Ecological Values Theory

Values are Multiple, Heterarchical, and Frustrated

Psychological acts of any complexity are jointly constrained and legitimated by many values rather than by a single value or a fixed hierarchy of values. Any given value cannot be realized in isolation, but only jointly as other values are also realized (Hodges & Baron, 1992). Driving demands accuracy (e.g., direction) and speed (e.g., efficiency), but without attention to safety, justice, tolerance, and many other values, accuracy and speed would quickly be undermined. The working hypothesis of ecological values theory is that values function as a community of criteria, mutually constraining each other so that each of them is realizable in the long run only if the other values are also honored.

Values function this way because they are heterarchically organized. Each value is constrained by all the other values, without control being vested in any one of them (Hodges & Baron, 1992^; Turvey et al., 1978). McCulloch (1945, p. 3), the first to use the term, claimed that “for values there can be no common scale.” If this is so, then values are not scaled in linear hierarchies or fixed, polar oppositions, as is often claimed (e.g., Rokeach, 1973^; Schwartz, 1994); rather, they are interdependent and mutually constraining, supporting each other in a dynamic tension. Each value leads and follows other values at various times during a task, as well as across tasks; hierarchical relations are temporary and reversible. In driving, for example, speed and accuracy may take priority over safety at one place or time, but then reverse at another. These are not simple tradeoffs; sometimes, safety might require increased speed and tighter tolerances, thus requiring greater accuracy (Hodges, 2009).

The heterarchical relation of values leads to skilled activities that can be seen as frustrated (Hodges, 2009). The term was first used in physics (e.g., Sherrington, 2010) to describe systems that are “subject simultaneously to very many different physical requirements that they cannot possibly satisfy fully” (Beek et al., 1992, p. 91). Given that values are in tension and yet function cooperatively, the result is that no one value can be fully realized, in the way that a goal might be achieved (Lewin, 1951, p. 41). Control of action is distributed, a balanced tension between order and disorder. As one becomes more skilled (e.g., juggling; Beek et al., 1992), one comes closer to the edge of what is possible rather than settling into some sort of “comfort zone.” Skilled action of any complexity works to realize a juggled balance of values that is continuously creative rather than being some sort of rule-following procedure that follows the same trajectory every time (i.e., repetitive).

Values are Obligatory and Enabling

Values are ecological; they are the “global constraints on an ecosystem,” or the “boundary conditions that provide for the dynamics of the system and the directedness of animate activity within it” (Hodges, 2007a, p. 590). They are not natural laws, rules, or goals, although they are related to each of these. Values are obligatory. Drivers, for example, are obligated to look where they are going. The obligation is not a social convention, or a natural law (i.e., one is not caused to look where one ought to be looking). Looking is not the goal of driving either; rather, it is a responsibility. It is required because the values of driving (e.g., accuracy; tolerance; safety) cannot be realized otherwise; it is a necessary constraint that makes the freedom of driving possible. Values work together in concert as a set of boundary conditions that create a task-space, or a field of action, that is open to creative activity within those constraints.

In one sense, values restrain, but in another sense, they are enabling. Driving in a way that is both accurate and flexible is demanding, but these demands enable driving to be both fast and safe. Paradoxically, freedom of action emerges from constraints, promoting creativity and robust functionality (e.g., see Anderson, 2015^; Russo, 2015 for examples of enabling constraints enhancing visual systems and social organizations).

Although values obligate responsible action, the tolerances for failing to realize values are often more relaxed than those for ignoring lawful relations. Drivers can sometimes be inaccurate, unfair, or unsafe in their actions, without immediate negative consequences; nevertheless, such actions always threaten to collapse the field of action (e.g., collision; gridlock). The greater elasticity of values relative to laws has tempted some to think that values are really nothing but rules, that is, arbitrary guidelines that are socially constructed and enforced. However, safety, freedom, and truth, for example, do not exist only on someone’s say-so. This does not mean that humans cannot argue about how safety, freedom, and other values are best realized in specific circumstances, but values are real goods that are necessary criteria for evaluating activities. If this were not the case, then productive arguments themselves would not be possible, whether about values or anything else.

Goals are specifiable end states that can be achieved through action, and rules provide specified procedures for reaching those goals. However, goals and rules are too definite and too restricted to legitimate action (Pugh, 1977^; Taylor, 2007) without becoming subject to an infinite regress of “rules on rules” or a series of meta-goals (Martin et al., 1987). Rules always depend on a larger “good faith” context Schwartz, 1986, pp. 259–260) that depends on values (Midgley, 1993). Goals and rules are personal and social constructions, and must be chosen, organized, and justified. This becomes particularly clear when goals or rules need to be revised. Revisions are rarely arbitrary; rather, the changes are justified by appeals to justice, freedom, safety, economy, accuracy, or other values (e.g., Kuhn, 1977).

