The Invisible Gorilla: And Other Ways Our Intuitions Deceive Us

by Christopher Chabris

  Imagine you are strolling across a college campus and up ahead of you, you see a man holding a map and looking lost. The man approaches you and asks directions to the library. You start giving him directions, and as you’re pointing to the map, a couple of people behind you abruptly say “Excuse me, coming through,” and they rudely carry a big wooden door right between you and the lost pedestrian. Once they pass, you finish giving directions. Would you notice if the original lost pedestrian were replaced by a different person as the workers carried the door through? What if the two people wore different clothes, differed in height by about three inches, had different builds, and had noticeably different voices? You would have to be pretty oblivious to miss the change. After all, you were in the middle of a conversation with the man, and you had plenty of time to look at him. That’s certainly what the two Dans and Ulric Neisser thought.

  That’s also what more than 95 percent of undergraduates thought when asked whether they would notice.23 And they were all wrong. All of us, undergraduates as well as scientists familiar with all of the research leading up to these experiments, fell prey to the illusion of memory. All were convinced that only the rare, unusually oblivious person could possibly miss the change. Yet nearly 50 percent of the people in the original experiment did not notice that they were talking to different people before and after the interruption!24

  Serendipitously, one day several years later when we were conducting a followup experiment at Harvard, many of the undergraduate psychology students were attending a lecture in the basement of the building. During the lecture, Professor Stephen Kosslyn (Chris’s graduate school mentor and longtime collaborator) happened to describe the “door” study in detail as an example of research being conducted by other faculty members in the department. When they left the lecture, several students were overheard making comments like, “There’s no way I would have missed that change.” Our recruiter asked them if they would like to be in an experiment and sent them to the eighth floor. As they stood at a counter after filling out a form, the experimenter who had been talking to them ducked down behind the counter—ostensibly to file away some papers—and a different person stood up. All of the students missed the change!25

  Change blindness is a surprisingly pervasive phenomenon considering that it has only been studied intensely since the 1990s. It occurs for simple shapes on a computer display, for photographs of scenes, and for people in the middle of a real-world interaction.26 And the illusion of memory leads people to believe that they’re great at change detection even though they’re lousy. This illusion is so powerful that even change blindness researchers regularly experience it. We only came to recognize the limits of our intuitions about memories when our own data repeatedly showed us how wrong we could be. Similarly, filmmakers learn about the illusion of memory the hard way, by seeing evidence of their own mistakes on the big screen. Trudy Ramirez, the Hollywood script supervisor, has experienced this many times: “The way you remember something, how your memory shapes what you think you saw, as sure as you think you are … often it’s different if you can actually look back at it. There were times when I would have staked my life on something and later on realized I was wrong.”

  There are limits to change blindness, of course. When we spoke publicly about the early person-change studies, we were often asked whether people would notice if a man changed into a woman. “Of course they would,” we thought, but of course our certainty was another reflection of the illusion of memory. The only way to find out was to try it. Later experiments in Dan’s lab showed that people do in fact notice when you change a man into a woman or when you change the race of an actor in a movie. And people are more likely to notice a change to the identity of a person who is a member of their own social group.27 But most other changes often go undetected.

  Even when subjects notice the person swap in these real-world experiments, they’re far from perfect in picking the original experimenter from a photographic lineup. And people who miss the change do no better with the lineup than they would have done by just guessing randomly.28 In a brief encounter, we appear to store so little information about another person that we not only fail to see changes, but we also can’t even identify the person we saw just minutes earlier. When you interact briefly with a stranger, there are only a few pieces of general information you can be certain of retaining: sex, race, and social group (student, blue-collar worker, businessperson, and so on). Most of the rest of what you perceive about the person probably won’t make it into memory at all.

  Recall Leslie Meltzer and Tyce Palmaffy, who witnessed a knife attack from their car but recalled it differently just moments later. In light of the evidence that people sometimes fail to notice that a person has almost instantaneously been replaced by someone completely different, Leslie and Tyce’s discrepant eyewitness memories are unsurprising. After all, they were just observing the person from a distance, not standing face-to-face with him and giving him directions.

  “I Sat Next to Captain Picard”

  About ten years ago at a party Dan hosted, a colleague of ours named Ken Norman told us a funny story about sitting next to the actor Patrick Stewart (best known for his roles as Captain Jean-Luc Picard of Star Trek and Charles Xavier in the X-Men films) at a Legal Sea Food restaurant in Cambridge, Massachusetts. The story was prompted when Chris noticed that Dan had a small figurine of Captain Picard perched next to his television screen. “Can I buy your Captain Picard?” asked Chris. Dan said that it was not for sale. Chris offered five, then ten dollars. Dan refused. Chris eventually raised his bid to fifty dollars—for reasons that escape him now—but Dan still refused. (Neither of us remembers why Dan refused, but to this day, Picard has not left his place amid Dan’s electronics.)

