If we presented one voice in the target article to describe a goal-based model of behavior, then it seems the article elicited (following Whitman) multitudes of perspectives. Taken together, the twenty-six expert commentaries are a mirror that reflects back to us important complexities and debates inherent to the study of goals. We thank all the commentators for their expert feedback; we have learned tremendously from them.
In our response, we emphasize how the commentaries' diverse viewpoints complement each other, and the target article, more than they conflict. We clarify what the Selfish Goal model addresses (common goal structure) and what it does not evaluate (specific functions of consciousness). Critiques of the “selfish” metaphor are addressed by describing how both the similarities and the incongruities offered by the goal-to-gene analogy have the potential to inform future goal research.
With help from the commentaries, we then situate the model's contributions within motivational science and elaborate on additional mechanisms that may drive selfish goal behavior. We also respond to the theme of goal dynamics, which emerged in the commentaries, and make the preliminary observation that control systems theory may prove particularly instrumental when incorporating lateral and hierarchical goal relationships into our model of selfish goals. To conclude our response, we discuss the implications of the present framework as it applies to individuals within social institutions and environments.
R1. Clarifications about the role of conscious and unconscious processes
In the target article, we argue that goals are mental representations that operate in ways that can be described as selfish; consequently, the exact functions of consciousness fall outside this argument. Some commentators (e.g., Moskowitz & Balcetis) appear to have misinterpreted the article as making the claim that conscious processes are “dispensable”; to the contrary, we proposed a model that assumes a key role for the integrative functions of conscious states (sect. 2.1).
Part of the misunderstanding may have arisen from our closing remarks regarding the key role of conscious thought in creating rationalizations and managing inconsistencies (sect. 5.2), and we take this opportunity to clarify our intent. We wanted to (1) refer readers to one prominent idea about the function of consciousness (i.e., Gazzaniga Reference Gazzaniga2005) and (2) highlight that conscious understanding of unconsciously rooted behaviors can be mistaken, with the behavior interpreted in terms of plausibility rather than insight (our use of the word “rationalization” emphasized that accuracy is not a guaranteed part of this process; see Bar-Anan et al. Reference Bar-Anan, Wilson and Hassin2010).
It is also possible that misreadings of our intent may reflect the influence of another related and topical research tradition that questions the extent to which certain psychological phenomena are conscious or unconscious (e.g., Newell & Shanks Reference Newell and Shanks2014). However, we highlighted evidence for the existence of unconscious processes (sect. 3.2) and similarities between conscious and unconscious goals (sect. 3.4) to underscore a different argument – namely, that many interesting and theory-relevant similarities exist between the two forms of pursuit, and this knowledge should be brought to bear on current views in motivational science. With this clarification, we hope to minimize potential future misreadings of our model and refocus the conversation toward the structure and interaction of goals (both conscious and unconscious) as they organize a person's behavior.
R2. Does selfish = useful?
We especially appreciate the opportunity to discuss the “selfish” metaphor we borrowed from Dawkins (Reference Dawkins1976) to describe the relationship of goals to the person pursuing them. The analogy struck a chord with reviewers, whose responses ranged from enthusiastic support to strong reservations. At the early stages of model development, we believe a metaphor's usefulness in behavioral science might be judged by how it calls attention to key similarities and differences that guide further theorizing and study, and casts fresh eyes within an area of research (Gentner & Grudin Reference Gentner and Grudin1985; Hoffman et al. Reference Hoffman, Cochran, Nead and Leary1990). The analogy employed in the target article helps accomplish both tasks, and it also addresses other concerns voiced by the commentaries.
Specifically, the selfish metaphor conveys a number of key similarities between goals and genes, including notions of independence, the existence of a smaller “selectable” unit of analysis, multiple autonomous units co-existing within a single entity, and competition and cooperation between units. Many commentaries responded positively to these concepts and the model inspired by the analogy (e.g., Becker & Kenrick; Conroy-Beam & Buss; Cuzen, Fineberg, & Stein [Cuzen et al.]; Eitam & Higgins; Hirsh; Kay & Jost; Mazzone; and Neuberg & Schaller). Their insights, and the promising directions they offer for future research, are discussed in sections below.
Interestingly, clinically oriented commentators were especially likely to endorse the analogy and/or the model it inspired. This may be – at least partially – because multiple clinical interventions employ a similar focus on the mechanistic underpinnings of behavior when trying to effect change in the patient him or herself. For example, the disease model of addiction de-emphasizes the extent to which patients' destructive behavioral patterns (e.g., gambling) are framed as issues of personal responsibility or shame. Therapeutic interventions highlight the specific steps that a patient must follow to fix his or her current state (regardless of how that patient got there). Other clinical approaches, such as motivational interviewing, similarly shift focus away from substance use as a fundamental issue of willpower (or lack thereof), and instead encourage patients to understand the costs and benefits associated with using, and those involved in quitting. Consequently, our model may have especially appealed to researchers who regularly develop interventions based on the nuts and bolts of behavior.
