The Problem Of An Incomplete Narrative
People do not experience closing the gaps as bias. They experience it as reasoning.
On the night of 23 May 2010, thirteen and a half million people watched the final episode of Lost. By the following morning, Reddit had a certain level of anger even I found hard to quantify. Fan forums that had spent six years cataloguing every symbol, every number, every unexplained event were not celebrating resolution; they were pissed off.
The problem with the ending was how incomplete it left the audience. The mythology of the island was never explained. The numbers were never fully decoded. The causality of why these specific people, on this specific island, at this specific moment in history was suggested and then abandoned. The writers, Damon Lindelof and Carlton Cuse, later admitted they had been raising questions for six seasons without always knowing the answers.
Now I know this is a disinformation Substack, but stick with me here.
For Lost, the audience did not wait. Within days, one explanation had spread across every platform: they were dead the whole time. But Lindelof denied it explicitly in 2010, again in 2014, and again in a 2021 interview with Vulture, where he said the theory “negates the whole point of the show.”
For millions of people, that became the real ending, authored not by the writers but by the audience’s own schema completion, the brain’s documented tendency to resolve narrative gaps using whatever fits the existing frame rather than waiting for evidence.
Psychologists call the underlying process motivated reasoning. Ziva Kunda’s landmark 1990 research established that people do not reach for the nearest available explanation when a gap opens. Instead, they reach for the explanation they are already disposed to believe, then construct what feels like a rational justification for it. The completion feels like understanding. It is not experienced as a guess; it is experienced as recognition.
Now consider what the same process produces when the incomplete story is not a television finale but the opening days of war. In June 2025, Israel launched strikes on Iranian nuclear sites. The wound was agreed upon. The causality of who bore ultimate responsibility and why, was not. Within hours, both sides were distributing their own completions.
The mechanism is the same.
The stakes are not.
An Incomplete Table
In Lost, the characters were fully developed across six seasons. In the Israel-Iran conflict, they were cast simultaneously by both sides into opposite roles. Israel framed its June 2025 strikes as a necessary response to an existential nuclear threat. Iran framed the same strikes as foreign aggression against a sovereign state.
The same event but two complete and internally coherent narratives that left no gap to notice. Therefore, each audience received a finished story. The damage was not that people felt uncertain; it was that they felt certain, in opposite directions, about the same event.
Initial equilibrium is where the competing stories become structurally incompatible. For Israel, the conflict’s narrative began with decades of Iranian proxy warfare and nuclear development. For Iran, it started with the June 2025 strikes themselves. These two starting points are not just different perspectives; they produce different wounds, different villains, and different justified responses from identical raw material. Whoever establishes where the story begins controls what counts as an attack and what counts as a reaction.
Causality is the row that converts events into verdicts.
In Lost, it was deliberately withheld. The island’s purpose, the specific reason these people were chosen, and the true nature of the forces at work were hinted at across six seasons yet never delivered. In the Israel-Iran conflict, the Trump administration’s causality shifted between nuclear threat, imminent attack, missile capability, regime change, and personal retribution across separate statements in the same week, as documented by the Washington Post, NPR, and NBC News. Both cases produced the same outcome as the audience fills(ed) the gaps themselves.
Kunda’s 1990 research on motivated reasoning identified the specific mechanism. When people encounter a gap in a causal chain, they do not reach for the nearest available explanation. Rather, they reach for the one consistent with the conclusion they already want to reach, then construct what feels like a justification. Critically, the person doing this does not experience bias; they experience reasoning. The Lost viewer who concluded the characters were dead the whole time was not being careless. They were being thorough, with a frame that was already loaded before the finale aired.
The moral axis row shows this in its starkest form. In Lost, the writers deliberately destabilised the axis. Jacob was presented as the protector of the island and also the man who authorised the Dharma Purge, the killing of an entire research community. The Man in Black was the villain and also the only character whose imprisonment was portrayed as a genuine injustice. The audience did not sit with the ambiguity; they resolved it by choosing a side, then sorted every subsequent piece of information through that choice. In the Iran conflict, the axis was not ambiguous. It was fully formed and inverted between audiences.
Resolution is the row with the sharpest downstream consequence. In Lost, it was emotionally present but mythologically absent. The characters found peace but the numbers and the island’s true purpose did not. In the Iran conflict, the Twelve Day War ended with a US-brokered ceasefire on 24 June 2025, with no settled causality and no accepted verdict. Operation Epic Fury, the joint US-Israeli offensive launched on 28 February 2026, confirmed that the June ceasefire had resolved nothing.
