Shared-Reality Development in Childhood

Defining criteria of shared reality

As proposed by Echterhoff, Higgins, and Levine (2009), shared reality is the experience that you have an inner state about something (e.g., a feeling or belief or concern about something) that is shared by others (a person or group)—not just overt behaviors. One person mimicking the behavior of another is not creating a shared reality. When discussing humans as a species, Tomasello and his colleagues (e.g., Callaghan et al., 2011; Tomasello, 2014) emphasized that humans share their inner states, such as their intentions and experiences, rather than just observable behaviors. In addition, for shared reality to occur, the common inner state must be about something, such as a common attitude toward something. It is not a shared reality if two people happen to have in common that they are both in a bad mood. It is a shared reality, however, if they recognize that they are both angry about what a particular person just did or that they both find interesting what’s happening on the branch of a particular tree.

For shared reality to occur, people must experience that they have an inner state (e.g., feeling, belief) about something that is in common with the inner state that someone else is experiencing. The experience of this commonality often occurs during interpersonal communication. In their early article on the nature of shared reality, Hardin and Higgins (1996) emphasized the importance of social verification exchanges in the creation of shared reality. As William James (1890/2007) noted, “You accept my verification of one thing, I yours of another. We trade on each other’s truths.” When another person verifies what you feel or believe about something, you experience that you and the other person have a common feeling or belief about something—a shared reality. Hardin and Higgins (1996) also argued that this shared reality makes you perceive your feeling or belief as being more valid and reliable—it is experienced as objective reality. Regarding this latter point, it should be noted that Heiphetz, Spelke, Harris, and Banaji (2013) suggested that this would be truer for beliefs that are considered facts than for ideological beliefs, such as religious beliefs, which children and adults treat as being not only like facts but also like preferences. With facts there can be only one right answer, whereas with preferences there can be more than one right answer. Ideological beliefs can fall in-between.

What is “shared” in shared reality? Echterhoff, Higgins, and Levine (2009) noted that the term “shared” has more than one meaning (see Cannon-Bowers & Salas, 2001; L. Thompson & Fine, 1999), and each meaning contributes to a fuller understanding of what shared reality is and is not. There is the meaning of “shared” as “communicated or disclosed to others,” as when people share what they find interesting about a person or event. This highlights the importance of interpersonal communication to the possibility of creating a shared reality. “Shared” can also mean “divided up into portions,” as when there is a division of labor on a common task with different individuals taking on different roles (Echterhoff, Higgins, & Levine, 2009). This definition highlights the potential importance of shared reality for joint tasks or organized games that require team members to have a common goal and coordinate their distinct roles.

As will be seen, the development of shared reality includes both of these kinds of “shared,” and they function as shared realities when interaction partners verify each other’s feelings, thoughts, or role expectations. When such social verification is included, “shared” means “held and experienced in common,” and all the criteria for having a shared reality is met (Echterhoff, Higgins, & Levine, 2009) because it explicitly includes the idea that the commonality is not only present but is also experienced as being present (see Bar-Tal, 2000). This finding is important because two animals could actually feel the same way about something, such as both fearing a snake, but unless they recognize that both feel the same way about the snake, there would be no shared reality. Perceiving that both feel the same way creates a shared reality. For example, when we say about a couple, “Anna and Joe share the enjoyment of Indian cuisine,” we imply that they experience the commonality of their positive feelings about this kind of food. They have a shared reality about it. As we will see, sharing feelings occur early in children’s development of shared reality.

What is “reality” in shared reality? The “reality” part of shared reality is also an experience. It does not refer to actual reality but to the experience of sharing a reality about something with another person or persons. What changes developmentally is that in different phases a new mode of sharing about something emerges, such as an infant girl sharing with her father her interest in something that is happening in front of them. Although from a scientific perspective what is experienced as real about something is a subjective experience, once it is experienced as being shared with another person it is now experienced as being objective (Hardin & Higgins, 1996)—what really is the case? For example, now that Daddy is also interested in what is happening in front of us, it is really interesting.

As we will see, this effect from creating a shared reality has significant benefits for humans, as well as costs, because what often matters is simply that people, independent of the evidence, experience their shared realities as being objectively real (see Brickman, 1978), although this can depend on the nature of the shared reality. Even children, for example, distinguish between the objective reality of a shared imaginary or pretend event or character versus an everyday life nonpretend event or character (Harris, 2000; Taylor, 1999). In addition, as noted earlier, shared preferences are not treated as being objective like other shared beliefs. Nonetheless, everyday beliefs that are shared with others can be treated as objective facts even when the evidence does not justify it. Indeed, the motivation to achieve the experience of shared reality is so strong that people can prefer less over more scientifically accurate knowledge if less scientific knowledge better supports a shared reality (see Higgins, 2012), as can occur with political and religious ideologies (see, e.g., Jost, Ledgerwood, & Hardin, 2007).

