Mind in Motion

by Barbara Tversky

  The new new thing keeps changing, that’s a given. But the oldest old thing keeps changing too. As new archaeological sites are discovered, the oldest known map keeps moving backward. Of course, we are unlikely to find the really and truly oldest maps because they were probably arrangements of sticks and stones or drawn in the sand or gestured in the air and are unlikely to have survived. Those that have survived are in sturdy forms, inscribed in stone or painted on the walls of caves. The current oldest, a carved stone found in a cave in northern Spain, dates back 13,600 years, far predating written language. You can see it in Chapter Eight. It shows both the terrain—mountains, rivers, paths—and animals, encouraging speculation that it was a hunting aid or perhaps a tale of hunting. What’s remarkable about this one as well as many of the previous contestants for oldest map is that they show two perspectives at once, overviews of paths, rivers, and mountains and frontal views of significant features of the environments, mountains, buildings, and in this case, animals. Many maps throughout history and to the present day do the same. They by no means conform to the standards of official cartography. Rather they appear to tell stories of the territory they depict, stories that can be entered and exited in diverse places. They can be used to tell many stories.

  MAPS IN THE BRAIN: SPACE AND BEYOND

  So, too, the maps in the brain. We go now from mind, from the thoughts and judgments and behavior of organisms, to brain, to the neural substrates presumed to underlie them. It’s almost impossible to separate mind and brain. They are connected through behavior; it’s by looking at the brain as organisms behave that we are able to infer the neural substrates. Now a pivotal example, the neural underpinnings of navigation. It will pivot us from real spaces to conceptual spaces.

  Years ago, neuroscientists learned how to place tiny electrodes into single neurons to record what excited each one. The procedure has frequently been used in rats and allows them to roam freely, oblivious to the apparatus appended to their heads. Rats live in nests and scavenge for food, so they need to keep track of where they are and how to get home from wherever they are. The key components are places, how to recognize stable points in an environment, and the spatial array of places, how you go from one place to another. Single-cell recordings from neurons in rat hippocampus and neighboring entorhinal cortex have revealed both, single cells that respond to places in the former and spatial arrays of cells that correspond to spatial arrays of places in the latter. Since the seventies, dozens of studies have shown cells in rat hippocampus that fire when rats are in particular places in a field they are exploring, some cells for one place, other cells for other places. However, the place cells are not spatially organized in the hippocampus, so how the hippocampus could serve as a cognitive map, as claimed, wasn’t clear. In the nineties, cells in rat brains that organized the place cells spatially were discovered bordering the hippocampus and richly interconnected with it, in entorhinal cortex. These cells have been called grid cells because they are laid out on a two-dimensional surface like the grid of a map and the pattern of firing looks grid-like. It is movement, the scampering of tiny feet as rats explore environments, that establishes the spatial layout of places on the grid cells. The grid cells then serve as a map, allowing the rat to go from one place to another in any order. The boundaries of the grid correspond to natural boundaries in the environment currently being explored. When rats explore a new environment, the set of grid cells is reused to map the new boundaries and locations of places. Those boundaries serve as a frame of reference for the set of places within, and both that frame of reference and the array it supports keep changing. The currently active set of grid cells is oriented with respect to the local environment, not to the coordinates of the world, as some of the hype implied. The important bottom line is that grid cells can be reused, recalibrated, and reoriented.

  Those discoveries, place cells and grid cells, earned a Nobel Prize in 2014 for John O’Keefe, May-Brit Moser, and Edvard Moser. O’Keefe and Nadel had done the seminal research on place cells as well as header cells that responded to the rat’s head direction, both in the hippocampus. The Mosers, as post-docs in O’Keefe’s lab, led the research on grid cells.

  Grid cells representing local cognitive maps are active in people when they explore environments, typically a virtual world in a scanner. In fact, the grid-cell substrate for representing cognitive maps seems quite similar in many mammals. What’s important is that the spatial representation in grid cells is place-to-place and egoless, that is, allocentric, and that allocentric representations of space are created from the get-go, even in infants.

  Those two components, place cells and grid cells, underlie spatial disorientation as well as spatial orientation. Losing either can be disorienting: not recognizing your surroundings, a feat that depends on place cells, or not understanding how your surroundings fit into the larger world, a feat that depends on grid cells.

  At the other extreme, there are the super-oriented. London taxi drivers are legendary. London is a sprawling city with paths going every which way connecting what once were dozens of villages. To become a taxi driver, candidates need to learn at least 320 basic routes through more than 25,000 streets including more than 20,000 landmarks. It typically takes two to four years of intense study to master. That intense education changes the brain! A headline-making piece of research reported that the posterior portion of the hippocampi of taxi drivers grows larger and larger the more years they drive taxis.

  Evolution likes to give new functions to old structures. Think of the mouth, originally for eating. We still use our mouths for eating, but we probably spend more time using our mouths for talking. Many of us learn to whistle and some of us can sing or play the flute. So for the brain: old structures pick up new functions both in evolution and in development. In rats, hippocampus and entorhinal cortex are used primarily for navigation, for remembering places and the paths among them. In humans, the hippocampus, entorhinal cortex, and other nearby cortical structures get used for remembering places and organizations of places, but different subregions are used for remembering and organizing many other things, including the events of our lives and the relations among ideas. Concepts and relations that are concrete and concepts and relations that are abstract alike.

  The tragic case of H.M. provided fascinating information about the roles of these structures in forming new memories. In 1953, before neuroscientists knew better, large parts of his hippocampus, entorhinal cortex, and other nearby regions of the brain were resected in an attempt to control his epilepsy. The damage was crippling; he could form no new memories, so every event of every hour of every day was a new one. People and places were unrecognized even after many encounters. He had to be cared for the rest of his life, and was.

  The hippocampus and entorhinal cortex are critical for remembering the past. It turns out that remembering the past is critical for planning the future. The same regions that are used to remember the past are also used to plan the future, so that damage in those regions affects both retrospective and prospective memory. It’s not that we need specific information from the past to plan the future; we can plan trips to places we have never been. The dual dependence on these brain structures for both remembering the past and planning the future is most likely due to the roles of these brain areas in organizing and representing the separate items of information and in integrating them in meaningful ways.

  FROM SPATIAL MAPS TO CONCEPTUAL MAPS

  Now some audacious neuro-speculation. A shameless and gross oversimplification that focuses only on a small part of the brain when far more of it is necessarily involved. In humans, the hippocampus and entorhinal cortex enlarge and differentiate to represent not just places and space but also episodes in time. Recent research has expanded those roles even further, to associative and conceptual spaces. There are two key facts, one about place cells, the second about grid cells, that allow them to represent different real spaces and, later in evolution, abstract spaces. Place cells in h
ippocampus represent integrated sets of features, whether places, episodes, plans, or ideas, as individuals, independent of how they are interrelated. Grid cells represent relations among those places or ideas, spatial, temporal, or conceptual. Like grid paper, grid cells are a template that can be reused, remapped. Voilà! the same neural foundation that serves spatial thought serves abstract thought. It’s as though the hippocampus created checkers or tokens for places or memories or ideas and entorhinal cortex provided a checkerboard for arraying the relations among them in space. Significantly, the array of grid cells, the checkerboard, is two-dimensional, flat, perhaps one reason why thinking in three dimensions is challenging for most people. I repeat: the same brain mechanisms in humans that represent actual places in real spaces also represent ideas in conceptual spaces. Spatial thinking enables abstract thinking.

  We are now ready to proclaim (trumpets, please!) the crucial, central, fundamental tenet of the book: Sixth Law of Cognition: Spatial thinking is the foundation of abstract thought. The foundation, not the entire edifice. We show some of the implications of the Sixth Law in the next section, which recounts many curious and systematic distortions in people’s cognitive maps, distortions that are mirrored in people’s social maps. But the main focus of the second half of the book, Chapter Six and onward, will be the spatial foundation of abstract thought.

  MAPS IN MINDS: COGNITIVE COLLAGES

  The exciting research on place cells and grid cells has shown forcefully that the brain does not have a file drawer of cognitive maps that it pulls out as needed. Instead, cognitive maps are constructed and reconstructed on the fly, from pieces that are distributed in the brain. Grid cells represent spatial relations, but approximately, not exactly, and relative to a frame of reference that keeps changing as the environment explored changes. Collecting disparate pieces to navigate or make judgments is true in spades for humans. But people have far more pieces to use to construct mental maps beyond personal exploration, beyond place and grid cells. People can use specific memories of places they have visited or routes they have taken, but they can also use descriptions of places and routes in language and depictions of them in maps. They now can use mobile phones and augmented reality and who knows what else in the future. People can use spatial schemas, general knowledge about layouts of cities and towns, not just of their own regions and countries but also of other regions and countries. Once I visited Prague and Budapest in succession and realized that they have the same map: a river running north and south, with an old town and a castle on the west bank and a “new” town and art nouveau museum on the east bank. I couldn’t use the maps interchangeably for finding the castles and the museums, but I could use the general schemas to understand the layouts of the cities. Japan has its own distinctive way of organizing cities: into quadrants, labeled the equivalent of NW, NE, SW, SE, with each quadrant divided geometrically into smaller units that are systematically labeled. Once you understand that organization, you can understand the clever and transparent address system, so different from Western systems.

  When making decisions about navigation or judgments of distance or direction or sketching maps, people can make explicit inferences as well as implicit ones. This makes spatial judgment and navigation the same as solving any problem: gather whatever information seems to be relevant and try to make sense out of it.