An absent resolution does not produce acceptance. It leaves a audience/population still waiting for the story to close, and in that state, the first fragment that arrives with a structurally complete answer does not need to be accurate; it just needs to fit. The IRGC’s false claim on 1 March 2026 that Iranian missiles had sunk the USS Abraham Lincoln reached eight million views before CENTCOM denied it. It spread that quickly not because it was convincing on its merits, but because it offered a form of resolution to an audience that had been holding an open story for eight months.
The reaction and moral debate reflect what all the previous gaps produced. In Lost, a false theory about a television show, denied by its creators, became embedded in popular understanding for over a decade. In the Iran conflict, competing populations carry different verdicts about who bears responsibility for the war, each verdict built from the same events, each feeling like the only reasonable conclusion.
The gaps in this table are not equivalent. A wrong theory about Lost costs nothing beyond a strong opinion. The same mechanism operating on the causality and resolution rows of an active military conflict shapes the terms on which people accept or reject what happens next.
The Brain Behind This
The brain does not experience an incomplete story as incomplete. In fact, the brain actively hates incomplete narratives.
Frederic C. Bartlett established why in 1932. In his Cambridge experiments on reconstructive memory, he gave British students a Native American folk tale called The War of the Ghosts and asked them to recall it over increasing time intervals. The story was culturally unfamiliar so what the students remembered was not the story as given. Rather, it was a version quietly edited to fit their existing cultural expectations. Unfamiliar details were replaced with familiar ones and gaps were filled without the students knowing they had filled them. They believed they were recalling accurately, but they were not; they were reconstructing, using schemas already in place to complete what the original material had not provided.
That is reconstructive memory. It explains how gaps are filled in the act of recall.
Motivated reasoning explains which direction the filling tends to go.
Ziva Kunda’s 1990 paper in Psychological Bulletin identified this specific mechanism. When people reason towards a desired conclusion, they do not simply accept what they want to believe; they construct what feels like a justification. They search for supporting evidence, weigh it favourably, and arrive at a position that feels earned. The bias is not in the conclusion itself; it is in the selection of what counts as evidence. Kunda’s finding was that people ask “Can I believe this?” for conclusions they are drawn towards, and “Must I believe this?” for conclusions they are not. The reasoning process feels identical in both cases, but the outcomes are not.
Applied to a missing narrative row, the implication is direct. When the causality row is absent, the search is not neutral. It is directed by the frame already in place. The actor already cast as the villain in the existing moral axis is the most likely recipient of the missing cause. The absence of the resolution row does not produce patience. Instead, it produces an active search for an ending that fits.
The second reason the wrong direction dominates is experiential. A correct gap-fill and an incorrect one are indistinguishable from the inside as both produce the same subjective sense of clarity. The Lost viewer who concluded the characters were dead the whole time was not guessing. They had constructed an explanation that accounted for everything they had seen across six seasons. That certainty did not shift because that kind of certainty is not anchored to the source of the information (in this case being the writers of the show). It is anchored to the auidence’s feeling of internal coherence.
This is why corrections so often fail. Walter and Tukachinsky’s 2020 meta-analysis of 32 studies covering 6,527 participants found that correction does not, on average, eliminate the effect of misinformation. It reduces it under specific conditions: when the correction is coherent, when it offers an alternative explanation, and when it comes from the same source as the original claim. When those conditions are not met, the original completion holds.
The gap does not stay empty.
The only question worth asking is what determines what fills it.
A Demonstration
Lost ran for six seasons with 121 episodes. That is roughly 90 hours of television built on a structural promise that the mysteries had answers. Every season added questions, and every season implied the causality row would eventually be filled.
It was not.
The finale, broadcast on 23 May 2010, provided emotional resolution for the characters while leaving the mythology largely unexplained. Showrunner Damon Lindelof later admitted that the writers had raised questions without always knowing the answers. The causality of the island, the resolution of its mythology, and the meaning behind six years of carefully planted details were absent when the story concluded.
The completion that spread fastest was the simplest: they were dead the whole time. One sentence that filled every missing row at once. It explained the island’s significance, the character selection, and the disconnected flash-sideways timeline. To the auidence, it was structurally complete.
Lindelof denied this story idea in 2010, in 2014, and in a 2021 Vulture interview where he stated the theory negated the whole point of the show. Carlton Cuse explained that the footage used to support it was intended as a narrative pause, not a revelation, and that the misperception became entrenched in the minds of many fans despite their explicit corrections. For a significant portion of the audience, it had become the real ending.