It is less clear to us, however, why the utility of the selfish metaphor should hinge on questions of evolutionary biology. Mattei criticized our model for relying on the empirical validity of Dawkins's original selfish gene theory, which is “at best, [an] unproven biological paradigm.” In our target article we specified that the comparison between goals and genes was an analogy and that questions regarding specific mechanisms of evolution fall outside the scope of our argument.
A number of commentators voiced more serious concerns that the metaphor was misleading because discrepancies exist between a gene's interests (vis-à-vis the individual) and a goal's interests (e.g., Fedyk & Kushnir; and Merker). For example, Merker noted that goals cannot be selfish exactly as genes are said to be selfish, because genetic self-interest is defined by its host-independence (ability to exist independent of the individual organism), whereas a goal cannot exist outside of the person who pursues it. Spurrett additionally questioned how goals are to be “rewarded” as genes are rewarded through replication along a similar vein.
It is important to note that even imperfect analogies can prove useful if the dissimilarities help generate new ideas for research (e.g., Cornelissen Reference Cornelissen2004; Hoffman et al. Reference Hoffman, Cochran, Nead and Leary1990). Social science provides examples where both similarities and dissimilarities between compared domains lead to productive lines of research. For example, natural selection within population ecology has provided a useful platform to conceptualize organizational behavior (Amburgey & Rao Reference Amburgey and Rao1996; Hannan & Freeman Reference Hannan and Freeman1977). Understanding that organizations succeed or fail as if they were living beings highlights key similarities between the domains such as dependence on the environment, competition for limited resources, and stages of existence which relate to health and age. Of course, organizations are not exactly like species; species adapt through genetic mechanisms, whereas organizations change through learning processes. Knowing the mismatch exists, however, inspired the eventual identification of “active” and “passive” learning processes unique to organizational adaptation (e.g., Stopford Reference Stopford, Dierkes, Antal, Child and Nonaka2003).
The strength model of self-control (Baumeister et al. Reference Baumeister, Bratslavsky, Muraven and Tice1998) also provides examples in which evoked similarities and dissimilarities serve as foundations for productive research lines on willpower (Baumeister et al. Reference Baumeister, Schmeichel, Vohs, Kruglanski and Higgins2007; Hagger et al. Reference Hagger, Wood, Stiff and Chatzisarantis2010). Conceptualizing self-control as a muscle nicely captures similarities between the two domains, including dependence on a limited resource, possibilities for short-term impairment, vulnerability to extreme usage, and potential improvement given long-term training. Moreover, the dissimilarities between willpower and muscle-power are also interesting. Biological resources do not explain self-regulatory abilities exactly as they do physical acts, leading Job and colleagues (Reference Job, Dweck and Walton2010; see also Molden et al. Reference Molden, Hui, Scholar, Meier, Noreen, D'Agostino and Maritn2012) to discover that participants' lay theories about willpower have the unique potential to predict whether they will be successful at self-regulatory tasks. Importantly, the authors themselves saw these results as offering promising lines of future research, rather than necessarily challenging the overall utility of the muscle metaphor.
Just as organizations adapt in a different manner than species-related change, and willpower is distinct from muscle strength, a goal's selfishness will differ from the exact definition of self-interest as applied to genes (Dawkins Reference Dawkins1976). Consequently, although we acknowledge Merker's point that a goal cannot be self-interested exactly as Dawkins defined it within the gene-individual relationship, we nevertheless maintain that an appropriately contextualized concept of goal-selfishness provides directions for insights into motivational science.
Specifically, key comparisons should be made between genes and the (physical) individual organism, and goals and an individual's corpus of behavior. Appreciating that any selfishness of goals must occur within an individual's corpus of behavior (as opposed to across generations of individual organisms) focuses our inquiry on observable replication-like effects within a person's lifetime. Indeed, research suggests that a long-term consequence of achieving a goal is that one typically becomes more strongly motivated to pursue that goal in the future (as held by self-efficacy and other theories; e.g., Bandura Reference Bandura1977). A combination of effects associated with success at goal pursuit can theoretically produce a goal-perpetuation effect, or literally a replication of the goal into the future, in which goals that are successful become stronger, whereas those not as successful became weaker based on performance feedback from the environments frequented by the person. Put another way, what is “in it” for goals is to be pursued more often; and the prediction that success at a given goal pursuit begets future successful iterations is in harmony with other motivational models. As held by self-efficacy theory and supported by more recent research using affective conditioning by Custers and Aarts (Reference Custers and Aarts2005), the positive affect associated with goal attainment increases the incentive value of that goal in the future for the individual. Meanwhile, the subsequent formation of situational cue-goal associative linkages simultaneously increases the relative likelihood of pursuit (e.g., Bargh Reference Bargh, Higgins and Sorrentino1990; Gollwitzer Reference Gollwitzer1999; Veltkamp et al. Reference Veltkamp, Aarts and Custers2008) and enables readiness in precisely those situations in which the individual has historically encountered success at pursuit.
More generally, the selfish metaphor provides a newer perspective on current goal research, refocusing attention from possibly unresolvable debates regarding the relative strength of one goal compared to another (Kay & Jost). The analogy also links areas of research that are less easily incorporated into commentators' alternative models. Whereas Moskowitz & Balcetis suggested that unconscious goals are all originally conscious and become unconscious only from extensive experience (to then ultimately promote an individual's well-being), that alternative runs counter to the considerable evidence and theory regarding innate, evolutionarily acquired adaptive goals (Becker & Kenrick; Conroy-Beam & Buss; and Neuberg & Schaller), infant motivations acquired prior to any experience (e.g., Baillargeon et al. Reference Baillargeon, He, Setoh, Scott, Sloan, Yan, Banaji and Gelmanm2013), and extremely negative outcomes of addiction-related disorders (e.g., Cuzen et al.; Müller & Amato). Moreover, interpreting the goal literature from this “conscious first” perspective, or one which sees “individuals choosing between competing opportunities” (as suggested by von Hippel & von Hippel) faces the additional hurdle of explaining how a person chooses and actively pursues opportunities that promote well-being when he or she is unaware of having, or actively pursuing, the goal.
In contrast, conceiving of goal influences as operating selfishly offers a reliable way to describe how a person instigates, guides, and stops processes of which he or she might not be aware (because these processes do not require individual awareness for their operation). Moreover, a goal-driven perspective helps account for research that suggests even when people consciously instigate goals, they are not necessarily in control of all operational features of pursuit, such as which environmental stimuli will be affected by pursuit and when the end-state is attained (Bargh et al. Reference Bargh, Green and Fitzsimons2008).
Animate-being metaphors (where “ideas or aspects of the mind are likened to creatures,” as when goals are described as selfish; Gentner & Grudin Reference Gentner and Grudin1985, p.184) were frequently employed in psychological discourse during the turn of the century, but they disappeared with increased popularity of other types of metaphor (e.g., computer systems metaphors; Gentner & Grudin Reference Gentner and Grudin1985). We concede that the relative novelty of an animate-being metaphor highlights the need for interpretive caution, but we are far less convinced of Mattei's argument that it represents a return to the “naïve intellectual modus operandus” that earlier plagued prescientific reasoning. Although we are indeed in early stages, the merits of our model should be judged according to how it helps conceptualize findings in recent research and points to areas that need further theorizing and study.
R2.1. Goals as critical pieces of the whole
Von Hippel & von Hippel (with similar points raised by Mattei and Merker) questioned the theoretical novelty of the Selfish Goal model, observing redundancies between examples of selfish goal behavior and evolutionary trade-offs. Those commentaries correctly point out that some research cited in support of our model (e.g., studies that find an active mating goal changes how an individual prioritizes physical self-protection) are already understood as cases of sexual selection favoring suboptimal individual-level outcomes.
If the target article had sought to inform evolutionary theory, that critique would be fair; however, it did not. The selfish goal framework is less concerned with isolating when goal conflicts are likely to result in evolutionary trade-offs, or identifying which goals are ultimately best for the individual in those conflicts, than it is with examining the structural commonalities that span the pursuit of all goals (including, but not limited to, those involved in sexual selection). Toward those ends, evolutionarily adaptive goals simply provide powerful demonstrations of the reconfiguration principle. We cited experimental demonstrations of mating imperatives conflicting with self-protection goals to illustrate how people's attitudes can shift based on the temporary goal they are pursuing – even in cases where pursuit of that goal results in arguably negative health consequences.
Consequently, we agree with Merker that our target article concerns the “more mundane circumstances attending those [goal] pursuits” – although we might have used the adjective “diverse” instead to emphasize the importance of context in the study of goal structure. Charting how goal operation unfolds across multiple, everyday circumstances helps reveal fundamental patterns to its influence over individual human perception and behavior. To take as an example: the self-protection goal is associated with functional changes to perception and behavior that, in ancestral environments, increased the likelihood that a person would avoid potential threats (e.g., being quicker to notice and sidestep a snake in one's path; Kenrick et al. Reference Kenrick, Griskevicius, Neuberg and Schaller2010; Neuberg et al. Reference Neuberg, Kenrick, Maner, Schaller, Forgas and Williams2004). In current-day circumstances, people react in similar goal-functional ways toward stimuli that are far removed from actual harm (e.g., being quicker to pull a joystick away as the word “snake” is flashed across a computer screen). The wide breadth of goal pursuit, as it spans contexts of varying ecological validity, indicates the presence of general operational patterns. The Selfish Goal model predicts similar outcomes for evolved goals in past and current contexts, as well as for nonevolved goals in the current context – because, as noted by Neuberg & Schaller, goals today operate through the reconfiguration principle, just as they did in the evolutionary past.
Goal conflict is mentioned often in the target article, which may have led Ainslie to conclude that examples of selfish goals conflicting (e.g., eating a cake versus exercising) are already understood under the framework of intertemporal choice theory. It may be the case that many of our examples involve temporal conflict, such that short-term benefits are weighed more heavily as compared to long-term benefits; however, in everyday life, a person may feel conflict over goals of equivalent temporal qualities – as when one chooses between different hedonic experiences, different future courses of action, or goods associated with different motivations. Moreover, the Selfish Goal model offers additional predictions regarding how the chooser will see the world and act before the choice (becoming quicker to perceive and physically approach stimuli related to the most incentivized goal regardless of when specific benefits associated with goal pursuit can be realized) and immediately afterward (inhibition of mental constructs associated with pursuit).
R2.2. Pathways to selfishness
We are especially grateful to the commentators who were able to anticipate others' concerns regarding the limited treatment of neurophysiological mechanisms in our target article. Bliss-Moreau & Williams, Cuzen et al., Müller & Amato, and Pezzulo elaborate on the affective, neuropsychological, and computational mechanisms that contribute to the selfish pattern of pursuit. In so doing, they provide welcome contributions to the open task of incorporating the Selfish Goal model “along the recognized pathways for behavioural control” (Spurrett).
In the target article, we described a sequence of events whereby the more frequently a goal is pursued, the more likely it is to be incentivized for future pursuit (sect. 3.2.1); Bliss-Moreau & Williams nicely expanded on this topic. Their commentary addresses how positive affect (incentive) becomes associated with the end-state of that goal, which consequently deprioritizes other goals within the person's repertoire, in accordance with views in many motivational models that affect serves key functions across specific stages of pursuit including goal formation, phenomenal experiences of failure and success (e.g., Kruglanski et al. Reference Kruglanski, Shah, Fishbach, Friedman, Chun, Sleeth-Keppler and Zanna2002), and goal switching (Carver & Scheier Reference Carver and Scheier1998).
Müller & Amato home in on the striatal circuits and physiological processes in which the consequences of goal conflict can be observed. In extreme cases such as with addicts, drug consumption increases learning-like neuronal plasticity (e.g., increased dendritic branching in the brain's reward systems) for that pursuit. The mechanisms that typically reinforce the pursuit of other goals (e.g., establishment of plasticity and learning in the reward circuits) are prevented, thus leaving the more incentivized addiction goal to dominate. We concur with Müller & Amato's speculation that this chemical process may occur to a lesser degree even for the goals that are not as “unusually polarized” as those in substance addiction, and provide an example of “selfishness.” Some selfish goals may be strong enough to block physiological mechanisms that typically reinforce the operation of other goals.
Cuzen et al. forge a similar connection to the literature on impulsive and compulsive disorders. They observe that predictions derived from the Selfish Goal model are very similar to the outcomes observed in clinical populations when the habit system dominates behavior. Individuals with obsessive-compulsive disorder (OCD) bear out the inconsistency principle by acknowledging that their compulsions are unreasonable or harmful even as they continue engaging in those behaviors. Repetitive behaviors can be precipitated by exposure to an environmental cue and, once triggered, are difficult to control, consistent with the automaticity and reconfiguration principles.
Pezzulo translates the reconfiguration principle into the language of computational neuroscience. Specifically, during a phenomenon known as active interference, priors (or goals) have a similar “selfish” influence upon an individual by directing which potential future events have value (as in Ferguson Reference Ferguson2008), which prediction errors should be monitored and evaluated, and the lessons that should be taken from events related to successful and unsuccessful prediction of the future.
Whereas the above commentators elaborated on how a goal might demonstrate selfishness, Huebner & Rupert suggest avoiding such terms altogether. To those authors, constructs such as goals, selfishness, and conscious versus unconscious processes are perhaps too deeply rooted in “traditional folk-taxonomies” because various forms of representation have motivational force. We see conceptual similarities between this view and other commentaries that address expanding the goal construct. For example, Nanay makes the distinction between specific and nonspecific behavioral tendencies; whereas Eitam & Higgins's ROAR framework charts the ways in which a primed or activated latent tendency emerges into action only when a specific second condition is satisfied.
We disagree with Huebner & Rupert that building our model around the goal construct is “likely to inhibit progress in cognitive science.” The goal construct may have roots in folk taxonomy, but it also has a deep and meaningful tradition in psychology (for reviews, see Austin & Vancouver Reference Austin and Vancouver1996; Gollwitzer & Moskowitz Reference Gollwitzer, Moskowitz, Higgins and Kruglanski1996); the existence of this literature helps clarify specific mechanisms of goal selfishness and points to future directions of research. For example, the goal literature is intimately connected to concepts of reward, reinforcement, incentive, and affect not typically associated with cognitive representations (Kruglanski et al. Reference Kruglanski, Shah, Fishbach, Friedman, Chun, Sleeth-Keppler and Zanna2002). (Indeed, as previously mentioned, many of the commentators rightly noted our relative neglect of affect as a variable that helps explain “why” people pursue certain end-states over others.)
Moreover, the goal construct also implicates notions of reference values, discrepancy reduction, and hierarchical feedback loops, which are central concepts within the control theory literature (Carver & Scheier Reference Carver and Scheier1982) and are likely to guide future ideas about pursuit. Consequently, we share Pezzulo's belief that hierarchical control systems provide one of the most promising directions for future research in goal pursuit. For example, a model in which action becomes goal-directed because behaviors are steered in ways that minimize prediction errors between the goal (desired end-state) and the present (current state) is very much in harmony with Carver and Scheier's (Reference Carver and Scheier1982) seminal homeostatic model of goal pursuit via continual discrepancy reduction, and we elaborate upon this below.
R3. Goals in vivo
In the target article we argued that a single active goal can be considered “selfish” in reference to the person expressing it. Because that relation was our main focus, we included only a limited discussion of how goal conflict is resolved between goals, suggesting that a “winner take all” mechanism leads a single goal to dominate (sect. 3.2.3; similar to the “full-system” orientation mentioned in commentaries by Becker & Kenrick, Müller & Amato, and Neuberg & Schaller). Admittedly, the proposed mechanism may be most relevant for some goals (particularly powerful ones, such as evolutionarily adaptive motives, or addictions) as a result, we extend discussion of this topic here.
As the commentators recognized, there are many issues to ponder when considering mechanisms of goal conflict resolution, even in the initial stages of the task. To begin with, one should consider the broader body of research on motivational hierarchies and the specific instances where goals can be said to cooperate. Both topics were of particular emphasis in the commentaries, and here we suggest that cybernetic models of control (e.g., Perceptual Control Theory; Powers Reference Powers1973) may help address these issues, as well as the open questions regarding goal dynamics.
Although the literature on control systems is vast and encompasses a variety of paradigms, it may help explain how relatively autonomous units (goals) give rise to complex systems capable of accommodating environmental variability (individual behavior). (For more comprehensive treatments of control systems within psychology, see Carver & Scheier [1998] and Powers [1973].)
R3.1. Control systems
Broadly stated, control-systems models explain how a comparatively simple system can produce reliable patterns of behavior (i.e., attain goals) in the seemingly infinitely variable environment of the real world. The basic control-system unit of behavioral organization has four main components: (1) an input function (also called a “perception”); (2) a standard (or “reference signal”) against which the current state is compared; (3) comparison of the input against the standard that leads to action given sensed discrepancies; and (4) a feedback effect whereby the action changes the environment, thus updating the original perceptual signal (because actions make an aspect of the world come to a new state; Powers Reference Powers1973). These components stand in relation to one another, and together they form a “loop” whereby organisms act to control how the environment is affecting them (in both positive and negative ways; also called “negative” or “discrepancy-reducing” feedback loops; Carver & Scheier Reference Carver and Scheier2002). Over time, it has the effect of reinforcing (if the environmental feedback is positive) or diminishing (if the environmental feedback is negative) certain behaviors.
Along with others (Carver & Scheier Reference Carver and Scheier2002, p. 305), we find this framework congenial for developing more systems-based understandings of individual behavior because people can be viewed as organizations of self-regulating feedback systems. Input functions represent the current state of the individual (within the environment, and vis-à-vis other goals and inputs from other systems levels). Desired end-states are the standard against which a person's current experienced state is compared. If the input function indicates discrepancies between actual and desired behavior (i.e., the goal has not yet been fulfilled), the individual executes behaviors to try to minimize the detected discrepancy until it is eliminated. A person's streamlined actions can be interpreted as the result of control systems making continual adjustments based upon the feedback (perceptions) these actions produce from the environment.