What motives underlie wanting to create shared realities? Like human adults, human children are motivated to create shared realities with others. The purpose of this article is to reinterpret some key phenomena in the previous developmental literature in terms of this general motivation and to describe developmental changes in the kinds of shared realities that children create with others. The motives of older human children are different from both nonhuman animals and human babies. Central to this difference is the motivation to create shared realities with others (Higgins & Pittman, 2008). And, notably, the developments in shared reality should not be characterized as improvements because a new shared reality can have costs as well as benefits, including the development of language which can also have costs (see, e.g., Harter, 1999; D. N. Stern, 1985; for other costs of development, see Leahy, 1985).

What then is involved in the human motivation to create shared realities? Although there could be several motives, two critical ones are epistemic and social-relational motives (Bar-Tal, 2000; Echterhoff, Higgins, & Levine, 2009; Hardin & Higgins, 1996; Jost et al., 2007). The epistemic motive is the motive to establish what is real, correct, and right (for a review, see Higgins, 2012) and the need to achieve a valid and reliable understanding of the world (Hardin & Higgins, 1996). If shared reality is about something in the world, then creating a shared reality serves the epistemic motive of experiencing one’s understanding of something in the world as being objective—what something really is. This in turn increases one’s sense of predictability and subjective efficacy in dealing with the world (for a review, see Higgins, 2012).

The social-relational motive is the motive to affiliate and feel connected to others; that is, to belong. The psychological literature has described several positive consequences of feeling connected to others, including emotional well-being and experiencing nurturance and security (e.g., Baumeister & Leary, 1995; Diener & Seligman, 2002). Wanting connectedness can also increase motivation to identify with a positively valued group, which can in turn enhance self-esteem (e.g., Hogg & Abrams, 1988; Levine & Kerr, 2007).

Evidence of epistemic and relational motives for creating shared reality

The importance of both epistemic and social-relational motives to the creation of shared reality has been demonstrated in research on how communicators’ own memory is affected by their tuning or tailoring their message about a target person to match their audience’s attitude toward that person. In the original study (Higgins & Rholes, 1978), undergraduate participants in the role of communicators were asked to describe a target person, Donald, based on a short essay that described his behaviors. Because the behaviors were ambiguous, they could be labeled either positively or negatively; for example, a behavior could be labeled as either “stubborn” or “persistent,” or, for another behavior, as either “confident” or “conceited.” The participants were given a referential communication task in which they were told to describe the target person (without mentioning his name) to another person, the audience, who knew the target. In an offhand manner, the participant communicators were also told that their audience either liked or disliked the target. On the basis of the message description, the audience would try to identify the target person from among a set of other possible persons.

This study and many subsequent studies (see Echterhoff, Higgins, & Levine, 2009) found that the communicators tuned their message to suit the attitude of their audience, using more positive labels when the audience liked the target and more negative labels when the audience disliked the target. Later (as much as 2 weeks later), the participants were asked to recall as accurately as they could, that is, word for word, the information about Donald’s behaviors that was contained in the original essay they had been given. These studies found that the communicators’ own memory for the original essay information was distorted to match the evaluative tone of their message rather than accurately matching the ambiguous tone of the original information—the “saying-is-believing” effect (Higgins & Rholes, 1978).

Subsequent studies have shown that in this standard in-group version of this communication situation (i.e., communicators and audiences from the same college), communicators create a shared reality with their audience and, by so doing, their tuned message becomes the “truth” about the target person that influences their reconstructive memory of the target’s behaviors (see Echterhoff, Higgins, & Levine, 2009; Higgins, 1992). Rather than representing how the target actually behaved, their recall represents what the participants believe happened, consistent with the shared reality about the target that was created. How do we know, however, that the recall distortion is not simply due to the message itself being distorted because it was tuned toward the audience’s attitude? We know that experiencing the message as a shared reality is critical because later research revealed that when communicators have goals other than shared reality for tuning their message toward their audience’s attitude, such as just being polite, or just having fun, or wanting the audience to choose them later for a money-making task, then the saying-is-believing effect disappears even though there is at least as much message tuning (e.g., Echterhoff, Higgins, Kopietz, & Groll, 2008). Moreover, Echterhoff et al. (2008) also showed that whether the saying-is-believing effect occurred or not depended on whether communicators trusted their message as reflecting the target person’s real personality (i.e., message trust as mediator). In addition, their trust in the message was stronger when the goal was to create a shared reality rather than just to be polite, have fun, or make money.

