trusted the powers of engineering all his life ended with a prayer — because the system that should have decided was still, to the last, waiting for instructions that never came.

On Toc itself, only two people remained that night: a married couple, Fulvio and Carmela, who would not leave their house and their animals just above the dam wall.

All the other families had withdrawn by evening.

The wave took the couple; their bodies were never found.

Between the official warning and any action lay the difference of two lives that trusted the system exactly as much as it had let them — which was not enough to save them.

Four sites of a single slide

Put the stories together and a pattern emerges.

A warning, silenced legally.

The truth, softened hierarchically.

Science, perverted methodologically.

An individual, broken ethically.

And around them all, a fifth and quietest mechanism — narrative, which after the fact converts a decision into a natural disaster and absolves everyone who decided.

Every level functioned “normally.”

The court did its work.

The geologist filed his report.

The laboratory ran its experiment.

The director followed instructions.

No one needed a grand conspiracy. Each fragment of the system made its own small adjustment — and out of a legally confirmed lethal danger, official optimism was produced.

It is important that Vajont was not the deed of a single man.

The catastrophe fell exactly in the transitional period when Italy was nationalising its electricity sector.

The state-owned ENEL was created in December 1962; the Vajont facility passed to public hands in the spring of 1963, only months before the slide, while SADE continued to operate it.

Responsibility was dispersed among private interest, state modernisation and bureaucratic handover at the precise moment it should have been most concentrated.

In such a web, no one wanted to be the one to stop the project — and so no one stopped it.

This is quiet deviation in its purest form: a system collapses not from a single blind eye but because it has so many devices for soothing uncomfortable data that the data never reaches the point where it would have to become a different decision.

Only the judicial epilogue named what the system could not name in time.

In 1971 — days before the statute of limitations would have expired — the Court of Cassation confirmed in a binding ruling that the event was no natural disaster but a foreseeable catastrophe.

The myth of the will of God fell.

Vajont was not fate.

It was a decision.

But even justice arrived diluted: the trial had been moved far from the valley to L'Aquila “for serious reasons of public order,” and even before the first verdict, ENEL had offered survivors a settlement to renounce their claims.

Responsibility, like everything else in this story, could be processed into an acceptable form — paid out before it was ever proven.

A necessary objection: the story we tell about the story

So far I have told the familiar story: a profit-driven company, silenced warnings, a courageous journalist. That is how most of us know it — through Tina Merlin's book and Paolini's monologue. But if I doubt SADE, I have to turn the same doubt on this story too.

In a 2026 study, the organisational sociologist Maurizio Catino distinguishes three kinds of truth that bear on a disaster:

• judicial truth (who is to blame, established under the rules of a trial),

• historical truth (how and why it happened, established by the methods of history),

• and a third he calls narrative truth — a version that circulates because it is emotionally compelling and widely shared, amplified by media as echo chambers, accepted as true largely on the strength of how well it is told.

Narrative truth is not necessarily false, but it can quietly displace better-evidenced explanations. Catino argues that this is exactly what happened to Vajont.

His sharpest target is the so-called race to testing — that SADE rushed to test the dam in order to sell it to ENEL at a higher price. He refutes it with numbers. The money was not tied to testing; the compensation was based on the average share value over 1959–61; and by the night of the disaster the dam had been ENEL's property for seven months. There was neither a reason nor a mechanism for any "race."

I take this seriously. But it exonerates no one: the verdict stands — a foreseeable catastrophe caused by negligence, not fate. What Catino removes is only the motive of greed.

The engineers faced a genuine scientific enigma — the geologists disagreed, the knowledge was not enough, and no one foresaw the speed of the slide.

That is not a story about greed.

It is a story about a system that could not convert divided, uncertain, contested knowledge into a precautionary decision — which is the failure of coherence this entire essay is about.

You do not need anyone to be evil for the gap between knowing and doing to kill two thousand people.

Only a system that, faced with uncertainty, reaches for the more comfortable reading and lets the lake keep filling.

So I keep Merlin in this essay — but for what is documented, not for what grew up around her: she reported real fear, she was prosecuted, she was acquitted, and a court found the danger well-founded three years before the slide.