  At this point Ken told us that at Legal Sea Food, Patrick Stewart had been dining with an attractive younger woman who, based on snippets of overheard conversation, appeared to be a publicist or agent. For dessert Stewart ordered Baked Alaska—a choice that stood out in memory because it appears rarely on restaurant menus. Toward the end of his meal, another distinctive event happened: Two members of the kitchen staff came out to Stewart’s table and asked for his autograph, which he readily granted. Moments later, a manager appeared and apologized, explaining that the “Trekkie” cooks’ action was against restaurant policy. Stewart shrugged off the supposed offense, and he and his companion were soon on their way.

  The only problem with the story was that it had actually happened not to Ken, but to Chris. Ken had heard Chris tell the story some time before and had incorporated it into his own memory. In fact, Ken felt so strongly that the memory was his, and had so completely forgotten that Chris was the original raconteur, that even Chris’s presence when Ken retold the story did not jog his memory of the way in which he had actually “encountered” Captain Picard. But when Chris pointed out the error, Ken quickly realized that this memory was not his own. This anecdote illustrates another aspect of the illusion of memory: When we retrieve a memory, we can falsely believe that we are fetching a record of something that happened to us rather than someone else.

  Although we believe that our memories contain precise accounts of what we see and hear, in reality these records can be remarkably scanty. What we retrieve often is filled in based on gist, inference, and other influences; it is more like an improvised riff on a familiar melody than a digital recording of an original performance. We mistakenly believe that our memories are accurate and precise, and we cannot readily separate those aspects of our memory that accurately reflect what happened from those that were introduced later. That’s how Ken appropriated Chris’s story—he had a vivid memory for the event, but mistakenly attributed it to his own experience. In the scientific literature, this type of distortion is known as a failure of source memory. He forgot the source of his memory, but because it was so vivid, he assumed that it came from his own experience.

  Source memory failures contribute to many cases of unintentional plagiarism
. In the classes we teach, we occasionally encounter intentional plagiarism (or a gross misunderstanding of the right way to do research) when a student copies sections of a paper from Wikipedia or other sources. Unintentional plagiarism refers to cases in which people are convinced that an idea was their own when they actually learned about it from someone else. Recently, bestselling spiritual author Neale Donald Walsch was caught plagiarizing a story originally written by Candy Chand that had circulated on spirituality websites and blogs for more than a decade.29 The story describes a group of students who were using placards to spell out “Christmas Love” in a winter pageant rehearsal. One student accidentally held her letter “m” upside down, resulting in the phrase “Christ was Love.” Walsch posted the story to Beliefnet.com in December 2008 as if it had happened to his son Nicholas. But it actually happened to Chand’s son, who also is named Nicholas, twenty years earlier—before Walsch’s son was even born. In this case, it is clear that Walsch appropriated someone else’s story. The question, though, is whether he was plagiarizing intentionally or whether he merely misappropriated the memory. In acknowledging a “serious error,” Walsch stated:

  I am truly mystified and taken aback by this…. Someone must have sent it to me over the internet ten years or so ago…. Finding it utterly charming and its message indelible, I must have clipped and pasted it into my file of “stories to tell that have a message I want to share.” I have told the story verbally so many times over the years that I had it memorized … and then, somewhere along the way, internalized it as my own experience.

  This case bears all the hallmarks of a failure of source memory. Walsch remembered the story, having read and retold it many times. The fact that the child in the story had the same name as his son made it easier for him to come to believe that the memory was his. (Our friend Ken Norman probably picked up Chris’s story more readily because he had dined at the same Legal Sea Food restaurant.) Walsch kept a record of the story in his file and came to believe that he had written it. In his interview with the New York Times, Walsch said, “I am chagrined and astonished that my mind could play such a trick on me.” Chand, however, thinks the theft was intentional: “If he knew this was wrong, he should have known it was wrong before he got caught … quite frankly, I’m not buying it.” Both Chand’s indignation and Walsch’s astonishment are entirely consistent with the illusion of memory. Walsch doesn’t understand how he could have mistakenly appropriated another person’s memory, and Chand doesn’t believe that he could have done so innocently. They both think that memory must be more faithful to experience than it really is.

  Just as we cannot be certain that Kenny Conley suffered from inattentional blindness when he reported not seeing Michael Cox being beaten, we cannot say for certain whether Walsch’s plagiarism was intentional or accidental. What we can say, though, is that it is possible that Walsch internalized someone else’s memory and lost track of the source of the story. Such source memory failures are common, and they even can be created in the laboratory. In a clever study, psychologists Kimberly Wade, Maryanne Garry, Don Read, and Stephen Lindsay asked subjects to view a doctored photograph showing the subject enjoying a hot air balloon ride as a child.30 The subjects were each interviewed several times, and were asked each time to recall the event, or if they could not recall it, to imagine that it had happened to themselves. Although none of the subjects had ever taken a hot air balloon ride, the photograph and attempts to recall it led some of them to incorporate information about the image into their personal narrative memories. Half the subjects created a false memory about the balloon ride, some embellishing their memories substantially beyond what was shown in the photograph.