Using this same basic paradigm, other studies have shown the importance of communicators’ relational and epistemic motives for creating a shared reality (for further examples, including studies using other paradigms, see Echterhoff, Higgins, & Levine, 2009). As an example where both motivations matter, Echterhoff, Higgins, and Groll (2005) compared what happens when the communicator who is a student at the University of Cologne communicates to either an in-group member (another Cologne student) or an out-group member (a student at a nearby vocational school for hairdressers). They found that communicators tuned their message toward their audience’s attitude not only for the in-group audience but also for the out-group audience, probably out of politeness for the out-group audience. Communicating to an out-group audience is an interesting case because people do not feel connected to out-group members and also do not trust them as much as in-group members (Hogg & Abrams, 1993; Kruglanski et al., 2005; Kruglanski, Shah, Pierro, & Mannetti, 2002). Therefore, both social-relational and epistemic motives are weaker for an out-group audience and would reduce the creation of a shared reality. Indeed, despite there being significant audience tuning for the out-group audience, the saying-is-believing effect disappeared.

In another study, Echterhoff et al. (2005) examined the importance of epistemic truth by giving ostensible feedback on whether the audience was, in fact, able to identify the target person correctly based on the communicator’s message. When communicators received success feedback, the saying-is-believing effect was found as usual, but when communicators instead received “failure” feedback, the saying-is-believing effect disappeared. If the message was unsuccessful, how can you trust its epistemic truth? Finally, another study demonstrated that even when there is epistemic trust in an audience’s attitude, not feeling connected to that audience can interfere with creating a shared reality. Echterhoff, Lang, Krämer, and Higgins (2009) examined personnel assessment in a business organization. Student communicators described an employee to either an equal-status audience (a student temp) or a high-status audience (a company board member). The high-status audience clearly had epistemic expertise regarding personnel, but, perhaps not surprisingly, the communicators felt less connected to the high-status audience compared with the equal-status audience (less ready to affiliate or be close). Once again, although audience tuning occurred for both the high- and the equal-status conditions, the saying-is-believing effect disappeared for the high-status audience.

In sum, the communication literature provides evidence that relational and epistemic motives contribute to the motivation to create shared realities with others. It should be emphasized, however, that when these motives are fulfilled by creating shared realities, it does not produce only benefits for people. After all, the saying-is-believing effect itself involves memory distortions from creating a shared reality. Thus, in addition to discussing the critical benefits of creating different kinds of shared reality in childhood, the potential costs will also be discussed.

I also need to emphasize that this communication literature illustrates how individuals communicate with others to create a shared reality—they are not simply expressing a shared reality that they had to begin with. Indeed, in the original Higgins and Rholes (1978) study, in another condition where communicators prepared their message for an audience they knew liked or disliked Donald but ended up not actually delivering their message, there was no shared-reality effect on their recall of Donald’s behaviors. Without delivering the message, no shared reality was created. More generally, people use communication to learn what others know that they do not know—they create a shared reality by learning from others. This is certainly true for children as well, as evident in the research reviewed in Paul Harris’s (2012) excellent book, Trusting What You’re Told: How Children Learn From Others.

Finally, I also need to highlight that my purpose in describing this literature on the saying-is-believing effect is to provide some evidence for the role of relational and epistemic motives in the creation of shared realities with others. I am not claiming that the kind of audience tuning that was performed by the undergraduate participants in these studies would be performed by young children. Indeed, my own earlier research (Higgins, 1977) found that, compared with 13-year-olds, even 9-year-olds are relatively poor at tuning their message to the inner state of their audience because they too often assume that what they know their audience also knows, such as giving directions to a stranger as if the stranger knows the local landmarks like they do (see also Glucksberg, Krauss, & Higgins, 1975). Notably, such “egocentric” communication need not prevent children from creating a shared reality with others because often they are correct when they assume that what they know others also know.

Before beginning my discussion of the major changes that occur during childhood in the kinds of shared realities that children create with others, I also need to be clear about my objectives in this article and what I am and am not claiming about childhood development. There are historical classic descriptions of milestone changes during childhood in both the intellectual (cognitive-structural) development literature (e.g., Piaget, 1932/1965; Selman & Byrne, 1974; Turiel, 1983) and the socioemotional literature (e.g., Erikson, 1950/1963; S. Freud, 1938; Loevinger, 1976; Sullivan, 1953). Since these early descriptions, research evidence for and elaborations of such changes have continually grown. The changes described involve phases of development, such that in each phase qualitatively distinct ways of interacting with others emerge. Some phase changes have been emphasized more in the intellectual development literature, whereas others have been emphasized more in the socioemotional literature. Both kinds of phase changes will be discussed in this article.

