Believing

by Michael McGuire

  It can’t be emphasized enough how critical models are for managing the activities of daily life. There are literally hundreds of such models associated with almost everything we do: not changing the oil in your automobile will lead to motor damage, thoroughly chewing food prevents indigestion, good examination grades in high school will get one into a good college, and careful techniques produce good surgical outcomes. We believe them. As they become more precise and predictive of outcome, a reduction in computational requirements is likely coupled with their conversion to beliefs.

  MOSTLY INNATE MODELS

  Multiple studies of child development have established that children can think before they speak,19 and they are born with the capacity to use models to deal with how the world works. For example, by sixteen months of age, they can infer the causes of failed actions:20 if a light switch is flipped and the light doesn’t go on, it could be due to selecting the wrong switch or a broken bulb. Children can figure out the most likely option. They also have concepts of property rights21 and can grasp abstract geometric principles despite the absence of math training.22 They apply cause-and-effect models in their interpretations of events that are closely related in time, such as noise from the kitchen at dusk is due to preparation for dinner. In short, children enter the world with strong predispositions to process information in specific ways.23

  Model-related influences on behavior may extend throughout life. A person experiencing the sudden onset of anxiety while speaking for the first time to a large audience may avoid speaking in such a forum again. Exposure to high places accompanied by fear that one will fall or might jump often leads to the subsequent avoidance of heights. The presence of pain following physical activity may lead to a change in work or exercise habits. Models appear to be critical for intuition, especially when they are the source of near-instant explanations of complex events or information.

  Mostly innate models are so fundamental to the way we think that they are not easily unlearned. This is due in part to their predictive accuracy and utility and in part to the absence of compelling alternative models for much of what is experienced in everyday living. To maintain that coincidence explains two events that are closely connected in time is seldom convincing. Still, within limits, mostly innate models can be modified. Adding a temporal feature separating cause and effect is an example. The child who plants his first radish seed may expect signs of growth within hours. With experience, he will expect seeds to sprout after a week or two.

  There is a long list of familiar variations of cause-and-effect models. A sampling includes the following:

  (a) Repeat behavior associated with pleasure. Children act in ways that are consistent with this model. They need not be taught that ice cream is a source of pleasure: a model makes the connection. Information may have a similar outcome. The sight of food, for example, may be followed by a desire to eat. As mentioned, experiments show that brain activity in areas associated with pleasure and reward precede actions that have a history of pleasurable responses.24 This may explain why disregarding the behavior-influencing effects of mostly innate models associated with pleasure is so difficult. (One downside of pleasure-based associations is that they appear to be a significant contributing factor in drug addiction, high-risk behavior, and behavior perversions.)

  (b) Avoid behavior that is associated with displeasure or the absence of reward. Children act in ways that are consistent with this variation.

  (c) Repay those who are the cause of personal injury. While emotions and plans for retaliation are not uniformly followed by action, they rarely disappear overnight. Formerly married individuals who consider themselves victims of infidelities of their ex-spouses frequently fit this model. So too with those who have experienced insults, pain, or betrayal at the hands of others. A personal experience catches the flavor of these situations.

  During my senior year in college, I lived on the third story of an apartment building. There was an empty apartment next to mine. A new tenant was expected soon. One Sunday morning, as I departed from the apartment to get a donut and a cup of coffee, a car with a large hooded trailer arrived. I suspected it was the new tenant. I introduced myself. He did likewise. I asked if I could help him move in. He accepted. Two hours later and after our thirty trips up three flights of stairs, his belongings were in his apartment.

  Several months passed, during which we encountered each other infrequently. Then the time arrived for me to vacate my apartment. In anticipation of the move, I knocked on my neighbor’s door. When he opened the door I asked if he “would help me move the following Saturday?” “I don’t return favors” was his reply as he shut the door. I was consumed with anger.

  It is now decades later. Each time I recall the incident, I still fill anger. And to this day, I wish I had responded differently.

  (d) An intimate relationship with an attractive person will be pleasurable. While positive and inviting feelings about interacting with attractive persons are most intense during the teens and early twenties, they continue throughout life and are difficult to modify.

  (e) Members of groups other than one’s own are dangerous. If this model isn’t mostly innate, how can the near-universal distrust and competition among groups be explained?

  None of this suggests that mostly innate models accurately depict how the world works. Nonetheless, they tend to narrow divides and all the more so when they consistently predict outcomes. The more they do so, the more likely they become beliefs.

  MOSTLY LEARNED MODELS

  It has been known for nearly a century that, during development, children change their model of balance scales—teeter-totters—several times. From an initial focus on weight to a later consideration of the distance dimension, they eventually settle for a model incorporating multiplication of the relevant factors affecting balance.25 As simple as a balance scale may seem to adults, it wasn’t always so. Models that develop this way are sometimes referred to as “bottom-up” models. These are applicable to learning selected skills and involve taking procedural or practical knowledge and transforming it into conceptual knowledge—that is, tool-kit-related knowledge. Such models contrast with “top-down” models in which conceptual knowledge is translated into procedural knowledge. This happens, for example, when receiving instruction about how to use a complex machine. Either way, the outcome may be similar: once models are learned, they are used by the brain to interpret information.26 And, typically, the steps of learning are usually lost to recall: for example, try recalling your first focus on weight as an explanation for the behavior of teeter-totters.