R3.2. (Self-) organization in goal systems
Within a control system, feedback loops can be linked hierarchically, with higher levels of control broadly correlating with the abstraction of the input (the input itself can span concrete events such as whether specific sensory nerve endings are stimulated or whether sequences of action can be understood according to an abstract standard such as honesty). Indeed, many commentators (e.g., Mazzone; Pezzulo; and Sripada, Swain, Ho, & Swain [Sripada et al.]) steered our attention to how factors at higher levels of processing influence pursuit of a focal goal.
A hierarchical organization of control has the potential to accommodate many of the abstract variables mentioned in the commentaries, including a motivational self (Baumeister & Winegard, Fishbach, and Hirsh) and the motivational relevance of representations (Eitam & Higgins). That also recalls Nanay's distinction between specific versus unspecific goals, insofar as goals at multiple levels of abstraction also have significantly different means and contexts in which they can be successfully pursued (an intuition echoed by Mazzone). We add that higher- and lower-level goals may have different relationships to their goal-means. For example, one can avoid physical harm from a predator through a limited number of ways, including fighting, fleeing, or freezing. A higher-level goal, however, has the potential to be satisfied in a variety of more abstract ways and may implicate a wider breadth of processes to reconfigure. One can cooperate both by volunteering one's time on a survey or by picking up pens dropped by an experimenter. At an even higher level of abstraction, one can make one's parent proud by achieving in school, becoming rich, doing good deeds, or winning athletic competitions.
If they agreed on the importance of goal integration, the commentators disagreed about which goals are most central and how they assemble together. Many commentaries focused on the importance of conscious processes for the integration of constituent goals. For example, Baumeister & Winegard (see also Fishbach) proposed that traditional notions of a “conscious self” can be seen as consisting partly of processes that favor some goals over others. Eitam & Higgins's commentary reviews how broader contextual factors related to control, relevance, and value similarly sway the relative goal priority. Hirsh's commentary describes how effort and self-reflection are critical processes to the creation of a more coherent “self-system,” whereas Sripada et al. provided thematically similar evidence from neuroscience, pointing out that the neuroregions traditionally involved with the resolution of goal competition are also implicated in higher-order processing and conscious self-regulation.
Conversely, Becker & Kenrick (as well as Conroy-Beam & Buss and Neuberg & Schaller) endorsed bottom-up processes that eventually create a hierarchical goal structure. Earlier-developing pursuits that are directly linked to adaptive outcomes (e.g., self-protection) serve as the foundation of a pyramid of universal human needs; they assume greatest priority for goal selection in proximal contexts. Goals that are higher up in the hierarchy (e.g., mate acquisition) develop later in life and are pursued only after successful pursuit of other, more foundational goals.
Carver and Scheier (Reference Carver and Scheier2002) note that within control systems, integration can occur through relatively autonomous, bottom-up processes (endemic to the self-organization capabilities of dynamic computer systems), as well as through top-down influences (traditionally associated with the term self-regulation). That leads to the interesting prediction that the ways in which goal hierarchies are formed may result in different dynamics between the goals organized within that system. Below, we review evidence that various forms of relationships exist between goals in a system, and predict how these patterns might vary according to their place in the hierarchical structure.
R3.3. Selfish goals cooperate
The idea that goals exert reliably selfish influences on the individual should not imply that they are always in competition with each other and never surrender the steering wheel (with pathological and physiologically powerful cases such as addictions being a potential exception; see Müller & Amato). That point was not sufficiently highlighted in our target article, and because leaving this point unclarified would underestimate the scope of behavior that could be incorporated into our model, we spend some time discussing it here.
Thankfully, the commentators had similar intuitions about the relevance of the body of research on goal cooperation; their responses helped correct for its absence in the target article. We outline these below, but we also direct readers to Kopetz, Hofmann, & Wiers (Kopetz et al.), Fishbach, and Carruthers for better descriptions of the empirical evidence than have space for in this reply.
The Selfish Goal model holds that “goals often can and do encourage behaviors that are consistent with (or at least not opposed to) other goals' end-states” (sect. 4, para.1). Indeed, recent social cognitive findings support the notion that goal activation can result not only in selfish effects on the person, but also in cooperative effects that aid in the pursuit of other goals. As mentioned by Kopetz et al., when a currently active goal and an alternate goal are perceived as unrelated to each other, priming the alternate goal undermines persistence and performance on the active goal (which can be seen as evidence for goals competing for the individual's limited cognitive resources). However, when goals are perceived to facilitate each other, priming the alternate goal actually increases persistence and performance on the focal goal (Shah et al. Reference Shah, Friedman and Kruglanski2002) – a phenomenon that is simultaneously cooperative and self-interested.
Goals can also collaborate through more explicit forms of collusion, as when a person's expressed behaviors represent two different end-states (here parallel constraint satisfaction models are particularly relevant; see Kunda & Thagard Reference Kunda and Thagard1996). Carruthers describes a notable example of goal collusion in a series of studies. When people are induced to write opinion essays (as in classic cognitive dissonance paradigms), the expression of the goal to say what one believes is constrained by the requirements of the goal of positive self-presentation. The outcome of the goal conflict is an expressed attitude that falls midway between the attitudes associated with pursuing each goal independently.
Another instance of cooperative goal behavior (or at least behavior that is interpretable in that light) occurs when a goal representation becomes inhibited following completion of the goal pursuit attempt (Förster et al. Reference Förster, Liberman and Higgins2005). As Atkinson and Birch (Reference Atkinson and Birch1970) first argued, the function of the goal turn-off effect is to give other important goals their chances at attainment. Fishbach and Kopetz et al. significantly expand upon this idea, identifying cooperative behaviors in cases where goals are pursued in serial fashion and hence can be seen as cooperative. For example, the fulfillment of one goal may lead to the activation another, as when a person indulges in a chocolate dessert only if he or she has exercised earlier in the day, thereby “justifying” the current pursuit (i.e., changing its incentive value from negative to positive) through prior attainment of another goal.
If selfish goals can compete, operate simultaneously with another goal, or drive pursuit contingent on the completion of another goal, it remains to be seen what factors predict when these relational dynamics will occur. Future research may uncover different forms of goal cooperation and balancing depending on how (or when, in evolution or in development; see Fedyk & Kushnir) they were incorporated into higher-level structures. For example, the goal balancing phenomenon attributed to the operation of Fishbach's motivational self may not characterize the behavior of lateral goals that are linked, bottom-up fashion, in Becker & Kenrick's oligarchy. Exercising may release a person to eat junk food, but successfully avoiding disease may not induce a person to become more amenable to other forms of physical danger.
Developmental stage may also offer insight into when these relational dynamics will occur (Fedyk & Kushnir and Hirsh). Goals that were integrated early in the pursuer's developmental stage may be systematically different than those incorporated in relatively later stages of adulthood, when conscious capabilities can assist with the task. One might expect, for example, that the goal facilitation as described by Shah et al. (Reference Shah, Friedman and Kruglanski2002) (also reviewed by Fishbach and Kopetz et al.) may be particularly likely to characterize goals that are integrated relatively late in development. One might additionally expect that the ability to use one means to fulfill two goals (as described in Carruthers's commentary) becomes more effective with age and experience, as pursuers develop sophisticated ways to perceive how a single behavioral sequence can represent or satisfy two different end-states.
R3.4. The goal standards of society
Some commentators highlighted the societal implications that ought to be considered in light of the target article. People's actions are significantly constrained by social norms and sanctions that in large measure determine the incentive and disincentive value of the various goal pursuits. Behaviors are observed and reconciled within the context of a greater society of individuals, and as described in Washington & Kelly's commentary, the issues raised have implications for the moral institutions that uphold society. The commentary argues that if individuals engage in post hoc rationalizations (and the fact that the information can be used to inform future behavior) or use feedback from behaviors to guide future behavior, these acts suggest that at least some guidance-related subsystem of processes exists within a person that can be held accountable. Thus, even if people lack traditional, centralized psychological cores that can be held “blameable” for their actions, actors can nevertheless be held morally responsible.
Indeed, people have important social-identity relationships with the larger institutions to which they belong, and how they measure and maintain these self-defining relationships is affected by their temporary goals (e.g., threat avoidance, system justification; impression-management). Consequently, understanding the mechanisms of human goal pursuit offers additional insight regarding how people relate to other social institutions. Particularly strong cases of selfish goal influence, such as with drug addictions, may challenge individuals' standing as normative members of society (e.g., Cuzen et al. and Müller & Amato).
Research suggests that goals can affect how a person perceives his or her own place within those institutions, even in cases where the perceptions are self-defeating at the level of the individual person.
One particularly intriguing example of this counterintuitive goal influence is driven by people's need to believe that the institutional system they operate under is fair to them (Kay & Jost). Specifically, people have important goals to imbue their social, economic, and political systems with legitimacy, but often the unconscious pursuit of this goal ironically upholds the very institutional systems that may operate against their individual self-interests and thus oppress them. This dynamic results in a somewhat paradoxical effect in which the most socially disadvantaged members of society are the most likely to support the existing social system (e.g., people of low-income status are more likely to endorse the statement that economic inequality is legitimate and necessary, as compared to those of high-income status; Jost et al. Reference Jost, Pelham, Sheldon and Sullivan2003).
As Kay & Jost note, examining the general structure of goals (for example, their contextual dependence or their reconfiguring abilities over a person's perceptions and behaviors) offers advantages over a focus on goal content (i.e., comparing which of two goals is more primary) for practical reasons as well as theoretical reasons. Knowledge of goal structure can be applied to interventions that alter features of the environment to effect change (which can be a particularly effective intervention; e.g., Johnson & Goldstein Reference Johnson and Goldstein2003). We would only add that it is possible that environmental changes spur behavioral change in part because the intervention leverages the environmentally dependent and selfish quality of goals.
Goal operation can reconfigure one's very experience of the world – as, for example, altering how irresistible desserts appear to the would-be dieters who wish to avoid them, or emphasizing the proximity and instrumentality of some means for one's ends (even if those means are generally socially undesirable and the goal is temporary, as when violence is seen as a means for the status-maintenance goal; Griskevicius et al. Reference Griskevicius, Tybur, Gangestad, Perea, Shapiro and Kenrick2009). Consequently, eliminating the presence of opportunity conditions for unwanted goals decreases the likelihood that those pursuits will be activated or successfully pursued because it impedes the chain of processes that might otherwise perpetuate the unwanted pursuit in a person's corpus of behavior (see above in sect. R3.1).
More important, an environmental change allows for the possibility that another goal may be activated or pursued in that situation; and if we have argued for one thing, it is that goals, when activated, will influence a person in a manner which can be meaningfully understood as selfish. Consequently, once an alternate, equally selfish yet socially valued goal is activated, given successful attainment of the end-state and appropriate circumstances, that goal will become stronger (i.e., its influence over how the individual may become more automatic and reliable).
R4. Conclusions
In closing, we return to some particular themes of the target article. In it we argued for the primacy of the currently active goal as an important influence over human social behavior. Importantly, we suggested that this was the case whether that goal was selected and put into motion by conscious, intentional means or unconsciously by features of the environment, thereby reserving any stance on the specific function of consciousness within goal pursuit.
Extensive recent research and theory on evolutionary social cognition, as well as on infantile motivations and innate social expectations and preferences (e.g., for fairness and equity, as well as in-group favoritism; see Baillargeon et al. Reference Baillargeon, He, Setoh, Scott, Sloan, Yan, Banaji and Gelmanm2013), has expanded the domain of unconscious processes. Importantly, this emerging research also points to new insights regarding how goals come to be capable of unconscious operation in the first place. The evolutionary social psychologists certainly appreciated this point, but other commentaries gave us the impression that an outdated impression of unconscious processes may linger, carrying with it the notion that all goal pursuits start out as conscious and intentional and only become unconscious after considerable experience (i.e., the skill acquisition model of nonconscious processes). A quarter century ago the process of skill acquisition, or sublimation with frequent and consistent experience, was the only way one of us saw clear to the possibility of unconscious goal pursuit (Bargh Reference Bargh, Higgins and Sorrentino1990). Since then, however, the considerable and significant advances regarding evolutionary and early childhood social cognition and motivation call for an updating of any remaining anachronistic, exclusively “conscious first” notions of automaticity and unconscious processes, across psychology.
Finally, in addition to the autonomy of active goal pursuit, we would like to highlight the remarkable plasticity of the cognitive and evaluative apparatus shown by the power of the active goal to reconfigure the mental system to facilitate goal attainment. Other social psychologists have recently remarked on this plasticity (e.g., Cunningham et al. Reference Cunningham, Van Bavel and Johnsen2008; Fiske Reference Fiske2013). It is shown, for example, in the research on alliance formation and how it changes stereotypic evaluations of an out-group member (which was long believed to be a chronic and intransigent tendency) from negative to positive if that out-group member suddenly becomes a teammate (Sherif et al. Reference Sherif, Harvey, White, Hood and Sherif1961). Alliances shift and friends become foes and vice versa. For the sake of adaptability to the shifting and changing social landscape, our goal pursuits can be most successful if the subprocesses they invoke are not slaves to our pasts. We believe such plasticity is an important component of autonomous goal operation, one that is also consistent with emerging research findings on epigenetics, the “affective forecasting” research showing how quickly we adapt to dramatic changes in our life situation, as well as continuing advances in knowledge on the adaptability of the human infant to the particulars of the social and physical environment in which it happened to be born.
In closing, we again express our appreciation for the opportunity to argue for the selfishness of goal pursuit, both in the target article and here in this response. The commentators' contributions challenged and helped us to engage with our model in new ways. Regardless of whether readers are ultimately convinced of our central argument, as long as they have considered existing research in social psychology from this new perspective, we will have attained our own (selfish) goals.