I discuss four phase changes during childhood, defined as birth to roughly age 13 years. I propose that a new mode of shared reality emerges during each phase and this new shared reality has significant self-regulatory and social regulatory consequences. I am not claiming that what happens is either the appearance of sudden, discontinuous stages or constant continuous growth. My general position on the nature of this development is basically the position suggested by Flavell (1982) and Damon (1983). As Damon (1983, p. 378) stated, development involves “ordered sequences of qualitatively distinct modes of knowing”—in this case, qualitatively distinct modes of sharing reality with others. What I claim is simply that the sequence of emerging shared-reality modes is ordered in that Phase 1 occurs before Phase 2, Phase 2 before Phase 3, and so on, without skipping a phase. Although I will suggest typical age ranges in which a new phase occurs, for any particular child it could occur earlier or later than the range I mention, especially for later phases that can vary across cultures and across historical periods (see, e.g., Higgins & Parsons, 1983; Rogoff, 2003).

I am also not suggesting that the changes I describe are the only major changes that occur during childhood development. Clearly, there are other major cognitive, social-cognitive, and socioemotional changes that occur during childhood as well as maturational and biological changes. These other changes are very important for understanding the nature of childhood development. My only claim is that there are also important changes in shared reality that occur in childhood development. Thus, it is not my objective to discuss all the changes that occur in childhood development and certainly not to provide a general theory of development during childhood. Instead, the purpose of this article is to review some key phenomena involving changes during childhood and attempt to capture them within the single conceptual umbrella of changes in children’s shared realities with others.

It should also be noted that because this article is concerned specifically with those ordered sequences of qualitatively distinct modes of sharing reality with others rather than with developmental changes in general, I do not review other important changes that are happening developmentally between the phases that I discuss. Thus, age gaps that are not specifically discussed in the article, such as the gap between ages 6 and 9 years, do not imply that nothing is happening developmentally during that period. Important developmental changes are certainly occurring. What I am attempting to identify is generally when a qualitatively distinct mode of sharing reality with others has clearly emerged in childhood. Other important developmental changes are occurring between and within the phases that I discuss.

I also must emphasize that I am not claiming that children leave behind a mode of shared reality from an earlier phase when a new mode of shared reality emerges in a later phase. It is not like giving up crawling as a mode of moving from place to place for the new locomotion mode of walking. Once a mode of shared reality emerges, such as the early mode of sharing feelings with others, it remains a part of the lives of children and adults, while continuing to develop and transform. Indeed, I have described in an earlier article (Higgins, 1991) how the emergence in preschool children of identification with significant others’ goals and standards for them continues to develop before becoming the adult version of self-regulation and self-evaluative responses to success and failure to meet goals and standards. Although the major purpose of this article is to review the emergence of each new mode of sharing reality with others, rather than to discuss how each new mode then continues to change with development, the latter is an important developmental story because it reflects both continuity and a kind of discontinuity. There is continuity because once a mode of shared reality emerges, such as shared feelings, it continues to function and develop. But there can also be a kind of discontinuity because the later emergence of a new shared-reality mode in a subsequent phase can introduce a qualitatively new form of the earlier shared-reality mode. Because of the importance of this developmental story, I will discuss it after the different shared-reality phases have been described.

Similarly, I recognize that there are significant development changes that occur after the phases that I describe. My discussion of developmental changes in shared reality is restricted to the childhood period from birth to age 13 years. Again, there are certainly important developmental changes that occur after this during adolescence and adulthood. In addition, these changes include what could be construed as new forms of sharing reality with others. However, how to characterize what new mode of shared reality emerges in adolescence was not clear to me—especially how to distinguish it from the final childhood phase of shared coordinated roles that are discussed. For this reason, I chose to limit my review of developmental changes in shared reality to those that occur during childhood.

Finally, I should note that the major objective of this article is to identify the qualitative changes in children’s sharing reality with others. It is not the aim of this article to describe the mechanisms that underlie these changes, whether from maturational changes, cognitive-developmental changes, or social-learning changes. Social-learning changes would include moving through social-life phases, such as moving from a home-centered initial phase to a later schooling phase and transitions across different school grade environments (e.g., Eccles & Roeser, 2003; Higgins & Parsons, 1983; Pianta, Rimm-Kaufman, & Cox, 1999). There is every reason to believe that all of these sources of change are part of the story and they interact with one another (Damon, 1983; Harter, 1999; Higgins, 1991). However, although trying to account for the mechanisms underlying specific shared reality changes is not central to this article, it is an important issue nonetheless. Given this, I will generally consider and discuss this issue after the phases of shared-reality development have been presented. In addition, during my discussion of the phase changes, I will discuss some cognitive and social-cognitive changes that contribute to these changes.

I will describe four phase changes: Phase 1 (6–12 months) shared feelings; Phase 2 (18–24 months) shared practices; Phase 3 (3–5 years) shared self-guides; and Phase 4 (9–13 years) shared coordinated roles. I will discuss the nature of the change in each phase and will distinguish among the phases. I will discuss the changes themselves and their significance for children’s self-regulation and social interactions with others, including the benefits and potential costs to children that accrue from these changes.

