Mind in Motion

by Barbara Tversky

  PLACES: SCENES FOR ACTION AND EVENTS

  The first question that people seem to ask on the phone these days is: Where are you? Knowing where someone is tells us so much about their present state. And so many times when we arrive at a distant place, we call a friend who was last with us there. Places are different yet again, not like objects, not like faces, and not like bodies. As for objects, faces, and bodies, there are areas in the brain that are selective for scenes, the parahippocampal place area (PPA) and the retrosplenial complex (RSC) primary among them.

  Scenes surround us, usually richly populated. We are always embedded in a scene. Scenes are the settings for activities, and most of our activities take place in specific scenes. Scenes both constrain and enable activities, in part through the objects that are present in scenes. Scenes include many of them, and typically include faces and bodies. Scenes are the settings, the background, for all the events of our lives, for the stories of our lives, the mundane and the monumental. They are the settings for historical events, mundane or monumental. So, scenes are especially informative. What we see and do in a restaurant is very different from what we see and do in a post office or what we see and do in a classroom or what we see and do in our living rooms. Like objects, scenes have a basic level, the level of school, store, home, park, beach, forest, where parts, objects, and activities converge. Some are indoors, some outdoors, and indoors and outdoors are the superordinate categories that encompass the basic-level categories.

  Like our memory for faces, our memory for scenes is extraordinary, so extraordinary that for a number of years, there was a competition in the literature, how many pictures of scenes can people remember? The prize went to ten thousand pictures. In general, recognition of pictures was excellent, even better for vivid pictures, and in both cases, far better than recognition of words. Now a caveat. Memory was tested in two ways. In one design, participants saw pairs of pictures, one that had been viewed and one that was new. The task was to select the old familiar picture in each pair. In the other design, pictures were presented one by one. Half were old and half were new, mixed randomly. Participants judged each picture in turn as old or new. Typically, the new pictures of scenes were very different from the old ones, making the task of distinguishing the old and the new fairly easy. When pictures are more similar, recognition accuracy naturally goes down. Think of identical twins.

  Change blindness

  Now, what might seem to be a paradox. We’re so biased to pick up the gist of scenes that we ignore the details. We don’t notice when features of a scene change, even important ones, a phenomenon known as change blindness. There are many demonstrations online; try them. Several demonstrations show pairs of photographs of richly inhabited scenes, in a city, on a lake, at an airport, in rapid alternation. The paired photos are identical except for one important feature that has changed in one of them, for example, one of the engines of a large jet is missing or the battery of a motor boat has disappeared. Viewers often sense that something has changed but rarely know what it is even after seeing many repetitions of the pairs of pictures.

  Change blindness happens in real life in real time. In one impressive demonstration, passersby on a college campus were stopped by a twenty-something male student asking for directions. As the passerby explains the way by gesturing on a map, a crew carrying a large door passes between them. Using a magician’s trick, the inquiring student switches places with another student, another twenty-something male who had been part of the crew carrying the door. Fewer than half the passersby noticed the change. Yes, both were young men, of similar build and dress, and yes, the passerby was looking at the map, not at the student’s face. But still. This trick, of changing a central feature of a live situation without the participant noticing, has been demonstrated numerous times. When we look around us, we see a richly detailed world. But that’s just it, we’re seeing it, not remembering it. As long as we can see that rich world, we can reference it, and we don’t need to remember it. It’s there. We just need to know where we are.

  Change blindness is yet another example of the First Law of Cognition: There are no benefits without costs. Here the benefit is rapid ascent to meaning. If I know the meaning, I know what to expect and how to behave. The cost: the details. It’s a field in a park and there’s a game on the field, I’d better walk around it. Whether or not there are trash bins or lights for night play isn’t relevant at the moment. I’m helping yet another confused stranger find the way on campus, I’m concentrating on explaining the route on a map, I’m never going to meet that person again. It’s a shopping street I’m approaching. I’d better be prepared for crowds. Whether or not there’s a shoe repair shop or a dry cleaner isn’t important unless that’s what I’m looking for. It’s troops boarding a plane. Of course, engines are critical for the airplane, but not for me. Not now.