To highlight the necessity of values is not, however, to shortchange the importance of laws, rules, and goals. Laws function as widely distributed (e.g., universal) space-time stabilities, while rules function as more localized or time-limited stabilities. Both are nested within values-realizing dynamics. Goals can be a means to realizing values, and rules can be a means of realizing goals, all within the possibilities provided by natural laws. It is values—a driving ecosystem that makes speed, safety, accuracy, comfort, efficiency, and other values possible and probable—that provide the criteria for generating rules and conventions, and that hold participants responsible for taking lawful constraints, such as varying road conditions, into account.

Values are Emergent, Self-Critical, and Dialogical

Values have an intrinsic developmental dynamic. The heterarchical tensions among values work together cooperatively to motivate continued perceptual learning and skill development. Values are real but they are not objects waiting to be picked up by eye or hand, but relationships that emerge through engagement over time. There is always more to learn; for example, as one learns to drive, the precision and accuracy of actions increases; later, as one ages, there may be decreases in accuracy and comprehensiveness (e.g., leaving greater gaps between vehicles and driving only during daylight; Hodges, 2007b). In another sense, values themselves are increasingly revealed (i.e., realized) as development and history unfold (Martin et al., 1986).

This cooperative, developing dynamic is what allows values to be self-critical and therefore legitimating. Any goal one chooses to pursue cannot be criticized from the point of view of the goal itself (Pugh, 1977). On the other hand, “every attempt to explicate a value is subject to criticism from the point of view of the value itself” (Martin et al., 1987, p. 68). Thus, someone may come to a truer appreciation of truth, or come to realize that their safety practices are unsafe in some ways not noticed previously. Furthermore, each value is judged by all the values (Hodges & Baron, 1992). Values work because any one value does not function as an isolated criterion, but only in the context of mutually interdependent values (Kuhn, 1977, p. 332). Therefore, in an important sense, values are continually revising activities in self-organizing fashion as those activities unfold.

Participants contribute to the ecosystem and its development, as well as being constituted, guided, and judged (i.e., evaluated) by it. Values define fields of action within ecosystems. Such fields may be viewed as dialogical (or social): Fields of action require ongoing coordination and cooperation, since the fields are jointly instantiated and defined. For example, one person’s ability to converse depends upon what another’s utterances offer as possibilities, and vice versa (e.g., De Jaegher & Di Paolo, 2007); similarly, one driver’s field of action co-varies with other drivers’ fields. All agents are required to commit themselves to realizing values in the system (e.g., driving; conversing) and to trusting that other relevant agents will be (or have been) careful to realize values as well (Costall, 1995).

What Values Afford: Locating, Learning, and Legitimating

Values are not located in any isolated aspect of an ecosystem but are relationships that allow the ecosystem to function, develop, and flourish. To return to the driving example, safety, accuracy, justice, and the other values of driving are relational; they do not exist as a property of drivers, or their vehicles, or the roads they drive on. They are not beliefs; they are actual and potential relations in the world. Realizing values such as safety and accuracy depends on how drivers, vehicles, traffic systems, weather, and many other sets of variables interact. Values do not exist within components of the driving ecosystem, but across relationships within a properly functioning ecosystem as a whole.

The various values that define driving (e.g., speed, accuracy, and safety) are the relationships that make it a skill and thus, allow it to be evaluated and improved. Values are the large-scale requirements (boundary conditions) that indicate when some activity (e.g., driving; reasoning; listening) is, or is not, going right (i.e., contributing to the well-being of the ecosystem as a whole). Without a larger ecology of values, individuals and societies would not be able to critique themselves, including their goals, rules, and uses of laws. Nor would they have any basis for creative attempts at improving their way of life (e.g., Midgley, 1993). Asch (1952) argued that beliefs, goals, and desires depend on values, not the reverse; and De Rivera (1989, p. 13) observed that values “are capable of inducing valences that are not a result of the person’s own needs or will . . . . [They] may even command us to perform some activity not in our personal self-interest.” Since values are relational standards, rather than personal possessions, actions answer to values, and can be judged, both by those acting and by others observing. Individuals can criticize themselves or the groups of which they are a part, and work to improve their grasp of the appropriate relationships, or they can learn from the correction of others who have noted their shortcomings. The same is true of groups; they can critique themselves or learn from the challenges of other groups, to strengthen their practices and adjust their goals and rules. Without such standards that transcend sentiment, desire, and norms, genuine learning and development (i.e., getting better) would not be possible (Köhler, 1966^; MacIntyre, 1981^; Martin et al., 1986^; Murdoch, 1970^; Sabini & Silver, 1982).