The consequence was a flawed theory about a television show.
Now, lets run the same table through the Iran conflict.
In the days before Operation Epic Fury launched on 28 February 2026, the Trump administration moved its justification across multiple positions. Trump’s State of the Union on 24 February cited Iran developing missiles capable of striking the US mainland. Days earlier, envoy Steve Witkoff told Fox News that Iran was probably a week away from having industrial-grade bomb-making material. White House press secretary Karoline Leavitt simultaneously maintained that the June 2025 strikes had already obliterated Iran’s nuclear facilities. The Washington Post, NPR, and NBC News all documented the shifting rationales. The IAEA stated it had no evidence of a structured Iranian nuclear weapons programme at the time of the strikes. A May 2025 Defence Intelligence Agency assessment concluded Iran was years away from a long-range missile capable of reaching the US.
All of these positions cannot all be true simultaneously, therefore the audience/population was not left uncertain. I asses that they were left selecting. A Pew Research Center survey of 3,524 US adults conducted in late March 2026 found that 88% of Democrats said the US made the wrong decision in striking Iran, while 71% of Republicans said it was correct. The same strikes, the same shifting rationales, evaluated through opposite frames. A Quinnipiac poll found that 79% of Republicans believed the war would make the world safer, while 74% of Democrats believed it would make the world less safe. Nobody was sitting with an open question. The same incomplete causality row had been filled in, with confidence, in opposite directions. Each audience selected the rationale consistent with the conclusion they were already disposed to reach, then experienced that selection as reasoning.
A tv series and a real conflict are not equivalent. One is fiction; the other produced military strikes, thousands of deaths, and a regional conflict still unfolding at the time of writing. However, the gap between the wound and the causality context operated similarly in both cases. The audience did not experience incompleteness. What they filled it with was determined by the frame they were already carrying.
And the truth doesn’t matter to them.
Why The Wrong Direction Dominates
The rows most frequently left incomplete are not the hardest to fill. I believe they are the most consequential ones to control.
Setting is almost always established: a location, a date, a named event. It costs nothing and concedes nothing. Causality and resolution are different. These are the rows that convert events into verdicts and verdicts into action. The first completion of either row is the hardest to shift.
The USS Maine exploded in Havana harbour on 15 February 1898. Within two days, William Randolph Hearst’s New York Journal ran the headline: “DESTRUCTION OF THE WAR SHIP MAINE WAS THE WORK OF AN ENEMY.” No investigation had concluded this; no evidence had established it. Hearst’s paper had spent months amplifying reports of Spanish atrocities in Cuba to drive circulation, as documented by the Library of Congress’s own account of the period. The public received the causality assignment before any alternative account existed. A US Navy investigation conducted in 1974 concluded the explosion most likely originated from an internal coal bunker fire. That finding arrived 76 years after the war it could not prevent.
The Reichstag burned on 27 February 1933. Marinus van der Lubbe, a Dutch council communist, was arrested at the scene with firelighters in his possession. Hitler assigned blame within hours, framing the fire as the opening signal of a communist uprising. The Reichstag Fire Decree, suspending civil liberties, was signed the following morning. At the Leipzig Trial that followed, the Reich Supreme Court convicted van der Lubbe but acquitted the four other defendants, finding insufficient evidence of a broader communist conspiracy. The legal foundation of the Nazi dictatorship had already been established before that verdict was delivered.
In both cases, a wound was real, a cause was assigned before evidence existed to support it, and this assignment had its effect before any competing account reached the public.
Walter and Tukachinsky’s 2020 meta-analysis of 32 studies covering 6,527 participants found that corrections tend to be less effective when misinformation is repeated multiple times before correction, when time elapses between the original claim and the correction, and when the correction comes from a different source than the original. All three conditions are structurally common in fast-moving conflicts, which is why the original assignment often holds even after it has been directly challenged.
The Trump administration’s shifting rationales before Operation Epic Fury, launched on 28 February 2026 and documented by the Washington Post, NPR, and NBC News, shifted between nuclear threat, imminent attack, missile capability, regime change, and personal retribution throughout the same week. No single rationale remained stable long enough to be cornered and disproved. Different audiences selected the rationale that fit their existing frame and reached conclusions that felt like reasoning.
The absence of resolution creates a related vulnerability. An unresolved story keeps the audience searching, and in that state, a fragment offering closure spreads based on structural fit rather than accuracy.
When the gap is the most valuable aspect of the story, leaving it open is not an oversight.