  CATEGORIES ARE EASIER THAN DIMENSIONS

  The world gives us more than categories; it gives us continua, dimensions on which things differ. Food isn’t just sour or sweet; there are gradations of each. People aren’t short or tall; they have heights. What we’ve seen in this chapter is how quickly the mind forms categories of the people, places, and things most central to our lives. Many of those categories are overdetermined; that is, category members have many features in common and differ from other categories by many features. Shirts share many features and differ from pants on many features. Ditto potatoes and tomatoes, chairs and carpets, football stadiums and grocery stores, beaches and post offices. The features come in bundles; they are highly correlated. Members of frequent categories such as furniture and clothing typically have many features in common, features that are not shared by other categories. Birds have feathers, lay eggs, and fly. Dogs do none of that. And I’ve picked close examples, examples from the same superordinate. Musical instruments and vegetables are far more extreme in common and distinctive features; they have far fewer overlapping features. Sharing many features and differing from others on many features allows categorization to proceed with both speed and accuracy. It also allows for many expectations and inferences that are likely to be correct. Be careful with knives. Drink from glasses. Sit on chairs. Have a bite of an apple. Kick a soccer ball. Categorical thinking makes carrying out the activities of life easy and fluent. Once we know that something belongs to one of that kind of category, we know so much about it and about what it can do and what we can do with it. Categories are incredibly useful. We use them to organize our belongings as well as our minds and our lives.

  But there are important things in our lives that can’t be so neatly put in bins and boxes. There aren’t bundles and boundaries created by common and distinctive features. Rather, exemplars that belong to different categories might differ on an important feature, the one underlying the category, but might not differ on others. Yet categorical thinking persists. Take citizenship. Citizenship is a highly meaningful category, with large implications for our lives. It also allows certain inferences about ourselves and about others. The diet, housing, and clothing of people in countries in the far north are likely to differ from the diet, housing, and clothing of people in equatorial countries. There might even be evolutionary differences due to adaptation to the different climates. People in far north Sweden, whose diet heavily relied on milk from herding animals, developed genes that allowed them to digest milk in adulthood. People in China, whose diet was crop-based, remain lactose-intolerant. Nevertheless, people in Sweden and people in China have heads and arms and legs, eat, laugh, get dressed, grow up, form families, work, and enjoy friends.

  Citizenship, then, may allow some inferences and good guesses about correlated features, but because human beings across the world are the same and different in innumerable ways, innumerable features are not correlated with citizenship. The same is true for other ways we categorize, especially the most important things in our lives, people. Children and adults. Urban and rural. Politicians and journalists. Democrats and Republicans. Cubs fans and
White Sox fans. Christians and Muslims. Poor and rich. Female and male. Categories that are as charged as they are important. People who vote or worship one way or another might share some features, but there are far more ways they differ from each other and far more ways they are similar to those in other categories. The uncorrelated features are likely to be far greater in importance than the few that are correlated. Yet thinking about subtle gradations along dimensions is much harder than lumping things into gross categories; there’s so much more that has to be considered and kept in mind. Categories are easier on the mind. But beware the First Law of Cognition, benefits come with costs.

