Self-Regulatory Perspectives on Personality Research Paper

We were both trained as personality psychologists. Throughout our careers, however, our research interest has focused on a set of issues regarding the structure of behavior. These issues link the concept of personality and its functioning to a set of themes that might be regarded as representing the psychology of motivation. Our interest in how behavior occurs has taken us into a number of specific research domains—most recently health-related behavior and responses to stress (Carver & Scheier, 2001; Scheier, Carver, & Bridges, 2001). However, these specific explorations have almost always occurred in service to a more general interest in the structure of behavior.

What we mean by “the structure of behavior” is reflected in the issues underlying questions such as these: What is the most useful way to think about how people create actions from their intentions, plans, and desires? Once people have decided to do something, how do they stay on course? What is the relation between people’s values and their actions? What processes account for the existence of feelings as people make their way through the world?

As we have tried to address such questions, we have consistently returned to the idea that people are self-regulatory entities. That is, human behavior is an attempt to make something occur in action that is already held in mind. Similarly, affects serve as self-regulatory controls on what actions take place and with how much urgency.

The self-regulatory principles we emphasize in our writings were not conceived as being a model of personality. However, the principles do turn out to provide an interesting perspective on personality. They suggest some implications about how personality is organized and expressed in people’s actions. These principles also point to some of the issues that are involved in successfully negotiating the world. The principles we emphasize deal most explicitly with the “process” aspect of personality—the functions that make everyone a little bit alike—but they can also be seen to have implications for the individual differences that are part of personality psychology.

This research paper is organized as a series of conceptual themes that reflect this self-regulatory perspective on personality. We start with basic ideas about the nature of behavior and some of the processes by which we believe behavior is regulated. We then turn to emotion—how we think it arises and a way in which two classes of affects differ from each other. This leads to a discussion of the fact that people sometimes are unable to do what they set out to do and of what follows from that problem. The next sections are more speculative and reflect emerging themes in thinking about behavior. They deal with dynamic systems, connectionism, and catastrophe theory as models for behavior and how such models may influence how people such as ourselves view self-regulation.

Behavior as Goal Directed and Feedback Controlled

The view we take on behavior begins with the concept of goal and the process of feedback control, ideas we see as intimately linked. Our focus on goals is in line with a growing reemergence of goal constructs in personality psychology (e.g., Austin & Vancouver, 1996; Elliott & Dweck, 1988; Miller & Read, 1987; Pervin, 1989), constructs known by a variety of labels such as current concern (Klinger, 1975, 1977), personal strivings (Emmons, 1986), life task (Cantor & Kihlstrom, 1987), and personal project (Little, 1983). The goal construct is at its core very simple. Yet these theories all emphasize that it has room for great diversity and individualization. For example, any life task can be achieved in diverse ways. People presumably choose paths for achieving a given life task that are compatible with other aspects of their life situation (e.g., many concerns must usually be managed simultaneously) and with other aspects of their personality.

Two goal constructs that differ somewhat from those named thus far are the possible self (Markus & Nurius, 1986) and the self-guide (Higgins, 1987, 1996). These constructs were intended to bring a dynamic quality to conceptualization of the self-concept. In contrast to traditional views, but consistent with other goal frameworks, possible selves are future oriented. They concern how people think of their asyet-unrealized potential, the kind of people they might become. Self-guides similarly reflect dynamic aspects of the self-concept.

Despite differences among these various constructs (see Austin & Vancouver, 1996; Carver & Scheier, 1998), they are the same in many ways. All include the idea that goals energize and direct activities; all implicitly convey the sense that goals give meaning to people’s lives (cf. Baumeister, 1989). Each theory emphasizes the idea that understanding the person means in part understanding the person’s goals. Indeed, the view represented by these theories often implies that the self consists partly of the person’s goals and the organization among them.

Feedback Processes

How are goals used in behaving? We believe that goals serve as reference values for feedback loops (Wiener, 1948). A feedback loop, the unit of cybernetic control, is a system of four elements in a particular organization (cf. MacKay, 1956; Miller, Galanter, & Pribram, 1960). The elements are an input function, a reference value, a comparator, and an output function (see Figure 8.1).

An input function is a sensor.Think of it as perception.The reference value is a bit of information specified from within the system. Think of it as a goal. A comparator is something that makes continuous or repeated comparisons between the input and the reference value. The comparison yields one of two outcomes: values being compared either are or are not discriminably different from one another. Following the comparison is an output function. Think of this as behavior (although the behavior sometimes is internal). If the comparison yielded “no difference,” the output function remains whatever it was. If the comparison yielded “discrepancy,” the output changes.

There are two kinds of feedback loops, corresponding to two kinds of goals. In a discrepancy-reducing loop (a negative feedback loop), the output function is aimed at diminishing or eliminating any detected discrepancy between input and reference value. It yields conformity of input to reference. This conformity is seen in the attempt to approach or attain a valued goal.

The other kind of feedback loop is a discrepancy-enlarging loop (a positive feedback loop). The reference value here is not one to approach, but one to avoid. Think of it as an “antigoal.” An example is a feared possible self. Other examples would be traffic tickets, public ridicule, and the experience of being fired from your job. This loop senses present conditions, compares them to the anti-goal, and tries to enlarge the discrepancy. For example, a rebellious adolescent who wants to be different from his parents senses his own behavior, compares it to his parents’ behavior, and tries to make his own behavior as different from theirs as possible.

The action of discrepancy-enlarging processes in living systems is typically constrained in some way by discrepancyreducing loops (Figure 8.2). To put it differently, avoidance behaviors often lead into approach behaviors that are compatible with the avoidance. An avoidance loop creates pressure to increase distance from the anti-goal. The movement away occurs until it is captured by the influence of an approach loop. This loop then serves to pull the sensed input into its orbit. The rebellious adolescent, trying to be different from his parents, soon finds other adolescents to conform to, all of whom are actively deviating from their parents.

Our use of the word orbit in the last paragraph suggests a metaphor that may be useful for those to whom these concepts do not feel very intuitive. You might think of feedback processes as metaphorically equivalent to gravity and antigravity. The discrepancy-reducing loop exerts a kind of gravitational pull on the input it is controlling, pulling that input closer to its ground zero. The discrepancy-enlarging loop has a kind of antigravitational push, moving sensed values ever farther away. Remember, though, that this is a metaphor. More is involved here than a force field.

Note that situations are often more complex than the one in Figure 8.2 in that there often are several potential values to move toward. Thus, if several people try to deviate from a mutually disliked reference point, they may diverge from one another. For example, one adolescent trying to escape from his parents’ values may gravitate toward membership in a rock band, whereas another may gravitate toward the army. Presumably, the direction in which the person moves will depend in part on the fit between the available reference values and the person’s preexisting values, and in part on the direction the person takes initially to escape from the anti-goal.

Feedback processes have been studied for a long time in a variety of physical systems (cf. Wiener, 1948). With respect to living systems, they are commonly invoked regarding physiological systems, particularly those that maintain the equilibriums that sustain life. We all know of the existence of homeostatic systems that regulate, for example, temperature and blood pressure. It is a bit of a stretch to go from homeostatic maintenance processes to intentional behavior, but the stretch is not as great as some might think (see Miller et al., 1960; MacKay, 1956; Powers, 1973).

One key to this extrapolation is the realization that reference values for feedback loops need not be static. They can change gradually over time, and one can be substituted quickly for another. Thus, a feedback system need not be purely homeostatic. It can be highly dynamic—chasing (and avoiding) moving targets and changing targets. This is not too far from a description (albeit a very abstract one) of the events that make up human life.

Some years ago we argued that the comparator of a psychological feedback process is engaged by self-focused attention (Carver, 1979; Carver & Scheier, 1981). Indeed, the similarity between self-focus effects and feedback effects was one thing that attracted us to the feedback model in the first place. Self-focused attention leads to more comparisons with salient standards (Scheier & Carver, 1983) and to greater conformity to those standards. On the avoidance side, self-focus has led to rejection of attitudinal positions held by a negative reference group (Carver & Humphries, 1981) and to stronger reactance effects (Carver & Scheier, 1981).

The literature of self-awareness is not the only one in personality–social psychology that fits well the structure of the feedback loop. Another good example (Carver & Scheier, 1998) is the literature of social comparison. People use upward comparisons to help them pull themselves toward desired goals. People use downward comparisons to help them force themselves farther away from (upward from) those who are worse off than they are.

Re-emergent Interest in Approach and Avoidance

Our interest in the embodiment of these two different kinds of feedback processes in behavior is echoed in the recent emergence of interest in two modes of regulation in several other literatures. One of the most prominent of these literatures stems from a group of theories that are biological in focus. Their research base ranges from animal conditioning and behavioral pharmacology (Gray, 1982, 1987b) to studies of human brain activity (Davidson, 1992a, 1992b; Tomarken, Davidson, Wheeler, & Doss, 1992). These theories assume that two core biological systems (sometimes more) are involved in regulating behavior.

One system, managing approach behavior, is called the behavioral activation system (Cloninger, 1987; Fowles, 1980), behavioral approach system (Gray, 1987a, 1990), behavioral engagement system (Depue, Krauss, & Spoont, 1987), or behavioral facilitation system (Depue & Iacono, 1989). The other, dealing with withdrawal or avoidance, is usually called the behavioral inhibition system (Cloninger, 1987; Gray, 1987a,1990),andsometimesawithdrawalsystem(Davidson, 1992a, 1992b).The two systems are generally regarded as relatively independent, with different portions of the brain being most involved in their functioning.

Another literature with a dual-motive theme derives from self-discrepancy theory (Higgins, 1987, 1996). This theory holds that people relate their perceptions of their actual selves to several self-guides, particularly ideals and oughts. Ideals are qualities the person desires to embody: hopes, aspirations, positive wishes for the self. Living up to an ideal means attaining something desired. An ideal is clearly an approach goal.

Oughts, in contrast, are defined by a sense of duty, responsibility, or obligation. An ought is a self that one feels compelled to be, rather than intrinsically desires to be. The ought self is a positive value, in the sense that people try to conform to it. However, living up to an ought also implies acting to avoid a punishment—self-disapproval or the disapproval of others. In our view, oughts are more complex structurally than ideals. Oughts intrinsically imply both an avoidance process and an approach process. Their structure thus resembles what was illustrated earlier in Figure 8.2. Recent work has demonstrated the avoidance aspect of the dynamics behind the ought self (Carver, Lawrence, & Scheier, 1999; Higgins & Tykocinski, 1992).

A similar theme can be seen in the literature of self-determination theory (e.g., Deci & Ryan, 1985, 2000; Ryan & Deci, 2000). That theory focuses on the importance of having a sense of self-determination in one’s actions. Actions that are self-determined are engaged in because they are of intrinsic interest or because they reflect values that are incorporated within the self. Such behavior clearly represents a voluntary approach of positive goal values. In contrast to this is what is termed controlled behavior, meaning that the behavior occurs in response to some sort of coercive force. The coercion can be from outside, or it can be self-coercion. An illustration of the latter is doing something because you feel you have to do it in order not to feel guilty. Such introjected values are very similar to the oughts of self-discrepancy theory, and we have suggested that they similarly involve an avoidance process along with the approach (Carver & Scheier, 1999a, 2000).

Hierarchicality Among Goals

Another key issue in the translation of goals into behavior reflects the obvious fact that some goals are broader in scope thanothers.Howtothinkaboutthedifferenceinbreadthisnot always easy to put your finger on. Sometimes it is a difference in temporal commitment. Sometimes, though, it’s more than that: It’s a difference in the goal’s level of abstraction.

Differentiating Goals by Levels of Abstraction

The notion that goals differ in their level of abstraction is easy to illustrate. You may have the goal of being an honorable person or a self-sufficient person—goals at a fairly high level of abstraction. You may also have the goal of avoiding a person at work who gossips or of making dinner for yourself, which are at a lower level of abstraction. The first set concerns being a particular kind of person, whereas the second set concerns completing a particular kind of action. You can also think of goals that are even more concrete, such as the goal of walking quietly to your office and closing the door without being noticed, or the goal of slicing vegetables into a pan. These goals (which some would call plans or strategies) are closer to specifications of individual acts than was the second set, which consisted more of summary statements about the desired outcomes of intended action patterns.

As you may have noticed, the examples used to illustrate concrete goals relate directly to the examples of abstract goals. We did this to show how abstract goals join with concrete goals in a hierarchy of levels of abstraction. In 1973 William Powers argued that a hierarchical organization of feedback loops underlies the self-regulation of behavior, thus proposing a model of hierarchicality among goals.

His line of thought ran as follows: In a hierarchical organization of feedback systems, the output of a high-level system consists of the resetting of reference values at the next-lower level of abstraction. To put it differently, higher order systems “behave” by providing goals to the systems just below them. The reference values are more concrete and restricted as one moves from higher to lower levels. Each level regulates a quality that contributes to (though not entirely defining) the quality controlled at the next-higher level. Each level monitors input at the level of abstraction of its own functioning, and each level adjusts output to minimize its discrepancies. Structures at various levels presumably handle their concerns simultaneously.

Powers (1973) focused particularly on low levels of abstraction. He said much less about the levels we’re most interested in, though he did suggest labels for several of them. What he called sequences are strings of action that run off directly once cued. Programs, the next-higher level, are activities involving conscious decisions at various points. Programs are webs of sequences with an overall purpose that synthesizes the goals of the constituent sequences. The next level is principles, qualities that are abstracted from (or implemented by) programs. These are the kinds of qualities that are represented by trait labels. Powers gave the name system concepts to the highest level he considered. Goals there include the idealized sense of self, relationship, or group identity.

A simple way of portraying this hierarchy is shown in Figure 8.3. This diagram omits the loops of feedback processes, using lines to indicate only the links among goal values. The lines imply that moving toward a particular lower goal contributes to the attainment of some higher goal (or even several at once). Multiple lines to a given goal indicate that several lower level action qualities can contribute to its attainment. As indicated previously, there are goals to be a particular way and goals to do certain things (and at lower levels, goals to create physical movement).

Although the Powers hierarchy per se has not been studied empirically, research has been done from the perspective of another theory that strongly resembles it—Vallacher and Wegner’s (1985) action identification theory. This model is framed in terms of how people think about their actions, but it also conveys the sense that how people think about their actions is informative about the goals by which they are guiding the actions. People can identify a given action in many different ways, and the identifications can vary in level of abstraction. High-level identifications are abstract, whereas lower level identifications are more concrete. Lowlevel identifications tend to convey a sense of how an activity is done, whereas high-level identifications tend to convey a sense of why.

The Vallacher and Wegner (1985) model does not specify what qualities define various levels but simply assumes that where there is a potential emergent property, there is the potential for differing levels of identification. However, the examples used to illustrate the theory tend to map onto levels of the Powers hierarchy: sequences of acts, programs of actions (with variations of smaller scale and larger scale programs), and principles of being. Thus, work on action identification tends to suggest the reasonableness of these particular levels of abstraction in thinking about behavior.

Step back from this hierarchy for a moment to consider its broader implications. Our present interest is in linking these ideas to the construct of personality. It should be clear that this model provides a way to talk about how the values that are embedded in a person’s personality are manifested in that person’s actions. Values are the source of intentions to take certain patterns of actions, and those programmatic action plans are realized in an extended series of sequences of movement. This view also provides for a mechanism by which the actions themselves take place, which is not typically the case in models of personality.

Multiple Paths to High-Level Goals, Multiple Meanings from Concrete Acts

This hierarchy also has implications for several further issues in thinking about behavior (for more detail see Carver & Scheier, 1998, 1999a). In this view, goals at a given level can often be attained by a variety of means at lower levels. This addresses the fact that people sometimes shift radically the manner in which they try to reach a goal when the goal itself has not changed. This happens commonly when the emergent quality that is the higher order goal is implied in several lower order activities. For example, a person can be helpful by writing a donation check, picking up discards for a recycling center, volunteering at a charity, or holding a door open for someone else.

Just as a given goal can be obtained via multiple pathways, so can a specific act be performed in the service of diverse goals. For example, you could buy someone a gift to make her feel good, to repay a kindness, to put her in your debt, or to satisfy a perceived holiday-season role. Thus, a given act can have strikingly different meanings depending on the purpose it’s intended to serve. This is an important subtheme of this view on behavior: Behavior can be understood only by identifying the goals to which behavior is addressed. This is not always easy to do, either from an observer’s point of view (cf. Read, Druian, & Miller, 1989) or from the actor’s point of view.

Goals and the Self

Another point made by the notion of hierarchical organization concerns the fact that goals are not equivalent in their importance. The higher you go into the organization, the more fundamental to the overriding sense of self are the qualities encountered. Thus, goal qualities at higher levels would appear to be intrinsically more important than those at lower levels.

Goals at a given level are not necessarily equivalent to one another in importance, however. In a hierarchical system there are at least two ways in which importance accrues to a goal. The more directly an action contributes to attainment of some highly valued goal at a more abstract level, the more important is that action. Second, an act that contributes to the attainment of several goals at once is thereby more important than an act that contributes to the attainment of only one goal.

Relative importance of goals returns us again to the concept of self. In contemporary theory the self-concept has several aspects. One is the structure of knowledge about your personal history; another is knowledge about who you are now. Another is the self-guides or images of potential selves that are used to guide movement from the present into the future. As stated earlier, a broad implication of this view is that the self— indeed, personality—consists partly of a person’s goals.

Feedback Loops and Creation of Affect

We turn now to another aspect of human self-regulation: emotion. Here we add a layer of complexity that differs greatly from the complexity represented by hierarchicality. Again, the organizing principle is feedback control. But now the control is over a different quality.

What are feelings, and what makes them exist? Many have analyzed the information that feelings provide and situations in which affect arises (see, e.g., Frijda, 1986; Lazarus, 1991; Ortony, Clore, & Collins, 1988; Roseman, 1984; Scherer & Ekman, 1984). The question we address here is slightly different: What is the internal mechanism by which feelings arise?

Velocity Control

We have suggested that feelings arise within the functioning of another feedback process (Carver & Scheier, 1990). This process operates simultaneously with the behavior-guiding process and in parallel to it. One way to describe this second function is to say that it is checking on how well the behavior loop is doing at reducing its discrepancies. Thus, the input for this second loop is a representation of the rate of discrepancy reduction in the action system over time. We focus first on discrepancy-reducing loops and turn later to enlarging loops.

We find an analogy useful here. Because action implies change between states, think of behavior as analogous to distance. If the action loop deals with distance, and if the affectrelevant loop assesses the progress of the action loop, then the affect loop is dealing with the psychological equivalent of velocity, the first derivative of distance over time. To the extent that the analogy is meaningful, the perceptual input to this loop should be the first derivative over time of the input used by the action loop.

This input does not in itself create affect because a given rate of progress has different affective consequences under different circumstances. As in any feedback system, this input is compared against a reference value (cf. Frijda, 1986, 1988). In this case, the reference is an acceptable or desired rate of behavioral discrepancy reduction. As in other feedback loops, the comparison checks for a deviation from the standard. If there is one, the output function changes.

We suggest that the result of the comparison process in this loop (the error signal generated by its comparator) appears phenomenologically in two forms. One is a nonverbal sense of confidence or doubt (to which we turn later). The other is affect, feeling, a sense of positivity or negativity.

Research Evidence

Because this idea is relatively novel, we should devote some attention to whether any evidence supports it. Initial support came from Hsee and Abelson (1991), who arrived independently at the velocity hypothesis. They conducted two studies of velocity and satisfaction. In one, participants read descriptions of paired hypothetical scenarios and indicated which they would find more satisfying. For example, they chose whether they would be more satisfied if their class standing had gone from the 30th percentile to the 70th over the past 6 weeks, or if it had done so over the past 3 weeks. Given positive outcomes, they preferred improving to a high outcome over a constant high outcome; they preferred a fast velocity over a slow one; and they preferred fast small changes to slower larger changes. When the change was negative (e.g., salaries decreased), they preferred a constant low salary to a salary that started high and fell to the same low level; they preferred slow falls to fast falls; and they preferred large slow falls to small fast falls.

We have since conducted a study that conceptually replicates aspects of these findings but with an event that was personally experienced rather than hypothetical (Lawrence, Carver, & Scheier, in press). We manipulated success feedback on an ambiguous task over an extended period. The patterns of feedback converged such that block 6 was identical for all subjects at 50% correct. Subjects in a neutral condition had 50% on the first and last block, and 50% average across all blocks. Others had positive change in performance, starting poorly and gradually improving. Others had negative change, starting well and gradually worsening. All rated their mood before starting and again after block 6 (which they did not know ended the session). Those whose performances were improving reported mood improvement, whereas those whose performances were deteriorating reported mood deterioration, compared to those with a constant performance.

Another study that appears to bear on this view of affect was reported by Brunstein (1993). It examined subjective well-being among college students over the course of an academic term, as a function of several perceptions, including perception of progress toward goals. Of particular interest at present, perceived progress at each measurement point was strongly correlated with concurrent well-being.

Cruise Control Model

Although the theory may sound complex, the system we have proposed functions much the same as another device that is well known to many people: the cruise control on a car. If you are moving too slowly toward a goal, negative affect arises. You respond to this condition by putting more effort into your action, trying to speed up. If you are going faster than you need to, positive affect arises, and you pull back effort and coast. A car’s cruise control is very similar. You come to a hill, which slows you down. The cruise control responds by feeding the engine more gas to bring the speed back up. If you pass the crest of a hill and roll downhill too fast, the system pulls back on the gas, which eventually drags the speed back down.

This analogy is intriguing because it concerns regulation of the very quality that we believe the affect system is regulating: velocity. It is also intriguing that the analogy incorporates a similar asymmetry in the consequences of deviating from the set point. That is, both in a car’s cruise control and in human behavior, going too slow calls for investing greater effort and resources. Going too fast does not. It calls only for pulling back on resources. That is, the cruise control does not apply the brakes; it just cuts back on the gasoline. In this way it permits the car to coast gradually back to its velocity set point. In the same fashion, people do not respond to positive affect by trying to make it go away, but just by easing off.

Does positive affect actually lead people to withdraw effort? We are not aware of data that bear unambiguously on the question. To do so, a study must assess coasting with respect to the same goal as lies behind the affect. Many studies that might otherwise be seen as relevant to the question created positive affect in one context and assessed its impact on another task (see, e.g., Isen, 2000). The question thus seems to remain open, and to represent an important area for future work (for broader discussion of relevant issues see Carver, in press).

Affect from Discrepancy-Enlarging Loops

Thus far we have restricted ourselves to issues that arise in the context of approach. Now we turn to attempts to avoid a point of comparison, attempts to not-be or not-do: discrepancyenlarging loops.

Our earlier discussion should have made it clear that behavior regarding avoidance goals is just as intelligible as behavior regarding approach goals. We think the same is true of the affective accompaniments to behavior. Our model rests on the idea that positive affect comes when a behavioral system is doing well at what it is organized to do. Thus far we have considered only systems organized to close discrepancies. There seems no obvious reason, however, why the principle should not apply just as well to systems organized to enlarge discrepancies. If the system is doing well at what it is organized to do, positive affect should arise. If it is doing poorly at what it is organized to do, negative affect should arise.

That much would seem to be fully comparable across the two types of systems. But doing well at moving toward an incentive is not exactly the same experience as doing well at moving away from a threat. Both have the potential to induce positive feelings, by doing well. Both also have the potential to induce negative feelings, by doing poorly. Yet the two positives may not be quite the same as each other, nor the negatives quite the same as each other.

Our view of this difference derives partly from the insights of Higgins and his colleagues (Higgins, 1987, 1996). Following their lead, we suggest that the affect dimension relating to discrepancy reduction is (in its purest form) the dimension that runs from depression to elation (Figure 8.4). The affect that relates to discrepancy enlargement is (in its purest form) the dimension from anxiety to relief or contentment. As Higgins and his colleagues have noted, dejection-related and agitation-related affect may take several forms, but these two dimensions capture the core qualities behind those two classes of affect. Similarly, Roseman (1984) has argued that joy and sadness are related to appetitive (moving-toward) motives, whereas relief and distress are related to aversive (moving-away-from) motives.

Merging Affect and Action

Theories about emotion typically emphasize the idea that emotion is related to action. How do affect and action relate in this model? We see the regulation provided by these systems as forming a two-layered array, with both simultaneously at work (Carver & Scheier, 1998, 1999a, 1999b). The two layers are analogous to position and velocity controls in a two-layered engineering control system (e.g., Clark, 1996). Such a two-layered system in engineering has the quality of responding both quickly and accurately (without undue oscillation). There is reason to believe that the simultaneous functioning of the two layers has the same broad consequence for human behavior.

Another way of addressing the relation between affect and action is to ask about the nature of the output of the affect loop. Earlier we described affect as reflecting the error signal of a loop that has as input a perception of rate of progress. The resulting output thus must be an adjustment in rate of progress. This output therefore has a direct link to behavior because it means changing its pace.

What does it mean to adjust the rate of progress? In some cases it means literally changing velocity. If you are behind, go faster. Some adjustments are less straightforward. The rates of many behaviors in which personality–social psychologists are interested are not defined in terms of literal pace of motion. Rather, they are defined in terms of choices among actions, even potential programs of action. For example, increasing your rate of progress on a reading assignment may mean choosing to spend a weekend working rather than playing. Increasing your rate of manifestation of kindness means choosing to perform an action that reflects that value. Thus, adjustment in rate must often be translated into other terms, such as concentration or reallocation of time and effort.

Despite this complexity in implementing changes in rate, it should be apparent from this description that the action system and the velocity system are presumed to work in concert with one another. Both are involved in the flow of action. They influence different aspects of the action, but both are always involved. Thus, this view incorporates clear links between behavior and affect.

Comparison with Biological Models of Bases of Affect

It is useful to compare this model with the group of biologically focused theories mentioned earlier in the paper. As indicated earlier, those theories assume that two separate systems regulate approach and avoidance behavior. Many assume further that the two systems also underlie affect. Given cues of impending reward, the activity of the approach system creates positive feelings. Given cues of impending punishment, the avoidance system creates feelings of anxiety.

Data from a variety of sources fit this picture. Of particular relevance is work by Davidson and collaborators involving electroencephalography (EEG) recordings assessing changes in cortical activation in response to affective inducing stimuli. Among the findings are these: Subjects exposed to films inducing fear and disgust (Davidson, Ekman, Saron, Senulis, & Friesen, 1990) and confronted with possible punishment (Sobotka, Davidson, & Senulis, 1992) show elevations in right frontal activation. In contrast, subjects with a chance to obtain reward (Sobota et al., 1992), subjects presented with positive emotional adjectives (Cacioppo & Petty, 1980), and smiling 10-month olds viewing their approaching mothers (Fox & Davidson, 1988) show elevations in left frontal activation. From findings such as these, Davidson (1992a, 1992b) concluded that neural substrates for approach and withdrawal systems (and thus positive and negative affect) are located in the left and right frontal areas of the cortex, respectively.

Thus far the logic of the biological models resembles the logic of our model. At this point, however, there is a divergence. The key question is what regulatory processes are involved in—and what affects result from—failure to attain reward and failure to receive punishment. Gray (1987b, 1990) holds that the avoidance system is engaged by cues of punishment and cues of frustrative nonreward. It thus is responsible for negative feelings in response to either of these types of cues. Similarly, Gray holds that the approach system is engaged by cues of reward or cues of escape from (or avoidance of) punishment. It thus is responsible for positive feelings in response to either of these types of cues. In his view, then, each system creates affect of one hedonic tone (positive in one case, negative in the other), regardless of its source. This view is consistent with a picture of two unipolar affective dimensions, each linked to a distinct behavioral system. Others have taken a similar position (see Cacioppo, Gardner, & Berntson, 1999; Lang, 1995; Lang, Bradley, & Cuthbert, 1990; Watson, Wiese, Vaidya, & Tellegen, 1999).

Our position is different. We argue that both approach and avoidance systems can create affects of both hedonic tones because affect is a product of doing well or doing poorly. We think that the frustration and eventual depression that result from failure to attain desired goals involve the approach system (for similar predictions see Clark, Watson, & Mineka, 1994, p. 107; Cloninger, 1988, p. 103; Henriques & Davidson, 1991). A parallel line of reasoning suggests that relief, contentment, tranquility, and serenity relate to the avoidance system rather than to the approach system (see Carver, 2001).

Less information exists about the bases of these affects than about anxiety and happiness. Consider first relieftranquility. We know of two sources of evidence, both somewhat indirect. The first is a study in which people worked at a laboratory task and experienced either goal attainment or lack of attainment (Higgins, Shah, & Friedman, 1997, Study 4). Participants first were given either an approach orientation to the task (to try to attain success) or an avoidance orientation (to try to avoid failing). After the task outcome (which was manipulated), several feeling qualities were assessed. Among persons given an avoidance orientation, success caused an elevation in calmness, and failure caused an elevation in anxiety. These effects on calmness and anxiety did not occur, however, among those who had an approach orientation. This pattern suggests that calmness is linked to doing well at avoidance, rather than doing well at approach.

Another source is data reported many years ago by Watson and Tellegen (1985). In their analysis of multiple samples of mood data, they reported “calm” to be one of the 10 best markers (inversely) of negative affect (which was defined mostly by anxiety) in the majority of the data sets they examined. In contrast, “calm” never emerged as one of the top markers of positive affect in those data sets. This suggests that these feelings are linked to the functioning of a system of avoidance.

The same sources also provide information on the momentary experience of sadness. In the study by Higgins et al. (1997), failure elevated sadness and success elevated cheerfulness among persons with an approach orientation. These effects did not occur, however, among participants who had an avoidance orientation. The pattern suggests that sadness is linked to doing poorly at approach, rather than doing poorly at avoidance. Similarly, Watson and Tellegen (1985) reported “sad” to be one of the 10 best markers (inversely) of the factor that they called positive affect in the majority of the data sets they examined. In contrast, “sad” never emerged as one of the top markers of negative affect in those data sets. This pattern suggests that sad feelings are linked to the functioning of a system of approach.

This issue clearly represents an important difference among theoretical viewpoints (Carver, 2001). Just as clearly, it is not yet resolved. It seems likely that it will receive more attention in the near future.

Responding to Adversity: Persistence and Giving Up

In describing the genesis of affect, we suggested that one process yields two subjective experiences as readouts: affect and a sense of confidence versus doubt. We turn now to confidence and doubt—expectancies for the immediate future. We focus here on the behavioral and cognitive manifestations of the sense of confidence or doubt.

One likely consequence of momentary doubt is a search for more information. We have often suggested that when people experience adversity in trying to move toward goals, they periodically interrupt efforts in order to assess in a more deliberative way the likelihood of a successful outcome (e.g., Carver & Scheier, 1981, 1990, 1998). In effect, people suspend the behavioral stream, step outside it, and evaluate in a more deliberated way. This may happen once or often. It may be brief, or it may take a long time. In this assessment people presumably depend heavily on memories of prior outcomes in similar situations. They may also consider such things as additional resources they might bring to bear, alternative approaches that might be taken, and social comparison information (Wills, 1981; Wood, 1989).

These thoughts sometimes influence the expectancies that people hold. When people retrieve “chronic” expectancies from memory, the information already is expectancies— summaries of the products of previous behavior. In some cases, however, the process is more complex. People bring to mind possibilities for changing the situation and evaluate their consequences. This is often done by briefly playing the possibility through mentally as a behavioral scenario (cf. Taylor & Pham, 1996). Doing so can lead to conclusions that influence expectancies (“If I try doing it this way instead of that way, it should work better” or “This is the only thing I can see to do, and it will just make the situation worse”).

It seems reasonable that this mental simulation engages the same mechanism as handles the affect-creation process during actual overt behavior. When your progress is temporarily stalled, playing through a confident and optimistic scenario yields a higher rate of progress than is currently being experienced. The affect loop thus yields a more optimistic outcome assessment than is being derived from current action. If the scenario is negative and hopeless, it indicates a further reduction in progress, and the loop yields further doubt.

Behavioral Manifestations

Whether stemming from the immediate flow of experience or from a more thorough introspection, people’s expectancies are reflected in their behavior. If people expect a successful outcome, they continue exerting effort toward the goal. If doubts are strong enough, the result is an impetus to disengage from effort, and potentially from the goal itself (Carver & Scheier, 1981, 1990, 1998, 1999a; see also Klinger, 1975; Kukla, 1972; Wortman & Brehm, 1975). This theme—divergence in behavioral response as a function of expectancies—is an important one, applying to a surprisingly broad range of literatures (see Carver & Scheier, 1998, chap. 11).

Sometimes the disengagement that follows from doubt is overt, but sometimes disengagement takes the form of mental disengagement—off-task thinking, daydreaming, and so on. Although this can sometimes be useful (self-distraction from a feared stimulus may allow anxiety to abate), it can also create problems. Under time pressure, mental disengagement can impair performance, as time is spent on task-irrelevant thoughts. Consistent with this, interactions between selffocus and expectancies have been shown for measures of performance (Carver, Peterson, Follansbee, & Scheier, 1983; Carver & Scheier, 1982).

Often, mental disengagement cannot be sustained, as situational cues force the person to reconfront the problematic goal. In such cases, the result is a phenomenology of repetitive negative rumination, which often focuses on self-doubt and perceptions of inadequacy. This cycle is both unpleasant and performance-impairing.

Is Disengagement Good or Bad?

Is the disengagement tendency good or bad? Both and neither. On the one hand, disengagement (at some level, at least) is an absolute necessity. Disengagement is a natural and indispensable part of self-regulation (cf. Klinger, 1975). If people are ever to turn away from unattainable goals, to back out of blind alleys, they must be able to disengage, to give up and start over somewhere else.

The importance of disengagement is particularly obvious with regard to concrete, low-level goals: People must be able to remove themselves from literal blind alleys and wrong streets, give up plans that have become disrupted by unexpected events, even spend the night in the wrong city if they miss the last plane home. Disengagement is also important, however, with regard to more abstract and higher level goals. A vast literature attests to the importance of disengaging and moving on with life after the loss of close relationships (e.g., Orbuch, 1992; Stroebe, Stroebe, & Hansson, 1993; Weiss, 1988).

People sometimes must be willing to give up even values that are deeply embedded in the self if those values create too much conflict and distress in their lives.

However, the choice between continued effort and giving up presents opportunities for things to go awry. It is possible to stop trying too soon, thereby creating potentially serious problems for oneself (Carver & Scheier, 1998). It is also possible to hold on to goals too long, thereby preventing oneself from taking adaptive steps toward new goals. But both continued effort and giving up are necessary parts of the experience of adaptive self-regulation. Each plays an important role in the flow of behavior.

Hierarchicality and Importance Can Impede Disengagement

Disengagement is sometimes precluded by situational constraints. However, a broader aspect of this problem stems from the idea that behavior is hierarchically organized, with goals increasingly important higher in the hierarchy, and thus harder to disengage from.

Presumably, disengaging from concrete values is often easy. Lower order goals vary, however, in how closely they link to values at a higher level, and thus in how important they are. To disengage from low-level goals that are tightly linked to higher level goals causes discrepancy enlargement at the higher level. These higher order qualities are important, even central to one’s life. One cannot disengage from them, disregard them, or tolerate large discrepancies between them and current reality without reorganizing one’s value system (Greenwald, 1980; Kelly, 1955; McIntosh & Martin, 1992; Millar, Tesser, & Millar, 1988). In such a case, disengagement from even very concrete behavioral goals can be quite difficult.

Now recall again the affective consequences of being in this situation. The desire to disengage was prompted by unfavorable expectancies. These expectancies are paralleled by negative affect. In this situation, then, the person experiences negative feelings (because of an inability to make progress toward the goal) and is unable to do anything about the feelings (because of an inability to give up). The person simply stews in the feelings that arise from irreconcilable discrepancies. This kind of situation—commitment to unattainable goals—seems a sure prescription for distress.

Watersheds, Disjunctions, and Bifurcations Among Responses

An issue that bears some further mention is the divergence in the model of the behavioral and cognitive responses to favorable versus unfavorable expectancies. We have long argued for a psychological watershed among responses to adversity (Carver & Scheier, 1981). One set of responses consists of continued comparisons between present state and goal, and continued efforts. The other set consists of disengagement from comparisons and quitting. Just as rainwater falling on a mountain ridge ultimately flows to one side of the ridge or the other, so do behaviors ultimately flow to one of these sets or the other.

Our initial reason for taking this position stemmed largely from several demonstrations that self-focused attention creates diverging effects on information seeking and behavior as afunctionofexpectanciesofsuccess.Wearenottheonlyones to have emphasized a disjunction among responses, however. Anumber of others have done so, for reasons of their own.

Kukla (1972) proposed an early model that emphasized the idea of a disjunction in behavior. Another such model is the reactance–helplessness integration of Wortman and Brehm (1975): the argument that threats to control produce attempts to regain control and that perceptions of loss of control produce helplessness. Brehm and his collaborators (Brehm & Self, 1989; Wright & Brehm, 1989) developed an approach to task engagement that resembles that of Kukla (1972), but their way of approaching the description of the problem is somewhat different. Not all theories about persistence and giving up yield this dichotomy among responses. The fact that some do, however, is interesting. It becomes more so a bit later on.

Scaling Back Aspirations and Recalibration of the Affect System

The preceding sections dealt with the creation of affect and confidence and the concomitant effects on behavior. By implication, the time frames under discussion were quite narrow. In this section we broaden our view somewhat and indicate an importantwayinwhichreferencevalueschangeacrosslonger periods of time. These particular changes are changes in the stringency of the goals being sought after. We consider this issue both with respect to the reference values underlying the creation of affect and with respect to the goals of behavior.

Shifts in Velocity Standards

Reference values used by the affect system presumably can shift through time and experience. That is, as people accumulate experience in a given domain, adjustments can occur in the pacing that they expect and demand of themselves. There is a recentering of the system around the past experience, which occurs via shifts in the reference value (Carver & Scheier, 2000).

Consider first upward adjustments. As an example, a person who gains work-related skills often undertakes greater challenges, requiring quicker handling of action units. Upward adjustment of the rate standard means that the person now will be satisfied only with faster performance. Such a shift has the side effect of decreasing the potential for positive affect and increasing the potential for negative affect because there now is more room to fail to reach the rate standard and less room to exceed it. Recall, however, that the shift was induced by a gain in skills. The change in skill tends to counter the shift in regions of potential success and failure. Thus, the likelihood of negative affect (vs. positive affect or no affect) remains fairly constant.

Now consider a downward adjustment. For example, a person whose health is failing may find that it takes longer to get things done than it used to. This person will gradually come to use less stringent rate standards. A lower pace will then begin to be more satisfying. One consequence of this downward shift of standard is to increase the potential for experiencing positive affect and to decrease the potential for negative affect because there now is less room for failing to reach the rate standard and more room for exceeding it. The failing health, however, tends to counter the shift in regions of potential success and failure. Again, then, the net result is that the likelihood of negative affect (vs. positive and neutral) remains fairly constant.

Mechanism of Shift

Such changes in comparison value do not happen quickly or abruptly. Shifting the reference value downward is not people’s first response when they have trouble maintaining a demanding pace. First, they try harder to keep up. Only more gradually, if they continue to lag behind, does the rate-related standard shift to accommodate. Similarly, the immediate response when people’s pace exceeds the standard is not an upward shift in reference value. The more typical response is to coast for a while. Only when the overshoot is frequent does the standard shift upward.

We believe that adjustments in these standards occur automatically and involuntarily, but slowly. Such adjustments themselves appear to reflect a self-corrective feedback process (Figure 8.5). This feedback process is slower than the ones focused on thus far, involving a very gradually accumulating shift. It resembles what Solomon (1980; Solomon & Corbit, 1974) described as the long-term consequences of an opponent process system (see also Helson, 1964, regarding the concept of adaptation level).

As an illustration, assume for the moment that a signal to adjust the standard occurred every time there was a signal to change output, but that the former was much weaker than the latter—say, 5% of the latter. If so, it would take a fairly long time for the standard to change. Indeed, as long as the person deviated from the standard in both directions (under and over) with comparable frequency, the standard would never change noticeably, even over an extended period. Only with repeated deviation in the same direction could there be an appreciable effect on the standard.

This view has an interesting implication for affective experience across an extended period. Such shifts in reference value (and the resultant effects on affect) would imply a mechanism within the organism that prevents both the toofrequent occurrence of positive feeling and the too-frequent occurrence of negative feeling. That is, the (bidirectional) shifting of the rate criterion over time would tend to control pacing such that affect continues to vary in both directions around neutral, roughly as before. The person thus would experience more or less the same range of variation in affective experience over long times and changing circumstances (see Myers & Diener, 1995, for evidence of this). The organization would function as a gyroscope serving to keep people floating along within the framework of the affective reality with which they are familiar. It would provide for a continuous recalibration of the feeling system across changes in situation. It would repeatedly shift the balance point of a psychic teeter-totter so that rocking both up and down remains possible.

Scaling Back on Behavioral Goals

The principle of gradual adjustment of a standard also operates at the level of behavioral goals (Carver & Scheier, 1981, 1998). Sometimes progress is going poorly, expectancies of success are dim, and the person wants to quit. Rather than quit altogether, the person trades this goal for a less demanding one. This is a kind of limited disengagement in the sense that the person is giving up the first goal while adopting the lesser one. However, this limited disengagement keeps the person engaged in activity in the domain he or she had wanted to quit. By scaling back the goal—giving up in a small way— the person keeps trying to move ahead—thus not giving up, in a larger way.

Small-scale disengagement occurs often in the context of moving forward in broader ways. A particularly poignant example comes from research on couples in which one partner is becoming ill and dying from AIDS (Moskowitz, Folkman, Collette, & Vittinghoff, 1996). Some healthy participants initially had the goal of overcoming their partner’s illness and continuing active lives together. As the illness progressed and it became apparent that that goal would not be met, it was not uncommon for the healthy partners to scale back their aspirations. Now the goal was, for example, to do more limited activities during the course of a day. Choosing a more limited and manageable goal ensures that it will be possible to move toward it successfully. The result was that even in those difficult circumstances the person experienced more success than would otherwise have been the case and remained engaged behaviorally with efforts to move forward.

How does the scaling back of goals within a domain occur? We believe that the answer is the same as in the case of affect: If the loop’s output function is inadequate at moving the input toward the standard, a second (slower-acting) process moves the standard toward the input. The scaling back of behavioral goals thus would involve the same structural elements as are involved in the recalibration of the affect system.

Conflict and Restraint

In thinking about the self-regulation of behavior, another set of issues to be considered concerns the existence of conflict. Conflict arises whenever two incompatible goals are held simultaneously and both are salient (see also Carver & Scheier, 1998, 1999b). It sometimes is possible to move toward two goals simultaneously, but sometimes moving toward one interferes with one’s ability to move toward the other. For example, the woman who wants to develop her career and also spend time with her family faces a conflict imposed by the limited number of hours in the day and days in the week (Figure 8.6). The effort to attain one (e.g., further the career by working extra hours) can interfere with efforts to attain the other (by removing the time available for family activities).

Given this structure, the experience of conflict naturally produces negative feelings, as movement toward one of the goals is impeded. If movement toward the active goal is rapid (relative to the reference velocity) as movement toward the other goal is stifled, the person may have mixed feelings, feelings relating to each of the two goal values. It is no surprise that people typically try to balance their conflicting desires so that both goals are partly attained. It is also no surprise that this strategy often feels unsatisfying, as the person “almost” keeps up with goals in both domains but keeps up fully with neither of them.

Often there is no structural basis for viewing one goal as intrinsically more valuable than the other (as in Figure 8.6). Sometimes, however, one goal has a kind of primacy because it is reflected in an explicitly formulated intention to override efforts to move toward the other goal. Sometimes the tendencies involved are mental; sometimes they are behavioral. Often, the attempt to override works for a while (sometimes a long while), but sometimes it fails.

Ironic Processes in Mental Control

One literature bearing on this theme was developed by Wegner (e.g., 1994) and his colleagues. The study that began this work was simple. Some people were told not to think of a white bear for 5 minutes. Then they were told to think about the bear. When the thought was permitted, it came more frequently than it did for people who had not had to suppress the thought first. Something about trying not to think of the bear seemed to create pressure to think of it.

This study was followed by others. Most of this research looked not at rebounds, but at what goes on during people’s attempts to control their thoughts. The data consistently indicate that an instruction to exert mental control yields better control if the person has no other demands. If something else is going on, however (e.g., if the person is trying to remember a 9-digit number), the instruction backfires, and people tend to do the opposite of what they are trying to do.

Wegner (1994) interprets this as follows: Trying to suppress a thought engages two processes.An intentional process tries to suppress. An ironic monitoring process looks for the occurrence of whatever is being suppressed. If it finds it, it increases the effort of the first mechanism. The ironic monitor is sensitive, but it is automatic and does not require much in the way of mental resources. The intentional process requires more resources.Thus, any reduction in mental resources (e.g., being distracted by a second thought or task) disrupts the intentional process more than it disrupts the ironic monitor. The monitor, searching for lapses, in effect invites those lapses to occur.

This theory also applies to the opposite pattern—attempts to concentrate. In this case, the intentional process concentrates, and the ironic process looks for the occurrence of distractions. As in the first case, if the person’s mental resources are stretched thin, the ironic process seems to invite the undesired thought into consciousness. In this case, the thought is a distraction.

This research indicates that trying hard to do something (or suppress something) gets much harder when your mental resources are stretched thin. Not only does it get harder, but you may even begin to do the opposite of what you are trying to do.

Lapses in Self-Control

Another important literature bearing on this set of issues concerns what Baumeister and Heatherton (1996) termed selfregulatory failure, which we will term lapse in self-control. The potential for this kind of event arises when someone has both the desire to do something (e.g., overindulge in food or drink) and also the desire to restrain that impulse. Self-control of this sort is often especially hard, and sometimes the restrained impulse breaks free.

Considerbingeeatingasanexample.Thebingeeaterwants to eat but also wants to restrain that desire. If self-control lapses, the person stops trying to restrain the desire to eat, lets himself or herself go, and binges.

In characterizing the decision to quit trying to restrain, Baumeister and Heatherton noted that restraint is hard work and that mental fatigue plays a role; however, giving up the restraint attempt rarely requires that the person reach a state of total exhaustion. Rather, there is a point where the person has had enough and stops trying to control the impulse. We have suggested that confidence about resisting the impulse plays a role in whether the person stops trying (Carver & Scheier, 1998). The confident person continues the struggle to restrain. The person whose confidence has sagged is more likely to give up.

Muraven, Tice, and Baumeister (1998) have extended this line of thought to argue that self-control is a resource that not only is limited but also can become depleted by extended self-control efforts. When the resource is depleted, the person becomes vulnerable to a failure of self-control. This view also suggests that there is a shared pool of self-control resources, so that exhausting the resource with one kind of self-control (e.g., concentrating very hard for many hours on a writing assignment) can leave the person vulnerable to a lapse in a different domain (e.g., eating restraint).

It seems worthwhile to compare the cases considered in this section (lapses in self-control) with those described just earlier (mental control). Both sections dealt with efforts at self-control. In many ways the situations are structurally quite similar. Each is an attempt to override one process by another, which falters when mental resources are depleted. There even is a resemblance between the “overdoing” quality in the previously restrained behavior in Baumeister and Heatherton’s cases and the rebound quality in Wegner’s research.

One difference is that the cases emphasized by Baumeister and Heatherton explicitly involve desires that direct the person in opposing directions. In most cases studied by Wegner, there is no obvious reason why the suppressed thought (or the distractor) would be desirable. This difference between cases seems far from trivial. Yet the similarities in the findings in the two literatures are striking enough to warrant further thought about how the literatures are related.

Dynamic Systems and Self-Regulation

Recent years have seen the emergence in the psychological literature of new (or at least newly prominent) ideas about how to conceptualize natural systems. Several labels attach to these ideas: chaos, dynamic systems theory, complexity, catastrophe theory.Anumber of introductions to this body of thought have been written, some of which include applications to psychology (e.g., Brown, 1995; Gleick, 1987; Thelen & Smith, 1994; Vallacher & Nowak, 1994, 1997; Waldrop, 1992). These themes are of growing interest in several areas of psychology, including personality–social psychology. In this section we sketch some of the themes that are central to this way of thinking.

Nonlinearity

Dynamic systems theory holds that the behavior of a system reflects all the forces operating on (and within) it. It also emphasizes that the behavior of a complex system over any period but a brief one is very hard to predict. One reason for this is that the system’s behavior may be influenced by these forces in nonlinear ways. Thus, the behavior of the system— even though highly determined—can appear random.

Many people are used to thinking of relationships between variables as linear. But some relationships clearly are not. Familiar examples of nonlinear relationships are step functions (ice turning to water and water turning to steam as temperature increases), threshold functions, and floor and ceiling effects. Other examples of nonlinearity are interactions. In an interaction the effect of one predictor on the outcome differs as a function of the level of a second predictor. Thus the effect of the first predictor on the outcome is not linear.

Many personality psychologists think in terms of interactions much of the time. Threshold effects and interactions are nonlinearities that most of us take for granted, though perhaps not labeling them as such. Looking intentionally for nonlinearities, however, reveals others. For example, many psychologists now think that many developmental changes are dynamic rather than linear (Goldin-Meadow & Alibali, 1995; Ruble, 1994; Siegler & Jenkins, 1989; Thelen, 1992, 1995; van der Maas & Molenaar, 1992).

Sensitive Dependence on Initial Conditions

Nonlinearity is one reason for the difficulty in predicting complex systems. Two more reasons why prediction over any but the short term is difficult is that you never know all the influences on a system, and the ones you do know you never know with total precision. What you think is going on may not be quite what’s going on. That difference, even if it is small, can be very important.

This theme is identified with the phrase sensitive dependence on initial conditions. This means that a very small difference between two states of affairs can lead to divergence and ultimately to an absence of relation between the paths that are taken later on. The idea is (partly) that a small initial difference between systems causes a difference in what they encounter next, which produces slightly different outcomes (Lorenz, 1963). Through repeated iterations, the systems diverge, eventually moving on very different pathways. After a surprisingly brief period they no longer have any noticeable relation to one another.

How does the notion of sensitive dependence on initial conditions relate to human behavior? Most generally, it suggests that a person’s behavior will be hard to predict over a long period except in general terms. For example, although you might be confident that Mel usually eats lunch, you will not be able to predict as well what time, where, or what he will eat on the second Friday of next month. This does not mean Mel’s behavior is truly random or unlawful (cf. Epstein, 1979). It just means that small differences between the influences you think are affecting him and the influences that actually exist will ruin the predictability of moment-to-moment behavior.

This principle also holds for prediction of your own behavior. People apparently do not plan very far into the future most of the time (Anderson, 1990, pp. 203–205), even experts (Gobet & Simon, 1996). People seem to have goals in which the general form is sketched out but only a few steps toward it have been planned. Even attempts at relatively thorough planning appear to be recursive and “opportunistic,” changing— sometimes drastically—when new information becomes known (Hayes-Roth & Hayes-Roth, 1979).

The notion of sensitive dependence on initial conditions fits these tendencies. It is pointless (and maybe even counterproductive) to plan too far ahead too fully (cf. Kirschenbaum, 1985), because chaotic forces in play (forces that are hard to predict because of nonlinearities and sensitive dependence) can render much of the planning irrelevant. Thus, it makes sense to plan in general terms, chart a few steps, get there, reassess, and plan the next bits. This seems a perfect illustration of how people implicitly take chaos into account in their own lives.

Phase Space, Attractors, and Repellers

Another set of concepts important to dynamic-systems thinking are variations on the terms phase space and attractor (Brown, 1995; Vallacher & Nowak, 1997). A phase diagram is a depiction of the behavior of a system over time. Its states are plotted along two (sometimes three) axes, with time displayed as the progression of the line of the plot, rather than on an axis of its own. Aphase space is the array of states that the system occupies across a period of time. As the system changes states from one moment to the next, it traces a trajectory within its phase space—a path of the successive states it occupies across that period.

Phase spaces often contain regions called attractors. Attractors are areas that the system approaches, occupies, or tends toward more frequently than other areas. Attractors exertametaphoricalgravitationalpullonthesystem,bringing the system into proximity to them. Each attractor has a basin, which is the attractor’s region of attraction. Trajectories that enter the basin tend to move toward that attractor (Brown, 1995).

There are several kinds of attractors, some very simple, others more complex. In a point attractor, all trajectories converge onto some point in phase space, no matter where they begin (e.g., body temperature). Of greater interest are chaotic attractors. The pattern to which this term refers is an irregular and unpredictable movement around two or more attraction points. An example is the Lorenz attractor (Figure 8.7), named for the man who first plotted it (Lorenz, 1963). It has two attraction zones. Plotting the behavior of this system over time yields a tendency to loop around both attractors, but to do so unpredictably. Shifts from one basin to the other seem random.

The behavior of this system displays sensitivity to initial conditions. A small change in starting point changes the specific path of motion entirely. The general tendencies remain the same—that is, the revolving around both attractors. But details such as the number of revolutions around one before deflection to the other form an entirely different pattern. The trajectory over many iterations shows this same sensitivity to small differences. As the system continues, it often nearly repeats itself but never quite does, and what seem nearly identical paths sometimes diverge abruptly, with one path leading to one attractor and the adjacent path leading to the other.

A phase space also contains regions called repellers, regions that are hardly ever occupied. Indeed, these regions seem to be actively avoided. That is, wandering into the basin of a repeller leads to a rapid escape from that region of phase space.

Another Way of Picturing Attractors

The phase-space diagram gives a vivid visual sense of what an attractor looks and acts like. Another common depiction of attractors is shown in Figure 8.8. In this view, attractor basins are basins or valleys in a surface (more technically called local minima). Repellers are ridges. This view assumes a metaphoric “gravitational” drift downward in the diagram, but other forces are presumed to be operative in all directions. For simplicity, this portrayal usually is done in two dimensions (sometimes 3), but keep in mind that the diagram often assumes the merging of a large number of dimensions into the horizontal axis.

The behavior of the system at a given moment is represented as a ball on the surface. If the ball is in a valley (points 1 and 2 in panel Aof Figure 8.8), it is in an attractor basin and will tend to stay there unless disturbed. If it is on a hill (between 1 and 2), any slight movement in either direction will cause it to escape its current location and move to an adjacent attractor.

One strength of this portrayal is that it does a good job of creating a sense of how attractors vary in robustness. The breadth of a basin indicates the diversity of trajectories in phase space that are drawn into it. The broader is the basin (B-1 in Figure 8.8), the more trajectories are drawn in. The narrower the basin (B-2), the closer the ball has to come to its focal point to be drawn to it. The steepness of the valley indicates how abruptly a trajectory is drawn into it. The steeper the slope of the wall (B-2), the more sudden is the entry of a system that encounters that basin.

The depth of the valley indicates how firmly entrenched the system is, once drawn into the attractor. Figure 8.8, panel C, represents a system of attractors with fairly low stability (the valleys are shallow). One attractor represents a stable situation (valley 1), whereas the others are less so. It will take a lot more “energy” to free the ball from valley 1 than from the others.

There is a sense in which both breadth and depth suggest that a goal is important. Breadth does so because the system is drawn to the attractor from widely divergent trajectories. Depth does so because the system that has been drawn into the basin tends to stay there.

A weakness of this picture, compared to a phase-space portrait, is that it is not as good at giving a sense of the erratic motion from one attractor to another in a multiple-attractor system. You can regain some of that sense of erratic shifting, however, if you think of the surface in Figure 8.8 as a tambourine being continuously shaken (Figure 8.8, panel D). Even a little shaking causes the ball to bounce around in its well and may jostle it from one well to another, particularly if the attractors are not highly stable. An alternative would be to think of the ball as a jumping bean. These two characterizations would be analogous to jostling from situational influences and jostling from internal dynamics, respectively.

Goals as Attractors

The themes of dynamic systems thinking outlined here have had several applications in personality–social and even clinical psychology (Hayes & Strauss, 1998; Mahoney, 1991; Nowak & Vallacher, 1998; Vallacher & Nowak, 1997). Perhaps the easiest application of the attractor concept to self-regulatory models is to link it with the goal concept. Indeed, alert readers will have noticed that we used the same metaphor—gravity and antigravity—in describing both the goal construct at the beginning of the paper and in describing the attractor concept just earlier.

As we said at the beginning of the paper, goals are points around which behavior is regulated. People spend much of their time doing things that keep their behavior in close proximity to their goals. It seems reasonable to suggest, then, that a goal represents a kind of attractor. Further, if a goal is an attractor, it seems reasonable that an antigoal would represent a repeller.

This functional similarity between the goal construct and the attractor basin is very interesting. However, the similarity exists only with respect to the end product—that is, maintaining proximity to a value (or remaining distant from a value). The two views make radically different assumptions about the presence or absence of structure underlying the functions. The feedback model assumes a structure underlying and supporting the process, whereas the dynamic systems model does not necessarily incorporate such an assumption.

Connectionism

A related set of questions about the role of central control processes is raised by the literature of connectionism. Connectionist models simulate thought processes in networks of artificial units in which “processing” consists of passing activation among the units. As in neurons, the signal can be excitatory or inhibitory. Energy passes in only one direction (though some networks have feedback links). Processing proceeds entirely by the spread of activation—there is no higher order executive to direct traffic. In a distributed connectionist network, knowledge is not represented centrally, as nodes of information. Rather, knowledge is represented in terms of the pattern of activation of the network as a whole (Smith, 1996).

In networks with feedback relations, once the system receives input, the pattern of weights and activations is updated repeatedly across many cycles. Thus, modifications or updates are made iteratively throughout the network, both with respect to activation in each node and the weighting functions. Gradually, the various values asymptote, and the system “settles” into a configuration. The settling reflects the least amount of overall error the system has been able to create, given its starting inputs and weights.

Multiple Constraint Satisfaction

A useful way to think about this process is that the system simultaneously satisfies multiple constraints that the elements create on each other (Thagard, 1989; see also Kelso, 1995). For example, two mutually inhibitory nodes cannot both be highly active at the same time. Thus they constrain one another. Constraints among multiple nodes are settled out during the repeated updating of activation levels.

This idea of multiple constraint satisfaction is now having a substantial impact on how people in social psychology think about a variety of topics (Kunda & Thagard, 1996; Read, Vanman, & Miller, 1997; Schultz & Lepper, 1996). It is an idea that has a great deal of intuitive appeal. It captures well the introspective sense that people come to conclusions and decisions not by weighing the evidence, exactly, but rather by letting the evidence sort itself until it reaches a degree of internal consistency. The conclusion then pops into mind.

Another term that goes along with this picture is selforganization (e.g., Prigogine & Stengers, 1984). The idea behind this label is that multiple causal forces which have no intrinsic relation to each other can cause the spontaneous emergence of some property of the system as a whole that does not otherwise exist. The term is used to describe emergent qualities in a variety of scientific disciplines. A number of people have begun to invoke it as a basis for emergent properties in dynamic systems (Nowak & Vallacher, 1998; Prigogine & Stengers, 1984).

Self-Organization and Self-Regulation

Some would argue that models of self-organization in dynamic systems represent a serious challenge to the viability of the type of self-regulatory model with which we began. That is, it might be asserted that behavior only seems to be selfregulated—that behavior instead self-organizes from among surrounding forces, like foam appearing on roiling surf.

Do feedback processes actually reflect self-organization— a haphazard falling together of disparate forces? Or are there structures in the nervous system (and elsewhere) in living systems that carry out true feedback functions? In considering the relation between the two sets of ideas, it is of interest that MacKay (1956) anticipated the principle of self-organization many years ago when he described a system of feedback processesthatcouldevolveitsowngoals(seealsoBeer,1995; Maes & Brooks, 1990). Thus, MacKay found the principle of self-organization to be useful, but he found it useful explicitly within the framework of a self-regulatory model.

Our view is, similarly, that the concepts of attractors and trajectories within phase space complement the idea that behavior is guided by feedback processes but do not replace it (Carver & Scheier, in press). There do appear to be times and circumstances in which forces converge—unplanned—and induce acts to occur that were not intended beforehand. However, there also seem to be clear instances of intentionality in behavior and its management.

It is of interest in this regard that contemporary cognitive psychologists often assume the existence of both bottom-up organizational tendencies and top-down directive tendencies (see, e.g., Holyoak & Spellman, 1993; Shastri & Ajjanagadde, 1993; Sloman, 1996; Smolensky, 1988). That view would seem to fit a picture in which self-organization of action can occur, but where actions can also be planned and executed systematically, from the top down. Similar two-mode models of regulation have also appeared in several literatures in personality-social psychology (Chaiken & Trope, 1999). In short, there seems to be some degree of consensus that human experience is part self-organization and part self-regulation.

Even when the focus is on planful behavior, the two kinds of models seem to complement each other in a different way. The feedback model provides a mechanism through which goal-directed action is managed, which the phase-space model lacks. The phase-space model suggests ways of thinking about how multiple goals exist and how people shift among those multiple goals over time, an issue that is not dealt with as easily in terms of feedback processes.

That is, think of the landscape of chaotic attractors, but with many different basins rather than just two or three.This seems to capture rather well the sense of human behavior. No basin in this system ever becomes a point attractor. Behavior tends toward one goal and then another, never being completely captured by any goal. The person does one thing for a while, then something else. The goals are all predictable—in the sense that they all influence the person—and the influence is highly predictable when aggregated across time. But the shifts from one to another occur unpredictably (thus being chaotic).

Catastrophe Theory

Another set of ideas that has been around for some time but may be reemerging in influence is catastrophe theory, a mathematical model that bears on the creation of discontinuities, bifurcations, or splittings (Brown, 1995; Saunders, 1980; Stewart & Peregoy, 1983; van der Maas & Molenaar, 1992; Woodcock & Davis, 1978; Zeeman, 1977). A catastrophe occurs when a small change in one variable produces an abrupt (and usually large) change in another variable.

An abrupt change implies nonlinearity. This focus on nonlinearity is one of several themes that catastrophe theory shares with dynamic systems theory, though the two bodies of thought have different origins (and are seen by some as quite different from each other—see Kelso, 1995, chap. 2). The similarity is nicely expressed in the statement that the discontinuity in catastrophe theory reflects “the sudden disappearance of one attractor and its basin, combined with the dominant emergence of another attractor” (Brown, 1995, p. 51).

Though several types of catastrophe exist (Brown, 1995; Saunders, 1980; Woodcock & Davis, 1978), the one receiving most attention regarding behavior is the cusp catastrophe, in which two variables influence an outcome. Figure 8.9 portrays its three-dimensional surface. X and z are predictors, and y is the outcome. At low values of z, the surface of the figure shows a roughly linear relationship between x and y. As x increases, so does y. As z increases, the relationship between x and y becomes less linear. It first shifts toward something like a step function. With further increase in z, the x–y relationship becomes even more clearly discontinuous—the outcome is either on the top surface or on the bottom. Thus, changes in z cause a change in the way x relates to y.

Another theme that links catastrophe theory to dynamic systems is the idea of sensitive dependence on initial conditions. The cusp catastrophe displays this characteristic nicely. Consider the portion of Figure 8.9 where z has low values and x has a continuous relation to y (the system’s behavior). Points 1 and 2 on x are nearly identical, but not quite. Now track these points across the surface as z increases. For a while the two paths track each other closely, until suddenly they begin to be separated by the fold in the catastrophe. At higher levels of z, one track ultimately projects to the upper region of the surface, the other to the lower region. Thus, a very slight initial difference results in a substantial difference farther along.

Hysteresis

The preceding description also hinted at an interesting and important feature of a catastrophe known as hysteresis.Asimple characterization of what this term means is that at some levels of z, there is a kind of fold-over in the middle of the x–y relationship.Aregion of x exists in which more than one value of y exists. Another way to characterize hysteresis is that two regions of this surface are attractors and one is a repeller (Brown, 1995).This unstable area is illustrated in Figure 8.10. The dashed-line portion of Figure 8.10 that lies between values a and b on the x-axis—the region where the fold is going backward—repels trajectories (Brown, 1995), whereas the areas near values c and d attract trajectories. To put it more simply, you cannot be on the dashed part of this surface.

Yet another way of characterizing hysteresis is captured by the statement that the system’s behavior depends on the system’s recent history (Brown, 1995; Nowak & Lewenstein, 1994). That is, as you move into the zone of variable x that lies between points a and b in Figure 8.10, it matters which side of the figure you are coming from. If the system is moving from point c into the zone of hysteresis, it stays on the bottom surface until it reaches point b, where it jumps to the top surface. If the system is moving from d into the zone of hysteresis, it stays on the top surface until it reaches point a, where it jumps to the bottom surface.

An Application of Catastrophe Theory

How does catastrophe theory apply to the human behaviors of most interest to personality and social psychologists? Several applications of these ideas have been made in the past decade or so, and others seem obvious candidates for future study (for broader discussion see Carver & Scheier, 1998, chap. 16).

One interesting example concerns what we believe is a bifurcation between engagement in effort and giving up. Earlier we pointed to a set of theories that assume such a disjunction (Brehm & Self, 1989; Kukla, 1972; Wortman & Brehm, 1975). In all those models (as in ours), there is a point at which effort seems fruitless and the person stops trying. Earlier, we simply emphasized that the models all assumed a discontinuity. Now we look at the discontinuity more closely and suggest that the phenomena addressed by these theories may embody a catastrophe.

Figure 8.11 shows a slightly relabeled cross section of a cusp catastrophe similar to that in Figure 8.10. This figure displays a region of hysteresis in the engagement versus disengagement function. In that region, where task demands are close to people’s perceived limits to perform, there should be greater variability in effort or engagement, as some people are on the top surface of the catastrophe and others are on the bottom surface. Some people would be continuing to exert efforts at the same point where others would be exhibiting a giving-up response.

Recall that the catastrophe figure also conveys the sense that the history of the behavior matters. A person who enters the region of hysteresis from the direction of high confidence (who starts out confident but confronts many contradictory cues) will continue to display engagement and effort, even as the situational cues imply less and less basis for confidence. A person who enters that region from the direction of low confidence (who starts doubtful but confronts contradictory cues) will continue to display little effort, even as the cues imply a greater basis for confidence.

This model helps indicate why it can be so difficult to get someone with strong and chronic doubts about success in some domain of behavior to exert real effort and engagement in that domain. It also suggests why a confident person is so rarely put off by encountering difficulties in the domain where the confidence lies. To put it in terms of broader views about life in general, it helps show why optimists tend to stay optimistic and pessimists tend to stay pessimistic, even when the current circumstances of the two sorts of people are identical (i.e., in the region of hysteresis).

It is important to keep in mind that the catastrophe cross section (Figure 8.11) is the picture that emerges under catastrophe theory only once a clear region of hysteresis has begun to develop. Farther back, the model is more of a step function. An implication is that to see the fold-over it is important to engage the variable that is responsible for bringing out the bifurcation in the surface (i.e., axis z in Figure 8.9).

What is the variable that induces the bifurcation? We think that in the motivational models under discussion—and perhaps more broadly—the control parameter is importance. Importance arises from several sources, but there is a common thread among events seen as important. They demand mental resources. We suspect that almost any strong pressure that demands resources (time pressure, self-imposed pressure) will induce bifurcating effects.

Conclusion

In this research paper we sketched a set of ideas that we think are important in conceptualizing human self-regulation. We believe that behavior is goal directed and feedback controlled and that the goals underlying behavior form a hierarchy of abstractness. We believe that experiences of affect (and of confidence vs. doubt) also arise from a process of feedback control, but a feedback process that takes into account temporal constraints. We believe that confidence and doubt yield patterns of persistence versus giving up and that these two responses to adversity form a dichotomy in behavior. These ideas have been embedded in our self-regulatory viewpoint for some time.

We have also recently begun to consider some newer ideas, addressed in the latter parts of the paper. In those sections we described ideas from dynamic systems theory, connectionism, and catastrophe theory. We suggest that they represent useful tools for the analysis and construal of behavior. Our view is that they supplement rather than replace the tools now in use (though not everyone will agree on this point). We see many ways in which those ideas mesh with the ideas presented earlier, though space constraints limited us to discussing that integration only briefly.

In thinking about the structure of behavior, we have tried to draw on ideas from disparate sources while continuing to follow the thread of the logical model from which we started. The result is an aggregation of principles that we think have a good deal to say about how behavioral self-regulation takes place. In so doing, they also say something about personality and how it is manifested in people’s actions.

The conceptual model presented here is surely not complete, and many avenues exist for further discussion and indeed further conceptual development. For example, this research paper included little attention to the issue of how new goals are added to people’s hierarchies or of how to think about growth and change over time (but see Carver & Scheier, 1998, 1999a, 1999b). Similarly, the concepts addressed here bear in several ways on problems in behavior and behavior change, though space constraints prevent us from describing them in detail. For example, we suspect that many problems in people’s lives are, at their core, problems of disengagement versus engagement and the failure to disengage adaptively (Carver & Scheier, 1998). As another example, it may be useful to conceptualize problems as less-than-optimal adaptations in a multidimensional phase space, which require some jostling to bounce the person to a new attractor (Hayes & Strauss, 1998). These are all areas in which more work remains to be done.

These are just some of the ways in which we think the family of ideas described here will likely be explored in the near future. Further analyses of the self-regulation of behavior are likely to produce insights that transform the models from which the insights grew.As the models change, so will our understanding of motivational processes and of how human beings function as coherent, autonomous units. This we take to be one of the core pursuits of personality psychology.

Bibliography:

1. Anderson, J. R. (1990). The adaptive character of thought. Hillsdale, NJ: Erlbaum.
2. Austin, J. T., & Vancouver, J. B. (1996). Goal constructs in psychology: Structure, process, and content. Psychological Bulletin, 120, 338–375.
3. Baumeister, R. F. (1989). The problem of life’s meaning. In D. M. Buss & N. Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 138–148). New York: SpringerVerlag.
4. Baumeister, R. F., & Heatherton, T. F. (1996). Self-regulation failure: An overview. Psychological Inquiry, 7, 1–15.
5. Beer, R. D. (1995). A dynamical systems perspective on agentenvironment interaction. Artificial Intelligence, 72, 173–215.
6. Brehm, J. W., & Self, E. A. (1989). The intensity of motivation. Annual Review of Psychology, 40, 109–131.
7. Brown, C. (1995). Chaos and catastrophe theories. Thousand Oaks, CA: Sage.
8. Brunstein, J. C. (1993). Personal goals and subjective well-being: A longitudinal study. Journal of Personality and Social Psychology, 65, 1061–1070.
9. Cacioppo, J. T., Gardner, W. L., & Berntson, G. G. (1999). The affect system has parallel and integrative processing components: Form follows function. Journal of Personality and Social Psychology, 76, 839–855.
10. Cacioppo, J. T., & Petty, R. E. (1980). The effects of orienting task on differential hemispheric EEG activation. Neuropsychologia, 18, 675–683.
11. Cantor, N., & Kihlstrom, J. F. (1987). Personality and social intelligence. Englewood Cliffs, NJ: Prentice Hall.
12. Carver, C. S. (1979). Acybernetic model of self-attention processes. Journal of Personality and Social Psychology, 37, 1251–1281.
13. Carver, C. S. (2001). Affect and the functional bases of behavior: On the dimensional structure of affective experience. Personality and Social Psychology Review, 5, 345–356.
14. Carver, C. S. (in press). Pleasure as a sign you can attend to something else: Placing positive feelings within a general model of affect. Cognition and Emotion.
15. Carver, C. S., & Humphries, C. (1981). Havana daydreaming: A study of self-consciousness and the negative reference group among Cuban Americans. Journal of Personality and Social Psychology, 40, 545–552.
16. Carver, C. S., Lawrence, J. W., & Scheier, M. F. (1999). Selfdiscrepancies and affect: Incorporating the role of feared selves. Personality and Social Psychology Bulletin, 25, 783–792.
17. Carver, C. S., Peterson, L. M., Follansbee, D. J., & Scheier, M. F. (1983). Effects of self-directed attention on performance and persistence among persons high and low in test anxiety. Cognitive Therapy and Research, 7, 333–354.
18. Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-theory approach to human behavior. New York: Springer-Verlag.
19. Carver, C. S., & Scheier, M. F. (1982). Outcome expectancy, locus of attributions for expectancy, and self-directed attention as determinants of evaluations and performance. Journal of Experimental Social Psychology, 18, 184–200.
20. Carver, C. S., & Scheier, M. F. (1990). Origins and functions of positive and negative affect: A control-process view. Psychological Review, 97, 19–35.
21. Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. New York: Cambridge University Press.
22. Carver, C. S., & Scheier, M. F. (1999a). Themes and issues in the self-regulation of behavior. In R. S. Wyer, Jr. (Ed.), Advances in social cognition (pp. 1–105). Mahwah, NJ: Erlbaum.
23. Carver, C. S., & Scheier, M. F. (1999b). Several more themes, a lot more issues: Commentary on the commentaries. In R. S. Wyer, Jr. (Ed.), Advances in social cognition (Vol. 12, pp. 261–302). Mahwah, NJ: Erlbaum.
24. Carver,C.S.,&Scheier,M.F.(2000).Autonomyandself-regulation. Psychological Inquiry, 11, 284–291.
25. Carver, C. S., & Scheier, M. F. (2001). Optimism, pessimism, and self-regulation. In E. C. Chang (Ed.), Optimism and pessimism: Implications for theory, research, and practice (pp. 31–51). Washington, DC: American Psychological Association.
26. Carver, C. S., & Scheier, M. F. (in press). Control processes and self-organization as complementary principles underlying behavior. Personality and Social Psychology Review.
27. Chaiken, S. L., & Trope, Y. (Eds.). (1999). Dual-process theories in social psychology. New York: Guilford.
28. Clark, L. A., Watson, D., & Mineka, S. (1994). Temperament, personality, and the mood and anxiety disorders. Journal of Abnormal Psychology, 103, 103–116.
29. Clark, N. (1996). Control system dynamics. New York: Cambridge University Press.
30. Cloninger, C. R. (1987).Asystematic method for clinical description and classification of personality variants. Archives of General Psychiatry, 44, 573–588.
31. Cloninger, C. R. (1988). A unified biosocial theory of personality and its role in the development of anxiety states: A reply to commentaries. Psychiatric Developments, 2, 83–120.
32. Davidson, R. J. (1992a). Anterior cerebral asymmetry and the nature of emotion. Brain and Cognition, 20, 125–151.
33. Davidson, R. J. (1992b). Prolegomenon to the structure of emotion: Gleanings from neuropsychology. Cognition and Emotion, 6, 245–268.
34. Davidson, R. J., Ekman, P., Saron, C. D., Senulis, J. A., & Friesen, W. V. (1990). Approach–withdrawal and cerebral asymmetry: Emotional expression and brain physiology I. Journal of Personality and Social Psychology, 58, 330–341.
35. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and selfdetermination in human behavior. New York: Plenum.
36. Deci, E. L., & Ryan, R. M. (2000). The “what” and “why” of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11, 227–268.
37. Depue, R. A., & Iacono, W. G. (1989). Neurobehavioral aspects of affective disorders. Annual Review of Psychology, 40, 457–492.
38. Depue, R. A., Krauss, S. P., & Spoont, M. R. (1987). A twodimensional threshold model of seasonal bipolar affective disorder. In D. Magnusson & A. Öhman (Eds.), Psychopathology: An interactional perspective (pp. 95–123). Orlando, FL: Academic Press.
39. Elliott, E. S., & Dweck, C. S. (1988). Goals: An approach to motivation and achievement. Journal of Personality and Social Psychology, 54, 5–12.
40. Emmons, R. A. (1986). Personal strivings: An approach to personality and subjective well being. Journal of Personality and Social Psychology, 51, 1058–1068.
41. Epstein, S. (1979). The stability of behavior: I. On predicting most of the people much of the time. Journal of Personality and Social Psychology, 37, 1097–1126.
42. Fowles, D. C. (1980). The three arousal model: Implications of Gray’s two-factor learning theory for heart rate, electrodermal activity, and psychopathy. Psychophysiology, 17, 87–104.
43. Fox, N. A., & Davidson, R. J. (1988). Patterns of brain electrical activity during facial signs of emotion in 10-month old infants. Developmental Psychology, 24, 230–236.
44. Frijda, H. (1986). The emotions. Cambridge, England: Cambridge University Press.
45. Frijda, N. H. (1988). The laws of emotion. American Psychologist, 43, 349–358.
46. Gleick, J. (1987). Chaos: Making a new science. New York: Viking Penguin.
47. Gobet, F., & Simon, H.A. (1996). The roles of recognition processes and look-ahead search in time-constrained expert problem solving: Evidence from grand-master-level chess. Psychological Science, 7, 52–55.
48. Goldin-Meadow, S., & Alibali, M. W. (1995). Mechanisms of transition: Learning with a helping hand. In D. Medin (Ed.), The psychology of learning and motivation (Vol. 33, pp. 115–157). San Diego, CA: Academic Press.
49. Gray, J. A. (1982). The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system. New York: Oxford University Press.
50. Gray, J. A. (1987a). Perspectives on anxiety and impulsivity: A commentary. Journal of Research in Personality, 21, 493–509.
51. Gray, J. A. (1987b). The psychology of fear and stress. Cambridge, England: Cambridge University Press.
52. Gray, J. A. (1990). Brain systems that mediate both emotion and cognition. Cognition and Emotion, 4, 269–288.
53. Greenwald, A. G. (1980). The totalitarian ego: Fabrication and revision of personal history. American Psychologist, 35, 603–618.
54. Hayes, A. M., & Strauss, J. L. (1998). Dynamic systems theory as a paradigm for the study of change in psychotherapy: An application to cognitive therapy for depression. Journal of Consulting and Clinical Psychology, 66, 939–947.
55. Hayes-Roth, B., & Hayes-Roth, F. (1979). A cognitive model of planning. Cognitive Science, 3, 275–310.
56. Helson, H. (1964). Adaptation-level theory: An experimental and systematic approach to behavior. New York: Harper & Row.
57. Henriques, J. B., & Davidson, R. J. (1991). Left frontal hypoactivation in depression. Journal of Abnormal Psychology, 100, 535–545.
58. Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94, 319–340.
59. Higgins, E. T. (1996). Ideals, oughts, and regulatory focus: Affect and motivation from distinct pains and pleasures. In P. M. Gollwitzer & J. A. Bargh (Eds.), The psychology of action: Linking cognition and motivation to behavior (pp. 91–114). New York: Guilford.
60. Higgins, E.T., Shah, J., & Friedman, R. (1997). Emotional responses to goal attainment: Strength of regulatory focus as moderator. Journal of Personality and Social Psychology, 72, 515–525.
61. Higgins, E. T., & Tykocinski, O. (1992). Self-discrepancies and biographical memory: Personality and cognition at the level of psychological situation. Personality and Social Psychology Bulletin, 18, 527–535.
62. Holyoak, K. J., & Spellman, B. A. (1993). Thinking. Annual Review of Psychology, 44, 265–315.
63. Hsee, C. K., & Abelson, R. P. (1991). Velocity relation: Satisfaction as a function of the first derivative of outcome over time. Journal of Personality and Social Psychology, 60, 341–347.
64. Isen, A. M. (2000). Positive affect and decision making. In M. Lewis & J. M. Haviland-Jones (Eds.), Handbook of emotions (2nd ed., pp. 417–435). New York: Guilford.
65. Kelly, G. A. (1955). The psychology of personal constructs. New York: W. W. Norton.
66. Kelso, J. A. S. (1995). Dynamic patterns: The self-organization of brain and behavior. Cambridge, MA: MIT Press.
67. Kirschenbaum, D. S. (1985). Proximity and specificity of planning: A position paper. Cognitive Therapy and Research, 9, 489–506.
68. Klinger, E. (1975). Consequences of commitment to and disengagement from incentives. Psychological Review, 82,
69. Klinger, E. (1977). Meaning and void: Inner experience and the incentives in people’s lives. Minneapolis: University of Minnesota Press.
70. Kukla, A. (1972). Foundations of an attributional theory of performance. Psychological Review, 79, 454–470.
71. Kunda, Z., & Thagard, P. (1996). Forming impressions from stereotypes, traits, and behaviors: A parallel-constraint-satisfaction theory. Psychological Review, 103, 284–308.
72. Lang, P. J. (1995). The emotion probe: Studies of motivation and attention. American Psychologist, 50, 372–385.
73. Lang, P. J., Bradley, M., & Cuthbert, B. (1990). Emotion, attention, and the startle reflex. Psychological Review, 97, 377–395.
74. Lawrence, J. W., Carver, C. S., & Scheier, M. F. (in press). Velocity toward goal attainment in immediate experience as a determinant of affect. Journal of Applied Social Psychology.
75. Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press.
76. Little, B. R. (1983). Personal projects: A rationale and methods for investigation. Environment and Behavior, 15, 273–309.
77. Lorenz, E. N. (1963). Deterministic nonperiodic flow. Journal of Atmospheric Science, 20, 130–141.
78. MacKay, D. M. (1956). Towards an information-flow model of human behaviour. British Journal of Psychology, 47, 30–43.
79. Maes, P. (1994). Modeling adaptive autonomous agents. Artificial Life, 1, 135–162.
80. Maes, P., & Brooks, R. A. (1990). Learning to coordinate behaviors. Proceedings of the American Association of Artificial Intelligence (pp. 796–802). Los Alto, CA: Kaufmann.
81. Mahoney, M. J. (1991). Human change processes: The scientific foundations of psychotherapy. New York: Basic Books.
82. Markus, H., & Nurius, P. (1986). Possible selves. American Psychologist, 41, 954–969.
83. McIntosh, W. D., & Martin, L. L. (1992). The cybernetics of happiness: The relation of goal attainment, rumination, and affect. In M. S. Clark (Ed.), Review of personality and social psychology: Volume 14. Emotion and social behavior (pp. 222–246). Newbury Park, CA: Sage.
84. Melton, R. J. (1995). The role of positive affect in syllogism performance.PersonalityandSocialPsychologyBulletin,21,788–794.
85. Millar, K. U., Tesser, A., & Millar, M. G. (1988). The effects of a threatening life event on behavior sequences and intrusive thought: A self-disruption explanation. Cognitive Therapy and Research, 12, 441–458.
86. Miller, G. A., Galanter, E., & Pribram, K. H. (1960). Plans and the structure of behavior. New York: Holt, Rinehart, & Winston.
87. Miller, L. C., & Read, S. J. (1987). Why am I telling you this? Selfdisclosureinagoal-basedmodelofpersonality.InV.J.Derlega& J. Berg (Eds.), Self-disclosure: Theory, research, and therapy (pp. 35–58). NewYork: Plenum.
88. Moskowitz, J. T., Folkman, S., Collette, L., & Vittinghoff, E. (1996). Coping and mood during AIDS-related caregiving and bereavement. Annals of Behavioral Medicine, 18, 49–57.
89. Muraven, M., Tice, D. M., & Baumeister, R. F. (1998). Self-control as a limited resource: Regulatory depletion patterns. Journal of Personality and Social Psychology, 74, 774–789.
90. Myers, D. G., & Diener, E. (1995). Who is happy? Psychological Science, 6, 10–19.
91. Nowak, A., & Lewenstein, M. (1994). Dynamical systems: A tool for social psychology. In R. R. Vallacher & A. Nowak (Eds.), Dynamical systems in social psychology (pp. 17–53). San Diego, CA: Academic Press.
92. Nowak, A., & Vallacher, R. R. (1998). Dynamical social psychology. New York: Guilford.
93. Orbuch, T. L. (Ed.). (1992). Close relationship loss: Theoretical approaches. New York: Springer-Verlag.
94. Ortony, A., Clore, G. L., & Collins, A. (1988). The cognitive structure of emotions. Cambridge, England: Cambridge University Press.
95. Pervin, L. A. (Ed.). (1989). Goal concepts in personality and social psychology. Hillsdale, NJ: Erlbaum.
96. Powers, W. T. (1973). Behavior: The control of perception. Chicago: Aldine.
97. Prigogine, I., & Stengers, I. (1984). Order out of chaos: Man’s new dialogue with nature. New York: Random House.
98. Read, S. J., Druian, P. R., & Miller, L. C. (1989). The role of causal sequence in the meaning of action. British Journal of Social Psychology, 28, 341–351.
99. Read, S. J., Vanman, E. J., & Miller, L. C. (1997). Connectionism, parallel constraint satisfaction processes, and Gestalt principles: (Re)introducing cognitive dynamics to social psychology. Review of Personality and Social Psychology, 1, 26–53.
100. Roseman, I. J. (1984). Cognitive determinants of emotions: A structural theory. In P. Shaver (Ed.), Review of personality and social psychology (Vol. 5, pp. 11–36). Beverly Hills, CA: Sage.
101. Ruble, D. N. (1994). A phase model of transitions: Cognitive and motivational consequences. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 26, pp. 163–214). San Diego, CA: Academic Press.
102. Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well being. American Psychologist, 55, 68–78.
103. Saunders, P. T. (1980). An introduction to catastrophe theory. Cambridge, England: Cambridge University Press.
104. Scheier, M. F., & Carver, C. S. (1983). Self-directed attention and the comparison of self with standards. Journal of Experimental Social Psychology, 19, 205–222.
105. Scheier, M. F., Carver, C. S., & Bridges, M. W. (2001). Optimism, pessimism, and psychological well-being. In E. C. Chang (Ed.), Optimism and pessimism: Implications for theory, research, and practice (pp. 189–216). Washington, DC: American Psychological Association.
106. Scherer, K. R., & Ekman, P. (Eds.). (1984). Approaches to emotion. Hillsdale, NJ: Erlbaum.
107. Schultz, T. R., & Lepper, M. R. (1996). Cognitive dissonance reduction as constraint satisfaction. Psychological Review, 103, 219–240.
108. Shastri, L., & Ajjanagadde, V. (1993). From simple associations to systematic reasoning: A connectionist representation of rules, variables, and dynamic bindings using temporal synchrony. Behavioral and Brain Sciences, 16, 417–494.
109. Siegler, R. S., & Jenkins, E. A. (1989). How children discover new strategies. Hillsdale, NJ: Erlbaum.
110. Sloman, S. A. (1996). The empirical case for two forms of reasoning. Psychological Bulletin, 119, 3–22.
111. Smith, E. R. (1996). What do connectionism and social psychology offer each other? Journal of Personality and Social Psychology, 70, 893–912.
112. Smolensky, P. (1988). On the proper treatment of connectionism. Behavioral and Brain Sciences, 11, 1–23.
113. Sobotka, S. S., Davidson, R. J., & Senulis, J. A. (1992). Anterior brain electrical asymmetries in response to reward and punishment. Electroencephalography and Clinical Neurophysiology, 83, 236–247.
114. Solomon,R.L.(1980).Theopponent-processtheoryofacquiredmotivation: The costs of pleasure and the benefits of pain. American Psychologist, 35, 691–712.
115. Solomon, R. L., & Corbit, J. D. (1974). An opponent-process theory of motivation: III. Temporal dynamics of affect. Psychological Review, 81, 119–145.
116. Stewart, I. N., & Peregoy, P. L. (1983). Catastrophe theory modeling in psychology. Psychological Bulletin, 94, 336–362.
117. Stroebe, M. S., Stroebe, W., & Hansson, R. O. (Eds.). (1993). Handbook of bereavement: Theory, research, and intervention. Cambridge, England: Cambridge University Press.
118. Taylor, S. E., & Pham, L. B. (1996). Mental stimulation, motivation, and action. In P. M. Gollwitzer & J. A. Bargh (Eds.), The psychology of action: Linking cognition and motivation to behavior (pp. 219–235). New York: Guilford.
119. Thagard, P. (1989). Explanatory coherence. Behavioral and Brain Sciences, 12, 435–467.
120. Thelen, E. (1992). Development as a dynamic system. Current Directions in Psychological Science, 1, 189–193.
121. Thelen, E. (1995). Motor development: A new synthesis. American Psychologist, 50, 79–95.
122. Thelen, E., & Smith, L. B. (1994). A dynamic systems approach to the development of cognition and action. Cambridge, MA: MIT Press.
123. Tomarken, A. J., Davidson, R. J., Wheeler, R. E., & Doss, R. C. (1992). Individual differences in anterior brain asymmetry and fundamental dimensions of emotion. Journal of Personality and Social Psychology, 62, 676–687.
124. Vallacher, R. R., & Nowak, A. (Eds.). (1994). Dynamical systems in social psychology. San Diego, CA: Academic Press.
125. Vallacher, R. R., & Nowak, A. (1997). The emergence of dynamical social psychology. Psychological Inquiry, 8, 73–99.
126. Vallacher, R. R., & Wegner, D. M. (1985). A theory of action identification. Hillsdale, NJ: Erlbaum.
127. van der Maas, H. L. J., & Molenaar, P. C. M. (1992). Stagewise cognitive development: An application of catastrophe theory. Psychological Review, 99, 395–417.
128. Waldrop, M. (1992). Complexity: The emerging science at the edge of order and chaos. New York: Simon & Schuster.
129. Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood. Psychological Bulletin, 98, 219–235.
130. Watson, D., Wiese, D., Vaidya, J., & Tellegen, A. (1999). The two general activation systems of affect: Structural findings, evolutionary considerations, and psychobiological evidence. Journal of Personality and Social Psychology, 76, 820–838.
131. Wegner, D. M. (1994). Ironic processes of mental control. Psychological Review, 101, 34–52.
132. Weiss, R. S. (1988). Loss and recovery. Journal of Social Issues, 44, 37–52.
133. Wiener, N. (1948). Cybernetics: Control and communication in the animal and the machine. Cambridge, MA: MIT Press.
134. Wills, T. A. (1981). Downward comparison principles in social psychology. Psychological Bulletin, 90, 245–271.
135. Wood, J. V. (1989). Theory and research concerning social comparisons of personal attributes. Psychological Bulletin, 106, 231–248.
136. Woodcock, A., & Davis, M. (1978). Catastrophe theory. New York: E. P. Dutton.
137. Wortman, C. B., & Brehm, J. W. (1975). Responses to uncontrollable outcomes: An integration of reactance theory and the learned helplessness model. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 8, pp. 277–336). New York: Academic Press.
138. Wright, R. A., & Brehm, J. W. (1989). Energization and goal attractiveness. In L. A. Pervin (Ed.), Goal concepts in personality and social psychology (pp. 169–210). Hillsdale, NJ: Erlbaum.
139. Zeeman, E. C. (1977). Catastrophe theory: Selected papers 1972–1977. Reading, MA: Benjamin.