The Mandela Effect Is What Happens When Strangers Share the Same Wrong Memory

Pikachu’s tail was never tipped in black. The Monopoly man never wore a monocle. Yet people around the world are certain otherwise. What the Mandela Effect reveals about the architecture of memory.

Photo by Gregory Fullard on Unsplash

This article was originally published in Japanese on Ki to Oku. (Translated via Generative AI)

Where the Name Comes From

In 2009, American paranormal researcher Fiona Broome was at a conference, discussing with several other attendees a shared memory: that Nelson Mandela had died in prison sometime in the 1980s. Broome herself recalled watching news coverage of his death, even a memory of funeral footage.

Mandela, of course, was alive. Released in 1990, he went on to serve as South Africa’s first Black president from 1994 to 1999, and died of a respiratory infection in 2013.

Broome wrote about the experience on a website and asked whether others shared the memory. Strangers with no connection to one another responded with the same recollection. She called the phenomenon the Mandela Effect and began applying the term to similar cases. It now describes situations where unrelated people misremember the same false detail as fact, with the same confidence. In cognitive science, the formal term is collective false memory.

What People Get Wrong, and How

The examples are numerous. Four of the most widely shared are worth examining closely.

Pikachu’s tail. Many people who grew up with the character recall a black stripe at the tip of its tail. In the actual design, the tail is solid yellow with a slightly brown-tinted base — no black stripe exists.

The Monopoly mascot, Rich Uncle Pennybags. Many remember him wearing a monocle along with his top hat and mustache. No official version of the character has ever included one.

Darth Vader’s line in The Empire Strikes Back. “Luke, I am your father” is often cited as one of the most famous quotes in film history. The actual line omits the character’s name entirely.

A Japanese example: the opening sequence of the anime Kyojin no Hoshi (Star of the Giants). Many viewers recall a scene of the protagonist, Hyuma Hoshi, pulling a heavy ground roller during the opening credits. That scene does not appear in the opening. It appears once, mid-episode, in the body of the show itself, accompanied by the theme song — and the two memories appear to have merged into one.

Cases like the Kyojin no Hoshi example are culturally specific; the Mandela Effect includes plenty of local variants tied to a particular culture’s media.

What unites all four cases is that the error never strays far from plausible. A gentlemanly cartoon figure seems like the kind of character who would wear a monocle. If Pikachu’s ears are tipped in black, a black-tipped tail feels like it should follow. This sense of plausibility appears to be doing some of the work in reshaping the memory.

When Science Finally Took Notice

For years, these cases circulated mainly as internet memes rather than objects of serious psychological study. Part of the reason was that Broome, the person who coined the term, was a paranormal researcher who promoted her own explanation for the phenomenon on her site: that we are constantly crossing between parallel universes, and that our memories get rewritten with each crossing. What feels like a “false memory,” in this view, might be a residue of something that really happened in another timeline. The term itself was, from the start, tangled up with this occult framing.

The turning point came in 2022, when Wilma Bainbridge and Deepasri Prasad of the University of Chicago published a study in the journal Psychological Science — the first systematic experimental investigation of what they termed the “visual Mandela Effect.”

Across 40 well-known visual icons, participants asked to identify the correct version among several options got the answer right less than a third of the time for seven of them. And the errors weren’t random: participants converged on nearly the same wrong version, and did so with confidence and a strong sense of familiarity.

An even more striking finding followed. Researchers showed participants who had never seen Pikachu or the Monopoly man before the correct image, then tested their memory of it a short time later. These first-time viewers reproduced the same false details anyway — a black tail stripe, a monocle — even though they had just seen the accurate version moments earlier. This result rules out the parallel-universe explanation outright, and points to some mechanism in memory formation that operates independently of prior exposure. The finding has since been replicated by an independent research team.

The researchers also examined the images actually circulating online. In a scrape of Google Image search results, they found that incorrect versions of some icons — Star Wars’ C-3PO with an entirely gold body, missing his one silver leg — made up roughly half of all images returned. But for Pikachu’s tail and the Monopoly man’s monocle, images depicting the false feature were nearly absent: under 5% for Pikachu, and close to zero for the Monopoly man. In other words, the simple explanation — that repeated exposure to wrong images online corrupted people’s memories — doesn’t hold up, at least not for these two cases.

From this, the researchers suggest that the false versions themselves may carry some property that makes them more memorable than the accurate ones. A monocled Monopoly man may simply stand out more, as a piece of visual information, than the monocle-free original — and the brain may be more inclined to retain whatever stands out.

A Structural Weakness in the Brain

Why do so many people land on the exact same error? Researchers point to several overlapping mechanisms.

One is schema-driven error. A schema is a knowledge structure that shapes how memory works — helpful for comprehension, but also a source of distortion. Psychologist Frederic Bartlett demonstrated this experimentally in his 1932 book Remembering: A Study in Experimental and Social Psychology, showing that when people encounter unfamiliar information, they tend to convert it into a more familiar shape, dropping or altering details along the way. The monocled Monopoly man may owe something to Mr. Peanut, a contemporary gentleman-mascot whose defining feature actually is a monocle.

A second mechanism is source-monitoring failure — forgetting where a piece of information came from, and mistaking something heard secondhand, or seen in a different context, for direct personal experience. The Kyojin no Hoshi case fits this pattern well: a scene from the body of the show and the memory of its theme song appear to have fused into a single memory of “seeing it in the opening.”

A third is the social contagion of memory. False memories spread from person to person. Elizabeth Loftus’s well-known misinformation effect describes how memories can be altered by information received after the fact — a process that, on social media, now plays out constantly and at scale.

None of these mechanisms fully accounts for the Mandela Effect on its own. Bainbridge herself has said the cause is likely not singular, and that more research is needed. The fact that first-time viewers reproduce the same error immediately after seeing the correct image suggests something beyond schema effects or social contagion is at work.

Ken Drinkwater of Manchester Metropolitan University has suggested that periods of uncertainty — a pandemic, for instance — tend to make misinformation and conspiracy theories spread more easily. People are inclined to search for meaning in events, and belief in the Mandela Effect may draw some of its pull from that same impulse.

Memory does not play back the past like a video recording; it reconstructs it, each time, from scratch. Cognitive science has spent decades establishing this. What the Mandela Effect adds to that picture is that reconstruction doesn’t always stay private — it can take the same shape across total strangers. What it doesn’t yet explain is why only certain memories behave this way, and most don’t.