  The beloved physician and professor of global health at the Karolinska Institute in Sweden, Hans Rosling, was dismayed by the many misconceptions people, even distinguished political and economic leaders of the world, had about the state of the world. His TED talk, telling the dramatic story of world economic development in recent times as if it were an ongoing tight soccer match, went viral. Many of the misconceptions that people held about economic and social development came from categorical thinking, especially dividing the world into rich and poor. Poor countries had no electricity, education, clean water, or health care. Rich countries had all that and more. Nothing in between. Wrong! Highly educated people vastly overestimated the percentage of the world that is poor, failing to update what was true of the world in the sixties to the world of the twenty-first century. They also saw poverty as all-or-none, not in gradations. Rosling found two ways to help people think dimensionally rather than categorically. First, by dividing the world into four categories instead of two and labeling them levels to signify a fluid dimension, not two rigid bins. Next, he used vivid depictions. Population of the world depicted as icons of people, each representing a billion, seven in total. Levels of wealth depicted in photos showing how the people at each level get water, move about in the world, cook, and what they eat. Life at Level 1 does resemble our image of dire poverty: people earn one to two dollars a day, carry water by foot in buckets from rivers, grow the crops they eat, gather firewood for cooking. But in 2017, only one billion of the seven billion people in the world lived at that level, not the large numbers that most educated people estimated. The three billion at Level 2 earn four dollars a day, get around by bicycle, have cannisters of fuel for cooking, grow more than they can eat, and eat a varied diet. Importantly, the children attend school. Even the girls. The two billion at Level 3 earn sixteen dollars a day, have running water, electricity, and a motorcycle. They might work multiple jobs, but the kids go to high school. As for Level 4, we know it intimately; we are among the one billion at Level 4. Take note that the key to jumping from level to level is means of motion, of getting from place to place in the world: from foot to bicycle to motor scooter to automobile.

  Rosling relied on two compelling devices, both rooted in spatial cognition, to correct categorical thinking and misconceptions. The first was using a small number of levels rather than two dichotomous categories. Four categories fit neatly into short-term memory, we can hold four in mind, but holding many nuanced gradations is difficult. The second was creating understanding of the levels and the numbers through striking depictions. Perhaps Rosling’s methods can encourage dimensional thinking and dispel misconceptions in other domains of our lives.

  There’s another lesson to learn from Rosling’s analysis: economic mobility is tied to mobility in space. Moving farther in space opens economic possibilities, and even more, it opens opportunities to learn more pathways in space, to encounter new perspectives, to meet new people, places, and things, to learn from interacting with them. Moving farther in space opens opportunities to increase well-being in all aspects of what makes life worthwhile.

  THE MIND CAN OVERRIDE PERCEPTION: UNCERTAINTY, HYPOTHESES, AND CONFIRMATION BIAS

  Back to perception of the world around us. The various demonstrations of change blindness show two important phenomena: that we have the impression that we are taking in a vivid, clear, coherent, and complete image of the world and that in fact we aren’t. In many cases, the immediate impression is based on inference rather than perception: if it’s a kitchen, it has a sink and a refrigerator; if a classroom, desks and white boards. Plausible inference fills in for direct knowledge in every domain of thought. Hence: the Fourth Law of Cognition: The mind can override perception.

  That is, what’s in the mind can override what’s perceived, or more generally, hypotheses override perception. First, a striking demonstration from perception of familiar objects. Years ago, students viewed blurry, out-of-focus photographs of fairly familiar objects. The students were asked to guess what they were seeing and to keep guessing as the photographs were gradually brought into focus. Another group saw the focused photographs. Of course, you’d think that the group that had previews would be faster to recognize the objects, perhaps even before they were fully focused. You’d be wrong. Those who saw the blurry photos tended to generate false hypotheses about what they were seeing, and those false hypotheses got in the way of recognizing the objects in clear view. Presumably, viewers continued to interpret what they were seeing in terms of their original hypotheses.

  This phenomenon, that hypotheses can override facts, occurs in cognition as well as in perception, like so many perceptual phenomena. The hypotheses or presumptions or beliefs that we hold bias our interpretations of the facts, of what we see. Here’s another example; there are dozens and dozens. In a classic experiment, both Princeton and Dartmouth students were asked about an especially rough football game between the rivals, a game in which there were many penalties and several serious injuries. When asked who started the rough play, 86 percent of the Princeton students but only 36 percent of the Dartmouth students said Dartmouth did. Similarly, 93 percent of the Princeton students but only 42 percent of the Dartmouth students thought the game was rough and dirty. Viewing a film of the game later, Princeton students detected more than twice as many violations by the Dartmouth team as Dartmouth students. The inevitable conclusion is that Princeton and Dartmouth students didn’t see the same game.