  No doubt many learned that bottom-up and top-down models merge. This happens, for example, in learning to play tennis: if the trajectory of the tennis ball is other than one predicts, adjustments in one’s footing and tennis stroke follow, but within the options and constraints of the game’s rules. More technically, working memory is first transferred to movement and model development.27 This is followed by modification of models in response to model-variant experiences, a point that is predicted from studies that show characteristic neural activity in response to new information.28 For both bottom-up and top-down models, practice is critical for refinement. It offers the advantage of direct evidence regarding model predictability and may lead to model revision and use.

  Another type of learned model is analogous to an algebra equation: the rules for its application are specified, but what the symbols represent may differ. For example, the model A = B + C - D can apply to who is and isn’t invited to a dinner party, a new business plan, a recipe, or a design for a military action in a distant country. At times, revision occurs following model testing, such as tasting a meal cooked according to a recipe designed to achieve a special flavor and texture.

  Then there are idiosyncratic models, such as those developed by families dealing with how they should behave following the death of kin: A = B + C + D + E + F ÷ G - H, in which the symbols refer to such things as kin coming together, family members show
ing remorse, a funeral, eulogies, and burial. Most but not necessarily all of the variables in the model may be honored. Idiosyncratic models based on imaginings and beliefs are also active in daily life. The world of spirits, gods, angels, and animals with humanlike motives and personalities provide much of their content.

  Not surprisingly, the manner in which mostly learned models organize and structure behavior is subject to multiple influences. This occurs because imaginings, beliefs, brain reading, possibly mirroring, and often considerable trial and error differentially contribute to model development and modification. Nonrational features of reasoning and belief, such as those discussed in chapter 4, frequently complicate matters—apparently there is no standard formula for development and modification. Nonetheless, many such models possess a common property: they are attributed to others, animals, inanimate entities, or the supernatural in the process of explaining behavior.

  In what ways do models fit in the scenario of happenings forty-five thousand years ago? The likely answer is similar to that for stories: they increasingly contributed to symbolic and technological complexity of the period as well as beliefs. For mostly innate models, there may have been few differences from those of today. For mostly learned models, it’s highly probable that their number, complexity, and degree of refinement have increased significantly—their utility has been recognized. Moreover, at the present time, they are likely to outnumber mostly innate models: new mostly innate models presuppose genetic change, which is slow, while new mostly learned models can be created in moments.

  Before my first trip to West Africa, I visited the Los Angeles Zoo to familiarize myself with animals I might encounter. Both black and green mambas were of special interest. Lore had it that no one had ever survived a bite from one.

  After spending ten minutes looking through the glass-enclosed cage labeled “Black Mamba,” I couldn’t find a snake. Eventually a zoo employee came by and I asked for assistance. She pointed to the snake. It was resting on a branch of a tree with its head extended some eight inches beyond the end of the branch, thus appearing as an extension of the branch. If I couldn’t identify the snake two feet away in the cage, I wondered what my chances would be in the wild.

  For the following two weeks, I had a recurrent dream. I was somewhere in Africa, preparing to go to bed. I would take off my shirt and hang it on what appeared to be a branch of a tree. The branch, of course, was a black mamba, just as in the zoo in Los Angeles. Each time I tried to hang the shirt, it dropped to the ground below the branch. I would pick it up and hang it on the same branch again. Through all this, the mamba was becoming increasingly irritated and prepared to strike. Just as it was about to strike, I would wake up anxious and sweating. After two weeks, I no longer had the dream.

  A month later, I was in Africa. Throughout the trip, whenever I saw a branch that resembled what I had seen in Los Angeles, I experienced a moment of anxiety followed by very careful vigilance. The branch served as a trigger for a specific response both emotionally and cognitively. (I never saw a black or a green mamba on the trip.)

  Triggering occurs when either external or internal information (stimuli) initiates brain activity and specific states of awareness. It is a special case of the extended mind hypothesis. It is so commonplace and frequent that we rarely notice that it is occurring. It has an interesting if not profound implication however: it’s not quite that we are walking around the world as “freethinkers” and “impassionate observers”; rather, to a significant degree, what we experience in awareness is determined by the environment in which we find ourselves.

  INNATE RELEASING MECHANISMS

  The concept of triggering was developed by Nicholas Tinbergen during his studies of animal behavior: a specific stimulus—information—elicits a predictable behavioral response in another animal via its effects on the animal’s innate releasing mechanisms. Such behavior is definable by its form and structure. It is species universal in that innate releasing mechanisms have evolved to respond to specific releasing stimuli.1 Releasing stimuli can be cumulative when more than one stimulus is required to initiate a response, or they may be additive, as can occur when the same stimulus is repeated.2 Typical behavioral responses among animals include escape, prey catching, courtship, and fighting.

  At times, animals behave in ways that suggest that their behavior is initiated by innate releasing mechanisms. This happens when dogs are trained to follow commands such as “sit” and “fetch.” Certainly dogs don’t know the meaning of these commands when they are born, nor are they versed in the many languages and thus different stimuli in which such commands are given. Nonetheless, with training, commands are followed. Brain mechanisms associated with learning account for this type of behavior, which is not species universal.

  HUMAN PARALLELS AND SIMILARITIES

  Although some scholars question the presence of innate releasing mechanisms in humans, a reasonable case can be made that they exist.3 Examples include reflexes, such as blinking one’s eyes in response to the flash of a bright light and withdrawal from a painful stimulus. Mirroring also may qualify: a specific behavior by others initiates predictable neuronal activity among observers. Recently it has been postulated that information present in the brain, such as an image of God, can serve as a stimulus that initiates the nonvocal aspect of petitioning prayer.4

  Less straightforward examples are found in stimuli such as loud sounds, disgusting smells, and inviting sexual postures that initiate specific responses. Such responses occur frequently, but they can be modified and don’t appear to be species universal. (The fact that advertisements often include pictures of attractive women and men or locations that are unrelated to what is advertised suggests that advertisers believe that human purchasing behavior can be triggered—recall the discussion in chapter 5 dealing with neuromarketing.)

  Then there are behavior responses among humans that are primarily learned yet may be highly predictable. For example, people indoctrinated with religious stimuli early in life frequently retain their responses throughout their lives. For these individuals, a picture of a religious figure or listening to a hymn triggers a specific behavior and emotion. But religious indoctrination is not unique. A similar point applies to well-known friends, disliked enemies, infants and parents, and spouses: for example, the posture, the voice tone, and what is said by spouses often trigger specific states of awareness. Much the same applies to bosses and employees. Such responses may be viewed positively in that they facilitate social communication, job efficiency, and sensitivity to others. But there are other considerations.

  RESPONSES TO TRIGGERING

  Among humans, responses to social stimuli might develop under the following conditions. (1) A high percentage of the social stimuli of daily life are similar and predictable. (2) Predictable responses to these stimuli facilitate social interaction. (3) Predictable responses reduce the brain’s computational requirements and energy expenditure.

  That responses to triggers under condition (1) are primarily learned seems clear, although a well-functioning memory is essential. Studies show that repetition improves memory—recall learning the multiplication tables—and increases the likelihood that a specific stimulus will lead to the retrieval of a specific response.5 How this might take place in the brain is suggested by studies showing that memory repetition is associated with neural pattern similarity, which enhances episodic recall to the same stimulus.6 And, as noted earlier, the resting brain recapitulates activity patterns that occur during recent experiences, which appear to contribute to memory longevity, specificity, and preparation for similar events in the future.7

  Condition (2)—predictable responses to social triggers facilitate social interaction—also appears to be primarily learned. Such responses should considerably reduce the potential for ambiguity that can accompany such interactions.

  Condition (3)—predictable responses reduce the brain’s computational requirements and energy expenditure—is consistent with the view tha
t brain systems are biased in favor of conserving energy.

  Although it is seldom viewed this way, triggers are a type of brain control. If I respond in a predictable way to a verbal stimulus, the person responsible for the stimulus briefly controls activities in my brain. In familiar situations, undesirable consequences rarely follow. However, as is discussed in upcoming chapters, the effects and consequences of this type of brain control can be significant.

  The preceding points add up this way. When persons live in predictable social and physical environments, a large percentage of the stimuli they encounter are similar in form and structure. They respond to these stimuli in highly predictable ways. As a result, computational requirements dealing with such interactions are reduced. A rough estimate is that 40 percent of the social interactions that occur in daily life meet these criteria—this percentage hints at why it is stressful being in an unfamiliar social setting or physical environment. Two points follow. The prevalence of triggering and predictable responses is a rough measure of the degree to which humans are socialized and influence each other. The effects of triggering are exceedingly difficult to avoid or control.

  There are other types of triggering.

  The two guides who took me up the river where I met Chief Fred had very different talents. One was a natural linguist and spoke at least half a dozen local dialects along with English and French. He was the translator in the meeting with the chief. The other was an expert in the environment and its animals. For most of his thirty-five years, he had lived in the bush. He excelled as a hunter and in finding the occasional person lost in the forest. As we trekked through the forest in search of monkeys, he was our guide. His capacities seemed uncanny. We would walk for a time and not see an animal. The usual sounds were those of birds, the wind, and falling branches. Then, unexpectedly, he would hold up his hand, point to an area, and whisper, “Monkeys are there.” Invariably they were. Neither the linguist nor I were able to match his skills.