  The ability to change memories using doctored photographs has Orwellian ramifications. If we can induce false memories simply by editing images, it might be possible to literally revise history, changing the past by doctoring it. Using a similar approach, Dario Sacchi, Franca Agnoli, and Elizabeth Loftus showed subjects an edited version of the famous photograph of a single person standing in front of a column of tanks during the 1989 protests at Tiananmen Square in Beijing.31 In the original version of the photograph, only the lone protester was visible on the wide road. The doctored version shows crowds of people lining a narrower road on both sides of the tanks. When they were quizzed about the historical facts of Tiananmen Square only moments later, those who viewed the doctored photograph believed that far more people had been at the protest.

  Forgetting a Life-and-Death Matter

  Memory distortions are not limited to irrelevant details like whether or not books were in an office or particular words were part of a list. In fact, they can apply to life-and-death decisions, even those that you yourself have made. Australian psychologist Stefanie Sharman and her colleagues conducted an experiment that calls to mind the classic Seinfeld episode in which Kramer asks Elaine to help him and his lawyer work through a long list to decide the medical circumstances under which he would be willing to carry on living. (Lawyer: “OK. One lung, blind and you’re eating through a tube.” Kramer: “Naw, that’s not my style.” Elaine: “Borrrr-ing.”) The researchers interviewed adults and asked them to make (more realistic) decisions about which life-sustaining treatments they’d want if they were seriously ill.32 For example, would they want only CPR performed, or would they also want to be fed artificially if necessary? They interviewed the same people twelve months later using the same questions.

  Overall, 23 percent of all their decisions changed between the initial interview and the follow-up, meaning that people who said during the first interview that they would want a life-extending treatment said during the second interview that they wouldn’t want it (or vice versa). That people would change their preferences is not terribly surprising. Perhaps they had discussed the possibilities with friends, relatives, or doctors in the interim; maybe they encountered news stories about end-of-life issues. What is striking is that 75 percent of the people who changed their minds were unaware that they had done so! They thought that the decision they reported in the second interview was the same as their decision in the first interview. Their memory for what they had said earlier was rewritten to match their current beliefs.

  The illusion of memory leads us to assume—unless we receive direct evidence to the contrary—that our memories, beliefs, and actions are mutually consistent and stable over time. Amid the national grief after President Kennedy was assassinated, a poll showed that two-thirds of people claimed they had voted for him in the 50/50 squeaker election of 1960.33 At least some of them must have revised their memories of how they voted three years earlier, probably to make them consistent with the positive feelings they had about their fallen leader. More broadly, we tend to assume that everything in our world is stable and unchanging unless something draws our attention to a discrepancy. When our beliefs change, though, our memories can change along with them. A living will you produced a few years ago may not reflect your current preferences—but you are likely to misremember its contents and assume that it expresses what you want today. If you become seriously ill and are unable to communicate, doctors will rely on this document and may inadvertently take actions that contradict your wishes.

  Where Were You on 9/11?

  Try to recall exactly where you were when you first heard about the attacks of September 11, 2001. If you’re like us, you have a vivid memory of how you learned about the attacks, where you were, who else was with you, what you were doing immediately beforehand, and what you did immediately afterward. Chris recalls waking up late that morning, after the first plane had hit the World Trade Center. He listened to the Howard Stern Show on the radio until it ended around noon, at which point he turned on the TV. He got in touch with an Israeli colleague, who told him it was already obvious who the perpetrators must be, and he received an e-mail update from a friend who was living in Brooklyn, watching the events safely from her rooftop. He received another e-mail from the manager of his office building at Harvard, Wil
liam James Hall, recommending evacuation.

  Dan recalls working in his office that morning when his graduate student Stephen Mitroff came in to tell him that a plane hit the first tower. He spent the next few minutes seeking information online, and when the second plane hit, he turned on the television in his lab and he and his three graduate students watched the towers collapse. He then spent a few frantic minutes on the phone trying to reach his brother David’s girlfriend because David was traveling back from New York to Boston that morning (he was sitting on a plane waiting to take off from LaGuardia Airport when the attacks happened). Dan remembers becoming concerned that the fifteen-story building he was in might also be a target. He left before noon to pick up his wife in downtown Boston and they went home together and watched the television coverage for the rest of the day.

  Neither of us has any idea what we were doing or whom we talked to the day before 9/11. We suspect that the same is true for you. Your memories of 9/11 are more vivid, detailed, and emotional than your memories of more ordinary events from that time period. Memories of dramatic events of personal or national importance often are recalled in greater detail. Some significant events appear to be imprinted in our minds in a way that lets us play them back in video-like detail, perfectly preserved despite the passage of time. This intuition is powerful and pervasive. It is also wrong.

  Such detailed memories for a significant event were first studied systematically in 1899 by Frederick Colgrove as part of his doctoral research at Clark University. Colgrove asked 179 middle-aged and older adults where they were when they heard about the assassination of Abraham Lincoln.34 Even though he asked people to recall events that happened more than thirty years earlier, 70 percent remembered where they were and how they heard about it, and some provided exceptional amounts of detail.