Although the actions of individuals and the practices of cultures answer to values, there is space and time for considerable variability in realizing values. There is more than one right way to explore France, to converse with a friend, or to advance scientific knowledge, but not all possible ways are equally good, and some may be bad (i.e., sooner or later they will undermine the integrity of the activity). Researchers are often tempted to make strong judgments about the actions of those they study (e.g., doing x is an “error”). However, sometimes this is done without adequate consideration of the larger context of action and its physical, social, and moral complexities (Carello & Turvey, 1991^; de Wit et al., 2015^; Mousavi & Gigerenzer, 2011). The intention of an ecological approach is to be a hedge against such tendencies, while, at the same time, providing resources for taking seriously the necessity of distinctions such as “right/wrong, meaningful/meaningless, beauty/ugliness, good/evil, [and] love/hate” (Deacon, 2012, p. 544) as crucial ones for psychology as a natural science (Reed, 1996).

We now turn to considering examples of how an ecological values account can contribute to improving research and theory in social psychology and perception-action tasks. These two domains were chosen because they are diverse in their topics and methods, and they show how ecological values theory provides fresh insights and generates novel hypotheses.

A Dissenting View: Speaking Truth to Power

Asch (1951^, 1956) asked individuals to answer relatively simple factual questions about line lengths, except on 12 of 18 trials everyone else in a small group of participants (who were confederates) gave the same incorrect answer before it was the individual’s turn to answer. In designing the experiments, Asch (1952) expected that participants would always dissent when others gave incorrect answers. This was because he believed he had created an all-or-nothing choice between right and wrong; independence was good, and consensus was “malignant” (Asch, 1952, pp. 495–496). Participants should ignore what others said and trust only themselves; their only social obligation was to the experimenter (Asch, 1956). Donald Campbell (1990), whom Asch (1990) thought understood his work better than most, disagreed. He proposed that people in an Asch dilemma should believe that the consensus was correct, but they should report their own personal perspective even though it was probably incorrect. The reason for believing one thing and doing another was a matter of epistemic responsibility; consensus helps to find the truth when people honestly report their individual viewpoints.

Most social psychologists have not followed either Asch’s or Campbell’s line of reasoning; they have shown little or no interest in dissent (e.g., Friend et al., 1990^; Jetten & Hornsey, 2014). Their concern has been to explain why people choose to agree with obviously incorrect answers (Hodges, 2017b). Two general approaches have been taken. One is informational influence: people are not sure about the correct answer, despite what they can clearly see, and think it is safer to agree with the consensus. The other is normative influence: people know the correct answer but are afraid to say it for fear of being scorned or ostracized (e.g., Bond & Smith, 1996^; Cialdini & Trost, 1998), or because they want to ingratiate themselves to others (Deutsch & Gerard, 1955). More recently, this distinction has been posed in terms of goals; people want to be right and to be liked (Campbell & Fairey, 1989). Sticking with the consensus seems an obvious way to gain both goals (Cialdini & Goldstein, 2004).

Asch’s results pose serious challenges for all of these accounts. First, participants dissented two-thirds of the time when there was a unanimous majority, and 95% of the time when the majority was 80% (Asch, 1956). These results undermine normative and informational influence accounts, which lead one to assume that agreeing with others is typical (Toelch & Dolan, 2015); yet dissent dominates rather than agreement. On the other hand, there is far too much agreement with wrong answers for Asch’s account; his zero-tolerance standard for agreeing with the majority was unrealistic (Krueger & Funder, 2004), and his surprise at his results has contributed to the misrepresentation of his research (Ceraso et al., 1990^; Friend et al., 1990). Second, the range of responses was extensive, varying from never agreeing with wrong answers (about 25% of participants) at one extreme to disagreeing occasionally and agreeing mostly (about 30%) at the other extreme. Third, 70% of participants dissented most of the time, illustrated by the median participant who dissented nine times and agreed with wrong answers three times. None of the accounts have addressed adequately the diversity of actions taken by participants; and, in particular, they have failed to address the largest and most typical group of participants that form the middle of the distribution (Hodges & Geyer, 2006). Fourth, few participants seem to have believed that the majority was correct on critical trials, even when they chose to agree with the majority’s answer (Allen, 1965); furthermore, agreement with the majority dropped dramatically when answers were not given publicly to peers (Asch, 1956). These findings are troublesome for Campbell’s (1990) analysis, as well as the informational influence account.