  Since that study, numerous other studies have shown that our own starting points, our perspectives or our hypotheses, bias our very perceptions. We are more likely to notice evidence that supports our hypotheses than evidence that refutes it. When we find counterevidence, we are likely to discount it, to explain it away as aberrations. We—everyone—are prone to confirmation bias: we actively seek confirmation of our hypotheses and ignore disconfirming evidence, and we do that even when we have no personal investment in the hypotheses. Gathering evidence that a claim is true seems to be at the very heart of understanding a claim. Off in the distance, is that the cousin we haven’t seen in years? Right height. Right hair. Right build. The surprising story we hear about a friend or news we read about a politician or a scientific discovery, could it be true? We first look for information consistent with the cousin’s appearance or the story. We can’t begin to reject a claim until we have some confidence in it. Looking for confirming evidence can make sense as a starting point; if there isn’t any confirming evidence at all, the hypothesis should be dropped. But looking for confirming evidence shouldn’t make us blind to counterevidence, fail to search for evidence that might counter the hypothesis, or dismiss it out of hand when confronted with it. Doing that can have dire consequences.

  One reviewer of dozens of experiments presenting evidence for, and some against, the confirmation bias put it this way:

  Finally, I have argued that the confirmation bias is pervasive and strong and have reviewed evidence that I believe supports this claim. The possibility will surely occur to the thoughtful reader that what I have done is itself an illustration of the confirmation bias at work. I can hardly rule the possibility out; to do so would be to deny the validity of what I am claiming to be a general rule.

  That we naturally seek support for our hypotheses and that our hypotheses make us blind to counterevidence is rampant in perception, and what i
s true of perception is true of all thought. Fifth Law of Cognition: Cognition mirrors perception.

  We’ll return to this theme over and over. Spatial thinking is mirrored in abstract thought, in social thought, in cognitive thought, in thought about what makes people tick, in thought about art and about science. Thinking is thinking, whatever the domain, and spatial thinking is core to our very existence. Getting food into our mouths, finding our ways in the world. Carrying out most of the quotidian tasks that living entails, arranging our belongings, moving about in the world. We are far from perfect spatial thinkers, we don’t perceive everything—too much is there, too much happens too fast—so we rely on plausible inferences. There is no guarantee that our inferences and judgments are veridical because we don’t have objective measurement instruments built into our bodies or our brains. So, we must rely on other mechanisms and they are imperfect and can be biased. We saw this in representations of the body and the space around the body, and there are more biases to come in representations of the larger world. Even so, we are far better and more experienced at spatial thinking than at abstract thinking. Abstract thought can be far more difficult in and of itself, but fortunately it can often be mapped onto spatial thought in one way or another. That way, spatial thinking can substitute for and scaffold abstract thought.

  PEOPLE, PLACES, THINGS

  We are surrounded by the things most central to our lives: people, places, things. They compose the settings for the events of our lives. We distinguish people we know from those we don’t know; we divine what they are thinking, feeling, doing, and communicating from their faces and their bodies. We know how to interact with the objects around us—the objects tell us. We know what objects and actions are likely in the places around us. The brain has specialized areas for recognizing people, places, and things, allowing us to identify what they are in the briefest glance with nary a movement of the eyes. Both the brain and the mind like to put things into boxes. Boxes are so much easier than dimensions. What’s surprising is how much meaning they carry, right there on the surface. Faces and bodies carry identity, emotion, intention, action, communication. Things carry affordances: what they can do for us and what we can do with them. Places tell us what they are and what we can do in them. Meaning arrives without words, too fast for words. We use that meaning to understand what is happening and to orchestrate our own behavior. And we use it to imagine worlds that have never been. Just as meanings can be wordless, so can thinking, an insight Richard Feynman came to as a child: