Affect and Information
Emotion is a topic so central that it will be treated at length in Chapter 8. Here we need to see in a general way how and why emotion, or affect more generally, plays a pivotal role in the mental architecture of Homo prospectus.
Emotion has been contrasted with cognition and reason by philosophers and psychologists alike. But the last few decades have seen a revolution in thinking about affect, a revolution that places affect squarely in the middle of our capacity to represent the physical and social world, to evaluate prospects or perils we face there, and to guide thought and action accordingly (de Oliveira-Souza, Moll, & Grafman, 2011; Pessoa, 2008).
Why would affect (i.e., feeling) be suited for such a central role? Let's start with Robert Zajonc's work, which began the "affective revolution" in psychology. He offered experimental evidence that affective responses preceded and shaped cognitive responses—the hypothesis of "affective primacy"—and constituted a source of information relevant to the organism's needs and goals (Zajonc, 1980, 1984). Zajonc's painstaking experiments took place prior to the massive development in brain imaging capability in the 1990s and were largely based on the long-standing method of measuring reaction times. Evolutionarily, it makes a great deal of sense that a reaction of fear, anger, or surprise should take place fast enough to enable one to contend with the threat, harm, or unexpected event they signal. For example, within a 10th of a second after exposure to the presence of a threatening animal—even if the visual information is wholly unexpected and not at the center of attention—the mammalian fear response is already underway, beginning to reorient attention, intensify alertness, accelerate brain activity, prime memories of similar prior events, increase heart rate and slow digestion, ramp up action- readiness, and prepare a response (Luo et al., 2010). It is characteristic of affect that it can act simultaneously and directly upon all of the mind's key systems—attention, perception, cognition, memory, motivation, felt experience, and action—as well as the body itself. Affect therefore can orchestrate the individual's response to a wide variety of challenges or opportunities in the physical and social environment (for a summary, see Rolls, 2007).
Today in cognitive social psychology, it has become customary to speak of valence—degrees of positivity or negativity—as an essential and distinguishing feature of affect, making it possible for affect to function as the brain's currency for evaluation or appraisal (Shuman, Sander, & Scherer, 2013). The idea that affect is evaluative flows naturally from affective responses playing the same sort of role in adjusting one's thought and action to a situation that evaluation would play. Liking and loving, hoping and trusting, respecting and admiring, are positive in valence, directly inducing attraction, approach, acceptance, and credence. Fear and distrust, disgust and revulsion, hatred and contempt, are negative in valence, directly inducing aversion, avoidance, rejection, and disbelief. As Jonathan Haidt describes the affective revolution that has swept social psychology, "the basic point was that brains are always and automatically evaluating everything they perceive, and that higher-level thinking is preceded, permeated, and influenced by affective reactions" (Haidt, 2007, p. 998).
Some affective states, such as fear, anger, and surprise, involve arousal, shifting our psyche from "business as usual" to heightened and refocused activity. When consciously felt, these tend to have a distinctive, attention-grabbing phenomenology. Other affective states, however, are low arousal, such as confidence, interest, liking, and satisfaction, which underwrite everyday living and encourage us to get on with our lives. Such everyday, unaroused emotions function as a kind of "default" setting of the psyche, to which a healthy psyche returns when more agitated, aroused emotions fade. The default character of these attitudes helps explain their "thin" phenomenology—to do their job, they should (like the default programs of a personal computer) be able to process information and ready responses in systematic, task-relevant ways without creating a distraction on the user interface. For example, everyday confidence in one's eyes and ears sustains spontaneous perceptual learning and belief formation while being barely noticeable. Only when some unusual circumstance or condition undermines this confidence, as when vision suddenly becomes blurry, the ears seem to "hear voices" we know are not there, or an anxiety attack pulls the rug from under self-assurance, is it clear what it is like to have such baseline trust in our faculties and their deliverances.
It seems clear why the default emotions must be in the positive register on the whole: If we were generally disposed not to have confidence in our perceptions, thoughts, or capacities for action, not to trust others, not to enjoy satisfying our needs or accomplishing our goals, not to take a positive interest in life, we would be in a dangerously self-defeating condition. Unable to trust our senses, we would be unable to learn from experience; incapable of sustaining confidence in our thoughts, we would be unable to follow a line of reasoning or to draw on our memory; lacking desire or pleasure in satisfying our wants, we would lose motivation to keep ourselves well, maintain our relationships, or pursue our goals. We would be boxed out of life and learning by our own psyches. It is, therefore, interesting that across many nations and tens of thousands of informants, 86% of informants reported themselves above neutral in subjective well-being, with an average life satisfaction in developed countries ranging from 6 to 7 or higher on a 10-point scale (Diener & Diener, 1996). In short, if we were not generally positive, we would be in a state of affective dysfunction, similar in its effects to the ways chronic depression erodes learning, memory, thought, motivation, and action. As we will see in Chapter 10, there are profound links between affective dysfunctions and dysfunctions of prospection. Here we are concerned primarily with how affect works in tandem with prospection when things are going reasonably well.
The evaluative role of affect includes not only the learning and representation of value, but also the value-based guidance of thought and action. Much of the content of affect is itself prospective. Part of what it is to have a goal, for example, is to have a positive affective attitude toward the thought of attaining it and a positive affective expectation from its realization. These in turn translate into a positive affective attitude toward means to realize that goal. Because affect is intrinsically linked with motivation as well as valence, this favorable attitude in turn motivates the effort needed to perform the means of attaining the goal or overcome the obstacles that stand in the way. The anticipation of effort and obstacles, too, has an affective expression, and this dispreference motivates us to look for means that are less demanding or difficult, if they can be found. When obstacles prove difficult to overcome, the negative affect of frustration makes us discontent with accepting failure and motivates us to invest the additional effort needed to find a solution, or, if that proves unavailing, to revise our plans. By contrast, making progress toward a goal is satisfying, but as encouragement rather than mere satiation. Successful effort is thus rewarded, but further effort is made more attractive as well. In these ways and more, affect gives values the practical force needed to turn them into planning and action.
If the affective system is indeed evolved to inform and help guide action, then we should expect that it does more than shape how we respond to the environment; it should reshape itself in response to the environment in order to remain as informative as possible. Experiments with rats suggest that the brain's "affective keyboards" for risk and reward are not fixed, but can be "retuned" in response to changes in the level of risk and reward it faces. In a more dangerous environment, for example, a greater portion of the nucleus ac- cumbens shell becomes devoted to fear-generation, to permit greater discrimination among degrees of risk; when risk is removed and the environment made friendlier, a greater portion of the shell becomes devoted to the discrimination of degrees of positive valence and appetitive motivation (Reynolds & Berridge, 2008). Like any biological system, these tunings are not perfect; particular kinds of experiences or drugs can result in the keyboard being out of tune in various ways. Chronic insecurity and stress in early childhood or severe trauma in adolescence or adulthood can result in individuals being hypersensitive to particular kinds of situations or drugs. A normally flexible, adaptive system that helps individuals cope with the vicissitudes of life can become distorted in ways that make life much more difficult. This can cause various kinds of disorder or addiction (Herman, 1992; Knudsen, Heckman, Cameron, & Shonkoff, 2006; Lupien, McEwen, Gunnar, & Heim, 2009). Proper attunement of paired forms of affect, such as confidence and fear, liking and disliking, anger and amiability, is as essential in daily life as it is in emergency situations or situations of great risk. Excessive confidence can render us foolhardy, socially insensitive, overly certain of our own opinions, and oblivious to evidence and to the feelings of others, whereas excessive fear can render us anxious, withdrawn, irresolute, suspicious, defensive, and mean-spirited. Out-of-tune affective responses are often just as great an obstacle to successful navigation of the physical and social world as poor sensory perception or physical coordination. Indeed, recent research relates a number of psychological disorders to failures of the learning processes of the affective system.
What we increasingly find in contemporary psychology, then, is a conception of the affective system as an experientially calibrated information-processing system. This system makes a key contribution to our ability to learn about, anticipate, evaluate, estimate, and act upon the prospects or perils of the world (Nesse & Ellsworth, 2009). The idea of "affect as information"—shaping cognition and helping guide behavior—has gained increasing support (Schwarz & Clore, 1983, 2007). Thus, although everyday life does not seem full of emotion, once we take into account default, as well as aroused, affect and think about what life is like in those moments when we lose default affective responses for whatever reason, we realize that an emotional tone—when healthy, typically mildly positive—is always there.
This conception of the affective system also helps explain a number of puzzling findings in the psychology of happiness or, more precisely, subjective well-being, which is an average of an individual's self-evaluation of two variables: recent affective state and overall sense of life satisfaction. We have already seen why a positive score on the typical 10-point scale of subjective well-being should be expected to be the norm in most human societies, even those with relatively low standards of living. But social scientists have been puzzled that once individuals attain a reasonable level of material sufficiency (around $60,000^80,000 in the United States in 20082009), their expressed level of satisfaction increases only with exponential increases in income, and their overall positive affect does not really increase at all (Diener, Sandvik, Seidlitz, & Diener, 1993; Kahneman & Deaton, 2010). Why doesn't greater wealth make us happier? Equally intriguingly, even though average levels of subjective well-being are relatively constant within a population or over relatively long spans of individual lives, there is noticeable variation within individuals over shorter periods of time, such as the course of weeks and even days, in response to changes in their condition or information state (Eid & Diener, 1999, 2004; Schimmack & Oishi, 2005). When workers learn that they have just received a healthy raise, this characteristically results in a higher level of life satisfaction, but the effect usually fades with time, and after some number of months, they will typically report a level of subjective well-being essentially the same as it was before the raise. This has been called a "hedonic treadmill," as individuals strive to increase their well-being while always reverting to a personal "set point," which varies somewhat from individual to individual (Brickman & Campbell, 1971; Watson & Clark, 1994). At the same time, finely grained daily events cause fluctuations in subjective well-being (Suh, Diener, & Fukita, 1996). Thus, a morning frustration, even relatively small, can send one's life satisfaction down, while an afternoon success, even relatively small, can send it up. But not all of life's changes work this way—some changes in condition leads to a persistent loss of life satisfaction, as for example, among those involuntarily unemployed in highly developed countries (Lucas, Clark, Georgellis, & Diener, 2004).
This all seems very difficult to fit into a coherent picture of subjective well-being until one thinks of affect as information. Default affect, we argued, must typically fall in the positive range if the individual is to sustain the ordinary business of life. Fine-grained fluctuation, however, makes sense if subjective well-being is telling us how well things are going from moment to moment in our lives and whether our prospects are improving or deteriorating— whether we should make some changes or adjustments in our behavior or to simply keep doing what we are doing. Return to a set point is a generic feature of many effective information-sensing systems. If the system ratcheted upward with each gain and stayed there until some decrement occurred—in the case of wellbeing, say, if the better one's life had gone, the higher the level would be, right up to the top of the scale—then the "needle" on the gauge of the sensing system would be pushed up against the top of its range, with less and less ability to signal the next gain. However, if there are conditions that one must persistently seek to find a way to overcome (i.e., obstacles to getting by), such as involuntary unemployment, a more persistent decrement is found on subjective well-being, because one is not able to return to business as usual (Lucas et al., 2004). All this is speculative, but it suggests how thinking of affect as functioning primarily to provide thought- and action-guiding information related to the meeting of life's large and small challenges, affords us a new lens for viewing what had been a very unclear picture of a central fact about human life. Moreover, it points directly toward the prospective character of affect: Affective states coordinate mental and bodily states to prepare the individual for what is to come, not simply to register a response to what is present now or what has come before.
Think back on your reaction to the email invitation and your mulling over how to respond. Perhaps, when you first saw the message, you felt almost instantly that you'd like to decline. But no sooner had this idea crossed your mind than you felt unsure about it and unhappy with it. So your mind ranged a bit further and turned up a complex interplay of personal and social feelings, each of which can be seen as attentive to one or more dimensions of the problem and each of which exerts some force in shaping the overall "feeling" that eventually led to your decision to accept. These responses drew directly on your own "affective palette," but they imaginatively "tried on" possible responses and used this palette to paint a psychic portrait of what those possibilities might be like, to which, in turn, one could react positively or negatively, confidently or anxiously. This process involved an immediate and involuntary form of empathic projection, a simulation of how things might look or feel from your coworker's standpoint, using your own affective system as a "test bed," whether or not these reactions on his part would be justified in your view. Just as our perceptual system keeps us in touch with what is happening to the objects around us, empathy keeps us in touch with what is happening to the people around us, and each is a vital stream of information needed in real time if we are to flourish as social beings.
In the normal case, these streams of information flow continuously without need for conscious intervention: We'll "see" the anger or sadness on colleagues' faces when we encounter them in the hallway, even though we were concentrating on something else at the time. And during a fluent conversation, we will "see" each other's shifting responses to what has been said almost as soon as they occur. Indeed, when this breaks down, conversation can become awkward and ineffective: "Yes, we talked about it. But I don't think they heard what I was saying." Part of what makes your quandary over how to respond to your colleague's email difficult is that you must anticipate his reaction without the visual information face-to-face communication would make possible. Empathy has limits and is, of course, entwined with our own cognitive, imaginative, and affective limitations. Various circumstances and attitudes will tend to inhibit empathy, and we may lack the responsive range in our own affective system to understand each other spontaneously if our differences in culture or life experience are sufficiently great.
From the standpoint of prospection, what is notable about empathy, however, is that spontaneous empathic simulation of others generates a predictive internal model of them, which can be more or less accurate depending on the psychological resources we can bring to bear. The more effectively we empathize, the better we will be at modeling and predicting. In our discussion of linguistic competence, we noted that positing some implicit, organized body of knowledge that speakers can spontaneously draw on is thought to be necessary in order to understand how humans can so universally and effectively learn their native languages and go on to produce and understand an enormous number of novel sentences. But producing and understanding sentences is hardly a grammatical task alone: The hardest parts in conversation are knowing what to say, how to phrase it, when to speak, what others' words mean, or what they are thinking and feeling and how to adjust to this. For such sociolinguistic competence to be an open-ended capacity that most humans achieve to some degree—they could hardly get on in the social world or even make sense of one another's words if they did not—the amount of implicit, organized knowledge and information they are able to spontaneously draw on must be huge. This seems possible only if individuals also spontaneously form multidimensional internal models of one another as they move through the social world. As in the case of language, we can speculate that the human mind is set up by evolution to be able to extract information from others' behavior to form such generative internal models of what they are thinking and feeling. For example, evidence from developmental psychology suggests that even at a very early age, human infants are able to distinguish the feelings and intentions of others (e.g., Barrera & Maurer, 1981; Meltzoff, 1995; Onishi & Baillargeon, 2005; Serrano, Iglesias, & Loeches, 1992; Southgate, Senju, & Csibra, 2007).