We have taken the time to review key findings of Asch (1956) studies and to summarize three important interpretive frameworks for two reasons. One reason is that Asch’s results have been pervasively and persistently characterized in misleading ways (e.g., Friend et al., 1990^; Harris, 1985^; McCauley & Rozin, 2003). The experiments are routinely presented as a powerful example of the strength of conformity, showing people “blindly going along with the group” (Moscovici, 1985, p. 349). The massive amount of dissent is ignored, largely because social psychologists have unwittingly adopted Asch’s zero-tolerance norm for defining conformity (Hodges, 2004). Dissent and truth-telling have been treated as undeserving of explanation (Friend et al., 1990^; Hodges & Geyer, 2006). A second reason is that the interpretive frameworks themselves have often not been well understood. Asch is widely viewed as trying to demonstrate conformity, not challenge it (Friend et al., 1990). Campbell’s (1990) view is virtually undiscussed in the conformity literature. The informational and normative accounts were advanced to explain agreement with erroneous answers, but some psychologists assume that the concern for accuracy means trusting oneself (not the consensus), and the concern for acceptance means following the lead of others. This garbles the intended meaning of the distinction; it ends up being a restatement of the normative dilemma (i.e., wanting to give one’s own answer, but fearing the disapproval of others). Neither separately nor together can these approaches account for the full range of the results and the complexity of the situation.

Hodges and Geyer (2006) explored an alternative account based on values, which we will try to elaborate and sharpen. Asch’s dilemma is a field of action located within a number of interrelated systems (e.g., perceptual, social, linguistic, and cultural) that is more complex than Asch and others anticipated. Actions are constrained and guided by multiple values that are in tension; these include, among others, truth, social solidarity, justice, and trust. What are these values and how do they work themselves out in this context? Truth seems to be the one value that Asch (1952) had in view, but his view of it was simplistic, as Campbell’s (1990) more complex view reveals. For Asch, being truthful was simply saying what one sees without regard to others. By contrast, Campbell was aware that truth is social, not just personal; truth requires interdependence, not only independence. An ecological account pushes well beyond Asch and Campbell regarding truth. Participants need to be true to the situation as a whole. There are multiple relationships to which one must be true—one should be true to one’s own perceptual experience, but one must also be true to one’s peers, the experimenter, and those beyond the immediate situation to whom one is accountable (e.g., family; friends; scientific community).

The search for how to be true is constrained by other values. Social solidarity is the obligation to care for the communities of which one is a part, so that they too can realize values. It is not to be confused with or reduced to a desire to belong to a group for the sake of being accepted, liked, or esteemed by others (e.g., belongingness; Hirsch & Clark, 2019). The justice that is called for is for all parties to receive what is their due; that is, justice requires attention to and respect for the experimenter, one’s peers, one’s self, and one’s reference groups (e.g., communities of which one is a part). Finally, trust is essential for knowledge; without it, truth cannot be vouchsafed (McLeod, 2020). It is also a crucial form of respect, and it contributes to social solidarity. However, it cannot be reduced either to a simple acceptance of others’ claims or to an assumption of one’s own infallibility. Rather, trust is the ongoing activity of using all available resources to determine what leads toward truth and goodness (Hermes et al., 2018).

While Asch and Campbell tried to focus truth, the informational and normative influence accounts never really get values in their sights at all. There is no concern for truth as something other than achieving the goal of being correct in one’s answers, and possibly being recognized for being right. If accuracy is reduced to a goal and isolated from other values, it does not add up to truth in this situation (see Koriat & Goldsmith, 1996, for a similar problem in memory research). The normative account takes another potential good, namely being esteemed and belonging to a larger community, but without taking account of the correlative responsibilities of contributing to the well-being of the community, to enhancing its capacities for doing justice, engendering trust, and finding truth. Both informational and normative influence accounts are too individualistic and self-focused to capture the social and moral complexity of Asch’s dilemma.

If it were not for values and the tensions within and between them, there would be no dilemma at all. At first, Asch’s dilemma seems to afford no possible way to realize all of these values. How does one say something that honors the integrity of one’s own perception, demonstrates trust in others, does justice to the experimenter’s question, and is sensitive to the situation in which they all are embedded (i.e., a scientific experiment on perception, which is embedded in much larger ecosystems of community and culture)? On any given item, the only choice allowed by the experimenter’s rules was to agree or disagree with others’ answer. However, if participants engaged in heterarchical shifts, sometimes disagreeing and sometimes agreeing with the majority, they could communicate—awkwardly but effectively—their own views, as well as their respect for others’ differing views. The median pattern of nine dissents and three agreements illustrates this possibility: One makes clear one’s dissent, but occasionally one repeats what others have said to make it apparent that one is aware of their views, and that one intends to treat them as trustworthy partners in understanding the situation in which they all find themselves. The apparent inconsistency of agreeing and disagreeing with others truthfully captures the awkwardness and frustration of the larger situation in which they are located.

This analysis suggests that it is morally and socially appropriate to take others' views seriously, even when one thinks them mistaken, in the hope that continued conversation and exploration might reveal the source of the discrepancies. That is, ecological values theory takes the view that people generally understand that finding truth depends on working together. Rather than identifying with others solely for the sake of safety, participants are acknowledging values to which the group, as well as the individual, must answer. Variations over time can be seen not as inept inconsistency or as lapses in conviction or courage, but as ways of speaking truthfully rather than being overpowered either by consensus or by self-confidence. Rather than being trapped in an all-or-nothing choice, participants can complicate their actions over time to reveal more clearly and comprehensively the situation they are in and the values that define it. Hodges and Geyer (2006) proposed that participants were engaging in local errors in order to communicate a larger, deeper truth about the situation. They noted that Asch himself had done the same thing in one sense; in his effort to show the “power” of truth (i.e., that people would dissent from an errant majority; Asch, 1990, p. 55), he had designed an experiment that used deception (i.e., confederates giving false answers).

There were collective resources, as well, for Asch’s participants to create more nuanced, global characterizations of the dilemma and the multiple values constraining their actions. Participants could take advantage of their expectations about how others might respond to the dilemma in deciding what they themselves should do. This is brought into sharper relief by Hodges and Geyer (2006) observation that there were three distinct peaks (12, 9, and 4 dissents out of 12 possible) in the distribution of Asch’s (1951^, 1956) data that can be seen as reflecting three general ways that values can be dynamically balanced in heterarchical fashion as the experiment unfolds. They suggested that the “biological, social, and moral integrity of communities might best be served by the existence of a diversity of strategies that honor the multiplicity of values” (p. 14). Groups are likely to function most adaptively and appropriately over time if members vary how they balance various values in situations as confusing and tense as the Asch dilemma. In such situations, the collective response can do greater justice to the complexity of the situation than any individual’s response.

Claiming that there is no one best way of responding to Asch-like dilemmas is not a recipe for relativism. First, an ecological account rejects the all-or-nothing perspective of Asch and most social psychologists, which accuses participants who agree even a single time (while dissenting 11 times) of having conformed. Second, it suggests that Asch’s participants were not, in all likelihood, either conforming or acting independently. Rather, it is more likely that they were engaging in creative, cooperative action to engage in truth-telling, social solidarity, trusting, and justice. Third, it points to the possible dangers of the pure dissent which Asch (1951) admired. Always disagreeing with others runs the risk of communicating an arrogant and dismissive attitude that undermines the possibility of finding consensus. Dissent implicitly appeals to some common ground; without social solidarity and trust, there can be no fruitful dissent. Fourth, epistemic and ethical evaluation of any given individual or group needs to be scaled over time and across tasks, not simply based on isolated, momentary decisions. In all of these ways, and more, ecological values theory differs from accounts that assume people are always self-centered and defensive (e.g., fearful of being wrong, or being different), or accounts that focus only on goals or rules to explain actions.

The Asch (1956) dilemma is an example of a frustrated system. It is a challenge to the integrity and flexibility of ecosystem relations, but Asch’s results suggest that the observed variability in participants’ behavior was truthfully capturing the larger tensions of the situation, due to participants’ efforts to realize multiple values, heterarchically related, that emerge as the situation unfolds.

Alignment, Divergence, or Dialogue? How Values Increase Comprehensiveness and Coherence

The values-realizing analysis of the Asch (1956) dilemma has implications for many other issues in social coordination, social learning, and social anthropology. We can mention only a few. First, standard approaches to Asch-type situations predict that people will conform more to friends than to strangers (Deutsch & Gerard, 1955); however, an ecological account predicts the opposite, at least in many truth-telling situations (Hodges & Geyer, 2006). Evidence reveals more dissent, not less, among friends (Matsuda, 1985^; McKelvey & Kerr, 1988^; Takano & Sogon, 2008). Friends can challenge friends more bluntly about errors or dangers because trust and social solidarity are well established (Hodges & Packer, 2016^: Packer & Miners, 2014). This suggests that the relationship among values is not primarily in terms of tradeoffs; rather, it is cooperative and heterarchical.

Second, Asch (1956) thought he had created a dilemma about truth that invited dissent. What would happen if one were in a similar situation except that it invited agreement with one’s peers? Hodges et al. (2014) created a situation in which two people (A and B) were in good positions to answer correctly, while a third (D) was not and could only guess (see Figure 1). Would the participant in the position of ignorance follow the lead of A and B when they answered correctly? Hodges et al. made the counterintuitive prediction that participants would intentionally give disagreeing, incorrect answers a significant proportion of the time due to the same sorts of dynamics that led participants in an Asch situation to sometimes give agreeing answers. The hypothesis was confirmed: about 30% of the time participants made up their own wrong answer, a speaking-from-ignorance effect. Participants are in a dilemma much like Campbell’s analysis of Asch: They can say what they believe is correct or they can indicate what they see from their position. They feel obligated to trust others and to speak truthfully, which leads them to sometimes choose to create wrong answers that reveal the larger truth of the situation: that they are in a position of ignorance. When participants were primed to be particularly sensitive to the demands of truthfulness, the frequency of disagreeing, incorrect answers increased, even when there were financial incentives for correct answers. In both the Asch dilemma and the speaking-from-ignorance dilemma, participants sometimes give answers they know are incorrect, so that they can communicate the larger truth of the situation and realize values more fully. People are motivated by more than goals (e.g., to be correct and to be agreeable).

Open multimedia modal

Third, ecological values theory leads one to expect that phenomena such as mimicry, synchrony, conversational alignment, and social learning are likely to reveal dynamics that are similar to those revealed in social dilemmas. Rather than a simple story of convergence, matching, and alignment, a more complex dialogue between convergence and divergence, between alignment and nonalignment, will be found. An ecological values account predicts that these coordinative practices will resist explanation solely in terms of laws, rules, or goals, and will be characterized, not by dominance of one partner in the dialogue, but by heterarchical reversals in which the activities of leading and following are joint and reciprocal. Although much further work is needed, there is considerable evidence in favor of pursuing this line of inquiry (e.g., Hodges, 2017b).

For example, mimicry is often described as if it were automatic, unintentional, and irresistible (e.g., Chartrand & Lakin, 2013); however, research indicates that it is context-dependent and selective (e.g., Wróbel & Imbir, 2019). It is dramatically reduced or absent in situations where trust or social solidarity is in doubt (e.g., Stel et al., 2009^; Weyers et al., 2009^; Yabar et al., 2006). Synchrony reveals a similar pattern of dynamics (Fusaroli et al., 2014). It is undermined by disagreement or by one of the partners acting irresponsibly in the larger context in which actions need to be coordinated (Miles et al., 2010^; Paxton & Dale, 2013). Engaging in joint action with another requires dialogue, not mirroring (Reddy & Uithol, 2016). In spontaneous conversations, participants repeat each other’s syntactic constructions less often than expected by chance (Healey et al., 2014), and partners often create linguistic conventions on the fly that are idiosyncratic and flexible, becoming more complementary than convergent over time (Mills, 2014). Overall, it appears that selective alignment yields better joint actions (e.g., greater accuracy in a coordinative task) than indiscriminate alignment (Fusaroli et al., 2012). In all these cases, an ecological values-realizing account encourages moving to a higher, more dialogical, level of dynamics rather than focusing only on convergence/matching on the one hand, or divergence/asymmetry on the other.

Research on social learning paints a similar picture. Theorists in social anthropology believe humans’ tendency to follow the lead of other humans is very strong, especially compared to other apes (e.g., Richerson & Boyd, 2005^; Haun et al., 2014). However, people often lean on their own experience rather than adopting technologies and methods most frequently used by others (e.g., Efferson et al., 2008^; McElreath et al., 2005^; Eriksson et al., 2007). People are particularly unlikely to move to majority options when they are motivated to be accurate (Claidière & Whiten, 2012), and some theorists have claimed that a conformity bias cannot function as a stable adaptive strategy (Eriksson et al., 2007).

Research on social learning in children also points beyond conformity and toward values-realizing dynamics. Children are often viewed as easily influenced, gullible even, yet many recent studies show how committed to truth children are, and how sophisticated they are about who to trust, and when, in assessing others’ claims and actions (see Hoehl et al., 2019^, Koenig et al., 2019, for reviews). They take truth, justice, social solidarity, and trust into account in ways that have remarkable integrity and flexibility, even by adult standards. Two studies placed children in Asch-type dilemmas and found that young children overwhelmingly dissented—most of them, every time—although the authors chose to emphasize the minority of agreeing answers (Corriveau & Harris, 2010^; Haun & Tomasello, 2011). Similarly, on practical action tasks (e.g., opening a puzzle box) preschoolers mostly do not adopt the consensus view or procedure when it is unsuccessful or implausible, even when they identify themselves as part of the group (Schillaci & Kelemen, 2014^; Wilks et al., 2015). However, children are also sensitive to social solidarity and trust (e.g., Corriveau et al., 2009). Children and adults show flexibility and sensitivity in how they learn from others; they engage dialogically with others and with their surroundings to realize values.

Methodologically, ecological values theory predicts that similar patterns of convergence–divergence, symmetry–asymmetry, and stability–flexibility will emerge across a wide array of social coordination tasks if researchers can enlarge their explanatory canvas while noticing details of actions more closely (Reddy, 2019). Ways this can be done include (1) looking carefully for signs of selectivity and context-dependency in their phenomena (e.g., Over & Carpenter, 2012^; Wróbel & Imbir, 2019); (2) integrating findings from across various social coordination domains, rather than remaining in research silos (e.g., Hodges, 2017b^; Whiten, 2019); and (3) acknowledging the time dimension in research and moving questions and analyses to higher level dynamics that are dialogical rather than focusing only on convergence/matching on the one hand, or divergence/asymmetry on the other (e.g., Hodges, 2017a^; Keupp et al., 2018). To give one example of the surprises in store, consider that social psychologists generally treat conformity in individualistic terms and see it as a danger; by contrast, social anthropologists treat it in terms of collective patterns and see it as an essential aspect of cultural evolution (Claidière & Whiten, 2012). Ecological values theory suggests that both of these stories are too simple; putting them in dialogue will likely yield a richer, truer account.

Driving for More Than Accuracy

Driving was used earlier to illustrate the necessity of values, but how does paying attention to values affect the actual conduct of research? One of the most common weaknesses of many studies of driving and related sorts of skills has been an excessive focus on a single value, treating it as if it were a goal (Hodges, 2007b). For example, how accurately can a driver stop a vehicle from a designated target, such as another car? Often, in such conditions, there is precise optical information that specifies time-to-contact (TTC; Lee, 1976^; Rock et al., 2006); thus, drivers should be able to make highly accurate judgments. However, drivers persistently underestimate TTC and often women do so even more than men, both of which researchers have struggled to explain (e.g., Caird & Hancock, 1994^; Schiff & Oldak, 1990^; Stoffregen & Bardy, 2001). If there are natural laws that provide the right information for acting appropriately, why do drivers’ actions not reflect that fact (Harrison et al., 2016)?

From the perspective of values-realizing, claims that males are more accurate than females (e.g., McLeod & Ross, 1983) could just as easily have been described as “females are safer than males.” Seeing how close one can come to a stopped vehicle may be more accurate, but it is also less safe (Hodges, 2007b). Likewise, sensitivity to safety can also explain why TTC is usually underestimated: “Erring on the safe side” will generally lead to much better driving than “cutting it as close as possible.” Drivers also take into account their own abilities and proclivities in making TTC judgments. Older drivers tend to underestimate TTC more than other drivers, as do drivers who have impaired vision (e.g., Rusch et al., 2016). Although all of these can be interpreted as supporting the importance of safety in action, researchers find it difficult not to equate inaccuracy with a reduction in safety, even when the observed biases would generally contribute to safer driving (Hancock & Manser, 1997^; Rusch et al., 2016). From an ecological perspective, attending to only one value, as if it were a goal, leads to bad driving; the same is true of research that treats one value (e.g., accuracy) simply as a means to another (e.g., safety). Focusing only on goals, whatever the activity, neglects the heterarchical, frustrated nature of value-realizing action.

Working within an ecological framework, Fajen (2008b) has shown that even when a single dimension of driving (i.e., braking) is chosen and the experimental task focuses on a single value (i.e., accuracy), there are many other constraints that guide online control of action. The methodological move that opened a larger space for understanding various aspects of driving was to introduce many more variables into the “braking equation” than are usually included in driving experiments. This increased variability stresses the driving ecosystem; however, these increased demands also help to reveal the diversity of values that enable and guide driving. For example, even if drivers are not told about a sudden change in brake strength or dynamics, they quickly recalibrate their actions in light of what they pick up from optical variables, making appropriate alterations to their pedal behavior (Fajen, 2007b^, 2008a). Drivers’ selection of actions and ongoing control of them are not determined by any one value in a way that can be designated as a “law of control.” “There is no such thing as a preferred state in which the actor should strive to be at each point in time” (Fajen, 2007a, p. 406), which suggests that driving is not exclusively a goal-oriented or rule-following task; instead, it is characterized by an ongoing state of tension among many values.

Some artificial intelligence researchers have explicitly argued that autonomous driving systems will require value-based argumentation rather than following fixed rules, if they are to be functional in the complexity of open environments (e.g., Bench-Capon & Modgil, 2017). Whether that will be possible remains unknown. In any event, narrow-gauge approaches to driving can easily miss the complexity and richness of the skill, generating experimental anomalies and practical puzzles. Locating driving in a larger ecology of values can help to avoid such oversimplifications.

Carrying: Beyond Goals and Rules

Dozens of times during a day people thoughtlessly or deliberately pick up some item and move it from one place to another. In either case, carrying is an intentional act. It is to see where something is, and then to transport it to a place where it ought better to be; thus, carrying is a values-realizing action. The act of carrying is ordinarily intended to increase safety, justice, freedom, beauty, or other values (although it does not always succeed). Frequently, carrying is goal-directed, but the motivation and justification for any move or set of moves depends on values, not just goals.