Precursors to shared-reality development

I believe that Phase 1—shared feelings—represents the beginning of human children sharing reality with others. However, I am not denying that there are important phylogenetic or ontogenetic precursors to shared-reality development in children. Sharing reality with others did not and does not suddenly appear from nowhere. In this section, I consider briefly what might contribute to the development of shared reality in human children.

Humans have some distinctive physical characteristics as a species (for a fuller discussion of these, see Suddendorf, 2013). The white of human eyes makes following others’ eye gaze easier, which facilitates attention coordination as well as feedback for the success or failure of this coordination activity. Infants are able to detect the direction of an adult’s visual gaze when it occurs within the baby’s visual field, and the target of the gaze also has attention-capturing properties (Butterworth & Jarrett, 1991). In addition, human babies are born before full brain development, which means they need to be taken care of longer than other animals and require very close caregiver-child connections for several months. Flavell (1999) mentioned that babies are born with or acquire early several predispositions that help them learn about people, such as finding human faces, voices, and movements especially interesting. In addition, the fact that human adults lived longer than other primates created the “grandmother” effect of there being an adult whose role was primarily to take care of children rather than gather food. All of these characteristics increased the emphasis on close social interaction and the need for adults to take care of the child for the child to survive. It also increased the dependency of the child on the adult and the need for caregiver-child action coordination.

Regarding the social connectedness aspect of shared reality, it is also worth noting that infants more readily imitate the action of another person than the action of an object. Legerstee (1991), for example, found that infants between 5 and 8 weeks old would imitate tongue protrusions and mouth openings that were modeled by an adult but would not imitate these gestures when they were simulated by an object (e.g., for the tongue protrusion, an elongated red object the size of a tongue moving back and forth through the center of a flat disk). Indeed, infants in the object condition were significantly more likely to do the opposite gesture to what the object did (e.g., performing a tongue protrusion after seeing a mouth opening). Positive nonverbal social interactions also occur by 6 months (Hay & Rheingold, 1983). During Stage 2 of attachment (5–11 months), infants show a definite preference for familiar people, such as their mother, father, older siblings, or frequent caregiver, than for a stranger. Around age 8 months, children often cry when the familiar person leaves the room. The impact of social familiarity on interaction preference even extends beyond the caretaking network. There is evidence, for example, that 10-month-olds reach for the identical toy more when it is offered by a native speaker of their language than when it is offered by a foreign-accented speaker (Kinzler, Dupoux, & Spelke, 2007).

All of these human characteristics could have contributed to the emergence of the motivation to share each other’s feelings and interests. Bruner (e.g., 1983) emphasized the importance of imitation and joint attention as being instrumental in children learning to communicate with others. And, certainly, as just mentioned, infants will copy and imitate what an adult does. By 9 months, children can copy even novel actions, such as combining a couple of objects to make a rattle (Suddendorf, 2013), and they are selective in what they imitate, taking into account other (simpler) options to making something happen than what they saw an adult do (Gergely, Bekkering, & Király, 2002). Thus, although not constituting shared reality, per se, imitation can be considered a precursor to shared reality.

Perhaps the closest precursor to the beginning of sharing reality with others is the face-to-face turn-taking interactions that infants have with their caregiver. When only a few months old, they will engage in complementary turn taking, such as behaving more passively when the adult is more active and behaving more actively when the adult is more passive (see Trevarthen, 1979). Infants will take turns with an adult in expressing feelings in a way that goes beyond just copying what the adult does, such as expressing happiness vocally after an adult expressed happiness facially (see D. N. Stern, 1985).

Finally, a cognitive-developmental change that occurs by the end of the first year may also be relevant to the development of sharing reality with others. Case (1985) referred to this change as the ability for bifocal coordination, the ability to coordinate two different reactions with each other. This means that infants can represent the relation between two events, such as the relation between a response that they have produced and someone’s response to it. Even at this early stage, then, children are capable of the preliminary form of role taking described by Mead (1934)—the ability to anticipate the responses of a significant other with whom the child is interacting. It is important to note that the response of the significant other is not just a response like a smile but rather the inner state of feeling interested that underlies the smile. The ability to infer such an inner state would allow infants to anticipate that what they are looking at and find interesting, a significant other could also find interesting if they pointed to it. This ability would support the first phase of shared-reality development—shared feelings.

In sum, there are several properties of human infants and their social interactions with others that could contribute to the early development of sharing reality with others. I turn now to Phase 1 of shared-reality development, shared feelings, a development that some consider to be unique to humans (e.g., Tomasello, 2014).