Vajont is so complete a failure of coherence that even our memory of it has struggled to turn evidence into an accurate story. The disaster of the diluted datum did not end in 1963 — it continues, quietly, every time the better-told story beats the better-evidenced one.

That is the most uncomfortable form of the lesson.

It is easy to see incoherence in someone else's organisation fifty years ago. It is harder to notice it in the story we ourselves hold dearest.

Vajont today

Visiting Vajont is still unsettling.

The catastrophe is not stored only in a museum and cut into a memorial plaque.

Like a raven, it spreads its wings between the dam, the mountain and the valley, as far as the eye can see from Giovanni Michelucci's church in Longarone.

The visitor climbs through it in two spirals, as if ascending out of Dante's circles, past three bells tuned to a mournful chord that resound in the bare concrete of the inner and outer amphitheatre.

Raw, heavy, painful, the space cries out for a view into the distance — toward the valley the destruction came from — and there, a small grey triangle far off, you catch sight of the dam. It stands eerily silent, perfect, untouched.

The old church had five bells; one survives whole, marked 1922. It rings once a year, on 9 October at 22:39, the hour of the catastrophe. In the lower church a few objects pulled from the mud are kept: a wrought-iron cross, a bell that lay among charred beams. The objects survived; the people did not — the exact reverse of what one would expect. In the museum, a survivor recalls searching for the place his house had stood and finding nothing left to anchor himself to, not one point by which to locate it. Only a few found an object that still tied them to a former life — a doll, an alarm clock, a sheet of paper, a spoon.

A tourist path now runs along the crest of the dam. People walk it, photograph the valley, lean over the rail — tens of thousands a year.

Part of this is necessary memory.

Part of it is a catastrophe turned into a scenic viewpoint.

The problem is not that people come.

The problem is how quickly a place where a person ought to fall silent becomes an attraction.

On the sixtieth anniversary, in October 2023, the President of Italy walked the cemetery at Fortogna alone — no speech, no entourage — among the white marble stones, one for each of the victims.

Beside that cemetery stands an inventory that no analysis of coherence, no graph, no report can answer:

1,910 victims;

773 families struck;

1,464 bodies buried at Fortogna, of whom 703 were identified and 761 were not;

many never found at all.

486 children between birth and fifteen.

33 not yet a year old.

28 born between the first of January and the ninth of October 1963.

Twenty-one — the number of days the youngest had lived.

And the children who were never born.

And yet perhaps this is exactly where the final lesson sits.

The dam still stands, because the concrete was not mistaken.

Those who were mistaken were the people deciding around the concrete — and their error is invisible, because an error of coherence never shows as a crack in the wall.

It shows only as the valley below.

So the question Vajont leaves is not whether we have enough data. We have more than ever. The question is sharper, and it is addressed to anyone who leads a project, an organisation, or simply faces a hard decision of their own: which piece of data have we already received — and is our system, even now, busy processing it into something more comfortable?

There is a question I have carried since I stood in Longarone, and I have since found it almost word for word in the work of the artists who keep returning to this valley: how do you see a catastrophe that is still only approaching?

The honest answer is that seeing was never the hard part.

At Vajont almost everyone, at some level, saw.

The hard part — the part we are still failing — is letting what we see change what we decide.

In the museum a child's poem is kept, titled My Daddy. The father worked far away and promised to come home for Christmas. He did come, a little early — but the child could not wait for him, because before Christmas the child, the mother and the grandmother were already gone. It was not our fault, Daddy. We will always love you.

No coherence model measures up to that. It is only meant to keep the sentence from ever having to be written again.

A note on the figures

Where the sources diverge — the volume of the slide (most estimates cluster near 270 million cubic metres), the speed of the collapse, the height of the wave, the exact toll — this essay uses cautious wording (around, more than, well under) to avoid false precision. The geological interpretation follows the most recent reassessments rather than the long-standing consensus: in particular the work of Paronuzzi and colleagues (2012–2021), Wolter and colleagues (2016), and Dykes and Bromhead (2018), together with the sixty-year research synthesis of Lapa and Bögöly (2025), which together have substantially revised the older account of the slide mechanism. The reading of the laboratory model draws on the 3D numerical reconstruction by Franci and colleagues (2020). The critical discussion of the dominant narrative draws on Catino (2026).