Carrying is a physical act, and thus, it is not surprising that weight affects the dynamics of lifting and carrying. Heavier weights require different movements than lighter weights, and those differences can be perceived by observers who can see only point-light (kinematic) displays of limb movements. Even when those carrying the load are instructed to engage in deceptive action, trying to make a heavy weight appear light or vice versa, properly attuned perceivers can identify to some extent both the true weight and the deceptive intent (Runeson & Frykholm, 1983), even when skilled actors and mimes are employed (Mark, 2007). Hodges and Lindhiem (2006) proposed that moral and social relationships would alter kinematic patterns of people carrying items in physical tasks that were challenging. Parents carried their child, or an equally weighted sack of groceries or trash across a series of uneven steps with gaps between them, holding them in the same way. Observers who saw point-light films of parents’ movements alone, with no knowledge of anything being carried, rated child trials as being more careful than groceries and trash trials. These results suggested that moral obligations of parents to children altered their carrying movements in ways identified as being more careful. Many values constrain walking and carrying, among them comfort, speed, safety, accuracy, stability, variability, and even social solidarity (e.g., parent carrying their child). Research that focuses only on measures of energy consumption (e.g., comfort) or stability are likely to be too limited to be adequate.

One important question that emerged from this research was: What variations in movement patterns indicate that one is acting carefully? Neither psychology nor kinesiology has addressed this question directly. Existing studies have focused on safety and stability, often defined in terms of the goal of “not falling” (Bruijn et al., 2013^; Hamacher et al., 2011). Whatever safe, stable walking is, it is unlikely to capture the meaning of carefulness because it does not address the nature of what is being carried and how it is cared for as it is being carried. Almost certainly, it will require dexterity, the capacity to generate novel action patterns that adaptively address situational demands as they emerge in real-time (Bernstein, 1996^; Harrison & Stergiou, 2015) as well. Values constraining careful carrying are not located only in the object carried, or in the steps, or in the parent, or in walking per se; rather values are distributed across the relationships that make up the ecosystem. It is important to note that the heterarchical nature of values means that children will not always be carried more carefully than groceries or other items, although virtually all parents indicate that they believe that their child is more valuable. There are variations between tasks, between moments within tasks, and there are individual differences among parents as well. For example, a child carried frequently, and who cooperates in being carried, may produce kinematic patterns that are rated as less careful than “dead weight” (e.g., groceries) carried infrequently. However, at any moment in the stepping, a disturbance can require a sudden change in the dynamics, reasserting the priority of safety over comfort, or accuracy over freedom. Although available evidence is limited, it favors ecological value theory’s claims: that values are not localized; that there is not a fixed hierarchy of constraints; and that conscious intentions and awareness are not sufficient to characterize the values constraining actions (Ventura & Hodges, 2021).

Finally, we consider recent, influential research on carrying to illustrate how attending to the broader, value-realizing context of actions can clarify and contextualize studies framed only in terms of goals, rules, and laws. Rosenbaum et al. (2014) presented participants with a simple, two-part goal task: Walk down a pathway which had pails stationed alongside, reach and pick up a pail with its contents, and carry it to a target destination. Besides varying in weight, pails were positioned closer to or farther from the target destination and on the right or left side of the pathway. The key rule was to do “whatever is easier.” Invoking Hull (1943) “law” that actions requiring less effort are preferred unless there are rewards that make the greater effort worth it (Potts et al., 2018), Rosenbaum et al. expected participants to pick up the pail closest to the target destination, since that would mean the carrying action would be shorter (thus, less effortful). However, participants mostly did the reverse. To explain this “suboptimal” action, they proposed a post-hoc hypothesis framed in terms of goals; participants wanted to achieve a sub-goal (i.e., picking up a pail) of the larger goal task as quickly as possible rather than waiting to pick it up.

An ecological values account approaches the situation differently. Researchers (Potts et al., 2018^; Rosenbaum et al., 2014) began their research assuming that one value (i.e., comfort), treated as a goal (do what is “easier”), would yield a lawful outcome (i.e., Hull’s least effort principle). When that did not work out as planned, they adopted another isolated value (i.e., speed) as a goal, treating it as fundamental (i.e., do part one of the task as quickly as possible).

In doing so, they traded one value for another. However, their results indicate that neither of these values is the fundamental constraint; sometimes, participants chose to carry heavier weights farther; other times, participants picked up weights later rather than sooner (Potts et al., 2018). The ecological claim is that there is no one fundamental constraint. If loads were heavier and the walking distance was longer, comfort (effort) would likely become a much more prominent constraint than doing a sub-goal task sooner rather than later. If the load being carried were more important (e.g., a child), it is not likely that comfort or the desire to get started sooner rather than later would weigh heavily in action choices and persistence. A methodological implication of this analysis is that researchers need to stress ecosystem relations more in order to discover the diversity of constraints that are relevant, and to pay close attention to the diversity of actions that occur within and across tasks.

Acknowledgements

The authors are grateful for helpful comments on earlier versions of this paper provided by Alan Costall, Laura Cuijpers, Adrian Frazier, Martin Fultot, Jonathan Gerber, Jason Gordon, Henry Harrison, Steven Harrison, Keith Holyoak, Blair Johnson, Maria Jarymowicz, Wim Pouw, Janusz Reykowski, Vasudevi Reddy, Nicole Rossmanith, and the reviewers of this journal. The article is dedicated to the memory of James E. Martin.