Phase 1 (6–12 months) shared feelings

Shared feelings as the first phase of shared-reality development occurs around ages 6 to 12 months. During this period children begin to point at something that attracts their attention and to urge others to attend to the same thing (see Tomasello, 2014). Originally, this phenomenon was called joint attention. Having experienced such pointing interactions with my one-year-old daughter years ago, I believe that there was more happening than her getting me to look at what she was seeing. It is important to note that the pointing often occurred on occasions where she clearly was not trying to get me to give her what she was pointing to, such as a colorful bird high up on a tree branch. It was not instrumental in this sense. Bates, Camaioni, and Volterra (1975) early on distinguished between pointing as a protoimperative gesture (pointing as a request to another person to do something for you) versus as a protodeclarative gesture (pointing as a comment to another person about something in the world; to draw another’s attention to something). What I believe is that she was looking at something that she found interesting and wanted me to look at it too and share how interesting it was. It was not joint attention in the service of obtaining a tangible outcome. Rather, it was joint interest—creating a shared feeling about something.

My account of my daughter’s pointing is called a rich interpretation. Other observers of such events also give a rich interpretation of such pointing, describing it as sharing interest with a communicative partner (e.g., Liszkowski, 2005; Suddendorf, 2013; Tomasello et al., 2005) or affective sharing (e.g., Saarni, Campos, Camras, & Witherington, 2006). Suddendorf (2013), for example, emphasized that although pointing can be used by one-year-olds instrumentally as a tool to get what they want, it can also be more than that. Children use pointing to draw others’ attention to things because they are interested in those things and want to share their interest. Bretherton (1991, p. 24) said that during this phase infants’ affective and gestural communications seem intended “to attract and direct the addressee’s attention to topics of mutual interest.”

Tomasello, Carpenter, and Liszkowski (2007) proposed taking a deeply social view on such pointing gestures that, they maintained, involve uniquely human skills and motivations for sharing attitudes and interests. They suggested that these gestures reflect a uniquely human motive “to simply share experience with others” (Tomasello et al., 2007, p. 9). The intrinsic motivation to share interest that underlies such pointing gestures is also suggested by the findings of a study by Liszkowski, Carpenter, Henning, Striano, and Tomasello (2004).

When a 12-month-old infant pointed to an object (toy; light; puppet) that had appeared through an opening in a screen and then moved, an experimenter immediately reacted by looking (a) back and forth between the infant’s face and the event while talking excitedly about what was going on in the event (joint attention), (b) just at the infant’s face (face), (c) just at the event (event), or (d) at neither the infant nor the event (ignore). The infants pointed more often across the trials in the joint attention condition than in the other conditions, as if the joint attention and joint interest were creating a connection between the child and the experimenter and increasing the motivation to create shared reality together. It is interesting to note that within a trial infants pointed more often in the non-joint attention conditions, as if they were trying to get the experimenter to create joint attention with them (create a shared reality).

Moreover, pointing at things is not the only way that children in Phase 1 try to share interest in something. Object sharing also emerges around age 8 months and then increases in frequency (Hay & Rheingold, 1983). One motivation for such object sharing, such as exchanging a toy with a peer, is to interest another person in the object that the child finds interesting (Eisenberg, Fabes, & Spinrad, 2006).

By around the end of their first year, children can also recognize that the adult’s emotional display is about a particular object or activity (Carpenter, Nagell, & Tomasello, 1998; Moses, Baldwin, Rosicky, & Tidball, 2001). They develop the ability to learn what the valence of an object or activity is like by reading an adult’s positive or negative facial or vocal expression when reacting to it (e.g., Mumme, Fernald, & Herrera, 1996). When the expression is negative, for example, they may avoid that object or activity, such as being less likely to cross an apparently deep cliff (with a solid but transparent surface) if their mother’s emotional facial expression is negative versus positive (Sorce, Emde, Campos, & Klinnert, 1985). Indeed, at least for negative emotional reactions, 12-month-olds even learn how to feel about a novel object from an adult’s emotional reaction to that object in a televised scenario (Mumme & Fernald, 2003). In this way, infants learn to share adults’ positive or negative feelings about different objects—social referencing to create shared realities about what to feel about different objects. It is also notable given shared reality’s relation to an epistemic motivation that infants are more responsive to the emotional signals of adults regarding toy objects when the children’s own initial emotional reactions to those objects are ambiguous rather than clearly positive or negative (see Kim & Kwak, 2011). Also perhaps reflecting infants’ epistemic shared-reality motivation is evidence that, in a laboratory setting, they choose to seek information about a novel toy object slowly approaching them on the table by directing their looks to the experimenter in the room rather than to their mother, as if the experimenter was the local “expert” (Stenberg, 2009).