  Spatial thought, abstract thought

  What follows are some of the proxies people use to make judgments about distances and directions in space or to sketch maps of space. Like our representations of the body, like our representations of the space around the body, the judgments are not random perturbations of physical measurements but systematically biased by the proxies and processes used to create them. Same for biases in the judgments of the larger spaces that we cannot see from one location. That’s significant on its own, but it attains greater significance because the biases in spatial judgments are directly mirrored in biases in social and cognitive judgments in accordance with the Fifth Law of Cognition: Cognition mirrors perception. Earlier, in the discussion of the space around the body, we noted that the mind creates spatial frameworks to keep track of where things are relative to each other. Those spatial frameworks are essentially networks and can be used to keep track of relationships among any set of ideas; ideas in conceptual spaces are like places in real spaces. Similarly, the same processes used in spatial judgments will be shown to be at work in abstract judgments, judgments that are social or cognitive. These parallels between spatial thought and abstract thought strongly support the premise that spatial thought is a foundation for abstract thought. Much more on that in Chapter Six.

  Rotation

  We know about people’s understandings of space from the judgments and inferences and decisions and sketch maps that they make. Now let’s look more closely at those tasks, their findings, and their implications. We begin with perceptual processes that bias judgments, processes rooted in the gestalt principles of perceptual organization, common fate and grouping. Common fate is the expectation that objects that are related should be oriented in the same way. If one slants, they all should. According to common fate, the slant or orientation of a geographic entity should be close to the orientation of its frame of reference. For geographic entities, the frame of reference would be the encompassing structure, in this case, the canonical cardinal directions. This implies that the mind should mentally rotate the geographic entity to be oriented more like that of its frame of reference, here, the cardinal directions. Naturally, the world did not evolve to conform either to grouping or to common fate; other forces determined its evolution. So it turns out that many (if not most) geographic entities, like South America, Italy, Long Island, the San Francisco Bay area, and Japan, are tilted with respect to their encompassing frame of reference, the cardinal directions. The encompassing frame of reference might be an approximate proxy for the true orientation, but it’s still approximate. People think of Milan as in the north and Naples as in the south of Italy, and that’s true, but Naples, on the Mediterranean side of Italy, is way east of Milan and even Venice, which is on the Adriatic side of Italy.

  To find out if minds use common fate even though geography doesn’t, we developed geography quizzes especially designed to make people get them wrong. One asked a group of Stanford students to draw a line to indicate the direction from Stanford, on the west Bay, to Berkeley, on the east Bay. Other Stanford students were asked to draw a line indicating the direction from Stanford, which is inland, to Santa Cruz, which is on the Pacific. The Bay Area actually slants relative to N-S. Even though the correct map was on the road maps that people used then and in the weather maps published in the daily papers, a significant majority of students who lived in the area got the answers wrong. The majority drew lines showing Stanford west of Berkeley and Santa Cruz west of Stanford. Incorrect in both cases. The wrong answers come from rotating the major axis of the Bay Area to be more north-south than it actually is. The Bay Area actually runs almost diagonally with respect to the cardinal directions, but people’s minds rotate it to be more upright relative to a north-south axis. In an informal test of an Italian audience, most raised their hands when asked if they thought that Naples (on the west coast of Italy) was west of Venice (on the east coast of Italy). Those Italians were wrong and surprised to learn that they were wrong. Just as Bay Area residents are surprised to learn that Berkeley is west of Stanford and that Palo Alto is west of Santa Cruz.

  Similarly, people upright South America in their minds; its true orientation seems tilted. When students were given cutouts of South America and asked to paste the cutouts in a rectangular frame oriented north up, a significant majority uprighted South America. You can check your own memory. We then taught students new maps, made-up ones in which the geographic entities were tilted with respect to the canonical axes. Sure enough, when participants were asked to remember the directions between pairs of cities or to orient the map from memory, they made errors in the direction of rotati
on, of using the canonical axes as proxies for the actual orientations. We found the same error when participants were presented with map-like blobs. Ditto for school-aged kids; for these errors, kids and adults look alike.

  Alignment

  A core gestalt organizing principle is grouping similar things together. Objects in close proximity are seen as grouped, and objects that are similar by some attribute such as shape or color or size are seen as grouped. Now another brief geography quiz, taken from an experiment. Which is farther south, Rome or Philadelphia? If you answered “Rome,” you are in excellent company. A majority give that answer. Their reasoning is reasonable, but the answer is wrong. For this question and many similar ones, people seem to rely on a perceptual inference, grouping by proximity. The mind groups the first row of x’s into two groups of three and the second row into three groups of two.

  xxx xxx

  xx xx xx

  When groups aren’t quite aligned, the mind treats them as virtually aligned. And so for large geographic bodies, like Europe and the United States. The mind groups them and aligns them along an east-west axis even though most of Europe is north of most of the United States. Philadelphia is located in the northern part of the United States, Rome is located in the southern part of Europe, so people make the not unreasonable inference that Philadelphia should be north of Rome. But it’s not. Grouping works north-south as well as east-west. Mentally aligning the United States and South America leads the majority to say that Boston is east of Rio because Boston is far east in the United States, and Rio is below Brazil’s bulge, so not on the easternmost part of South America. This bias, to group similar geographic entities, has been termed alignment.