Sources and further reading

Geology, mechanics and the laboratory model

• Dykes, A. P. & Bromhead, E. N. (2018). The Vaiont landslide: re-assessment of the evidence leads to rejection of the consensus. Landslides, 15, 1815–1832.

• Dykes, A. P. & Bromhead, E. N. (2018). New, simplified and improved interpretation of the Vaiont landslide mechanics. Landslides, 15, 2001–2015.

• Franci, A., Cremonesi, M., Perego, U., Oñate, E. & Crosta, G. (2020). 3D simulation of Vajont disaster. Part 2: Multi-failure scenarios. Engineering Geology, 279, 105856. (With Part 1: Numerical formulation and validation, 105854.)

• Genevois, R. & Ghirotti, M. (2005). The 1963 Vaiont Landslide. Giornale di Geologia Applicata, 1, 41–52.

• Hendron, A. J. & Patton, F. D. (1985). The Vaiont Slide: A Geotechnical Analysis Based on New Geologic Observations of the Failure Surface. Technical Report GL-85-5. Vicksburg, MS: U.S. Army Engineer Waterways Experiment Station.

• Lapa, R. O. & Bögöly, G. (2025). The Vajont Landslide: an overview of 60 years of research. Rock Mechanics Letters, 2(3), 21.

• Müller, L. (1964). The Rock Slide in the Vaiont Valley. Rock Mechanics and Engineering Geology, 2, 148–212.

• Paronuzzi, P. & Bolla, A. (2012). The prehistoric Vajont rockslide: an updated geological model. Geomorphology, 169–170, 165–191.

• Paronuzzi, P., Rigo, E. & Bolla, A. (2013). Influence of filling–drawdown cycles of the Vajont reservoir on Mt. Toc slope stability. Geomorphology, 191, 75–93.

• Paronuzzi, P., Bolla, A., Pinto, D., Lenaz, D. & Soccal, M. (2021). The clays involved in the 1963 Vajont landslide: genesis and geomechanical implications. Engineering Geology, 294, 106376.

• Semenza, E. & Ghirotti, M. (2000). History of the 1963 Vaiont Slide: the importance of geological factors. Bulletin of Engineering Geology and the Environment, 59, 87–97.

• Wolter, A., Stead, D., Ward, B. C., Clague, J. J. & Ghirotti, M. (2016). Engineering geomorphological characterisation of the Vajont Slide, and a new interpretation of the chronology and evolution of the landslide. Landslides, 13, 1067–1081.

History, memory and the sociology of the narrative

• Catino, M. (2026). Forms of Knowledge and Regimes of Truth: The Case of the Vajont Disaster. Sociologica, 20(1), 61–78.

• McCully, P. (1996). Silenced Rivers: The Ecology and Politics of Large Dams. London: Zed Books.

• Merlin, T. (1983). Sulla pelle viva. Come si costruisce una catastrofe. Milan: La Pietra. (Later editions: Cierre, 1997.)

• Paolini, M. & Vacis, G. (2014). Il racconto del Vajont. Milan: Garzanti.

• Reberschak, M. (ed.) (2013). Il grande Vajont (3rd, revised ed.). Sommacampagna: Cierre edizioni.

• Art, place and the question of the approaching catastrophe

• CALAMITA/À (2024). An investigation into the Vajont catastrophe. Amsterdam: Fw:Books. (Curated by G. Arena & M. Caneve; see also ArchAlp n. 15, 2025.)

Documents, museum and contemporary accounts

• Comune di Erto (1963). Avviso di pericolo continuato, 8 October 1963.

• Buzzati, D. (1963). Natura crudele. Corriere della Sera, 11 October 1963.

• Merlin, T. (1959). La SADE spadroneggia ma i montanari si difendono. L'Unità, 5 May 1959.

• Statistical inventory of the victims: inscription beside the Fortogna cemetery (renovated 2004). Child's poem Il mio papà: museum exhibit, Longarone.