There is evidence, then, that around the end of the first year, infants communicate with pointing gestures to create shared attention and interest in an object or event and they learn what to feel about different objects (at least positively vs. negatively) from observing adults’ emotional expressions toward them. It is for these reasons that I refer to this phase as “shared feelings.” Saarni et al. (2006) described a “feeling” as the registration that an event matters, and they discussed how feelings can be generated through perceiving the vocal, facial, or gestural expressions of others.

There is also evidence that infants as young as age 12 months display concern when another person is in distress and will even try to comfort a distressed person (see Eisenberg & Fabes, 1998; Zahn-Waxler, Radke-Yarrow, Wagner, & Chapman, 1992), where comforting in particular could be interpreted as a form of tuning a socioemotional response toward the feeling state of another. More generally, Phase 1 is a period when children increasingly coordinate their expressive actions to suit their surrounding emotion-eliciting circumstances (Saarni, 2000).

The importance to infants of shared feelings is also evident in their preference for individuals who share their feelings about things over those who do not. In a study by Hamlin, Mahajan, Liberman, and Wynn (2013), for example, infants’ preference for either graham crackers or green beans was first determined. Then rabbit puppets expressed their preference between these two foods, with one puppet preferring the same food as the infant (similar) and the other puppet preferring the other food. Then two dog puppets behaved in a way that either helped (helper) or harmed (harmer) the similar rabbit puppet. Finally, the infant was shown the helper and harmer dog puppets and which puppet the infant reached for was recorded as the measure of preference. The study found that 9-month-old infants, and especially 14-month-old infants, showed a preference for dog puppets that treated the similar rabbit puppet well. Thus, an infant was more likely to choose a dog puppet that treated well that rabbit puppet that had a shared food preference, that is, a shared reality, with the infant.

This Phase 1 development of shared feelings has epistemic and social-relational motivation benefits. Children and their caregivers share what each finds interesting about the world. These child interactions with good caregivers (see the “good enough parent” of Winnicott, 1958; see also Harter, 1999) make an important contribution to children by their learning what is considered significant within their cultural community. Caregivers point out what they, as representatives of their culture, consider worthy of notice, and they decide when to signal to the child that what the child has pointed out is, indeed, worthy of notice and worthy of shared interest. Thus, even at this young age, shared-reality interactions are teaching the child about what is and is not considered significant, what does and does not matter. The potential downside is that children could begin to filter out events they would have found interesting as no longer being relevant because there is no shared interest. Thus, what is worth paying attention to and feeling wonder about could be narrowed.

From a relational perspective, learning what adults find interesting (and not interesting) allows the child to get along better with other members of the community, and the community to get along better with the child. This is a major benefit to the child and society. But, again, it also has potential costs. What one community finds interesting or not interesting is not necessarily shared by a different community. What one community finds interesting another community might find irrelevant, silly, weird, or disgusting. Even as early as Phase 1, then, there is the potential for perceiving in-group/out-group differences that can create intergroup conflict.

These are not the only potential downsides of Phase 1. In their Phase 1 interactions, a caregiver can communicate to a child, “This is what I find to be interesting, and it is what I find interesting that matters and not what you find interesting.” Caregivers could consistently fail to verify the child’s interest in things, regularly communicating that they do not share what the child finds interesting. In addition, it is clear that although the frequency of maternal affirmations of children does increase during Phase 1, even at the end of Phase 1 there are significant differences among mothers in how frequently they affirm their children’s bids for attention (Tamis-LeMonda, Bornstein, & Baumwell, 2001).

Phase 2 (18–24 months) shared practices

For both Piaget (1926, 1951/1962) and Vygotsky (1962), and to a lesser extent Mead (1934), as well as many more recent cognitive-developmental theorists, the most significant change during development is the symbolic development that occurs around 18 to 24 months, especially as it relates to the development of language. For these theorists, it is this development that transforms how children think—the cognitive transformation that makes them human. It should be noted as well that this development is often described as being about the development of language, per se—the development of multiword utterances and the syntax and grammar underlying it. I believe that there is too much emphasis on language, per se, rather on the emergence of words as shared symbols—single word utterances, or names for things, that are symbolic rather than signals and constitute a shared reality with others.

To appreciate the significance for shared reality of the emergence of single word utterances that are symbolic, one must appreciate that the single word utterances that children use before this phase are not symbolic—they are signals. Children’s early words are signals that they use instrumentally, like motor gestures, to make something happen. They are equivalent to a pet dog scratching the door to be let outdoors. These signals constitute instrumental contingency knowledge, equivalent to, “When I point at something, Mommy looks at it.” For words it might be “juice,” as a request for juice, or “up” (often accompanied by arms lifted up), as a request to be picked up. The words are speech acts used as requests or demands—not as comments just to be informative. According to Terrace (2005), this is also the case for the hand gestures of the great apes (chimpanzees, gorillas) that have been taught sign language by human trainers. Around 18 to 24 months, however, there is a qualitative change in how children use words—a change that does not occur with other primates. The use of verbal utterances is not just “protoimperative,” that is, children recruiting help from another person to get what they want by using a verbal signal. As Terrace (2005) said, it is “tacts” and not just “mands.” “Tacts are not followed by primary reinforcement. Their sole function is to share knowledge with another person for its own sake” (Terrace, 2005, p. 99).

The qualitative difference between children’s symbolic versus signal single-word utterances is evident in what happens after the symbolic words emerge—children’s vocabulary explodes. This is the well-known “naming period” phenomenon. During this period, children excitedly point to objects to have someone older tell them their names. They ask for their names (Chouinard, 2007). Before the shift from word signals to word symbols, children learn a couple of words a week. Afterward, they learn 5 words a day (Nelson, 2005)—a phenomenal 150 words a month. The discovery that “everything has a name” is momentous for children. W. Stern (1914) was apparently the first to describe this discovery. He noted that around ages 18 to 24 months the child first realizes that each object has its permanent symbol, a sound pattern that identifies it—each thing has a name. At this phase, then, children utter sounds to refer to something objective. Vygotsky (1962) agreed with Stern that this turning point is observable from children asking for the name of objects and the subsequent sharp increase in children’s vocabulary, and he agreed that this is a decisive turning point in the child’s linguistic, cultural, and intellectual development.

From a mental representation viewpoint, symbols are special developmentally. They are, however, also special developmentally from a shared-reality viewpoint. During the naming period, children point to an object so that someone older will tell them which verbal sound category refers to that type of object. In this “original word game” (Brown, 1958), they also check with others which objects are included and excluded as referents of a verbal sound category. The fact that they do so indicates that they recognize that this older person has knowledge about which sound categories (names) refer to which object categories—knowledge that they do not have but want to have. As Harris (2012) pointed out, the fact that children ask others questions about what is the name of some object implies that they can conceive of an object having a name even though they do not know what the name is, and they believe that the person they question can, and will, tell them what is the name. They recognize that other people have knowledge that they do not yet have and they can acquire this knowledge. By asking for and then retaining this knowledge for names, they create a shared reality with others about how communication with words works. It is a shared reality about the practice of communicating with words in interpersonal interactions, as both encoders of words and decoders of words. In addition, Phase 2 children’s questioning of others to learn the names of things represents an information-gathering strategy that is distinct to humans and is used even before these children can use words properly to ask questions (Harris, 2012).

The impact of single-word labels on toddlers’ shared practices is nicely illustrated in a study by Jaswal and Markman (2007). Two-year-olds were given a key-looking hybrid object that could also be used as a spoon, together with photographs of a cereal bowl and a car. They were asked to demonstrate the function of the hybrid object. When the key-looking hybrid was not labeled, the children demonstrated its function to start the car. But when it was labeled “spoon,” the children demonstrated its function to eat from the cereal bowl. Thus these 2-year-olds accepted the shared practice suggested by the adult’s label. Jaswal (2007) also found that the 2-year-olds who had the largest vocabulary were also the children who most often accepted the unexpected label for a hybrid object. This finding suggests that children who trust that they can learn the right shared practice from adults’ communications acquire shared-reality labels more quickly.

It should also be noted that older toddlers at least (24 months old) are also sensitive to feedback from others—social cues of assent or dissent—about what is the real name for some object. In a study by Fusaro and Harris (2013), for example, two experimenters sat next to each other, opposite from the participant toddler. A pair of novel toys was held up one at a time, and one experimenter gave first one toy a novel name (e.g., “modi”) and then the other toy the same novel name (“modi”). The second experimenter, while looking at the child, gave a head gesture for one toy that indicated agreement with the first experimenter’s claim of name but gave a head gesture for the other toy that indicated disagreement that the name was correct for that toy. The first experimenter then asked the child to select one of the two objects as the answer to the question, “Which one is the modi?” The 2-year-olds chose the toy receiving the second experimenter’s assent more than the toy receiving the second experimenter’s dissent. Thus, the second experimenter broke the ambiguous tie to create a shared reality with the toddler.

The development of symbolic single word utterances (and other symbolic vehicles) is not the only new shared-reality learning that happens during the 18- to 24-month Phase 2. It is also the time for children to learn other social practices and social routines from their guided participation in community activities (Rogoff, 2003; R. A. Thompson, 2006; Whiting & Whiting, 1975)—what (Rogoff, 2003, p. 3) described as the “routine ways of doing things in any community’s approach to living.” This phase includes eating routines, dressing routines, bathing routines, as well as the toilet routines famously described in S. Freud’s (1938) “anal period.” According to Erikson (1950/1963), success in learning such routines contributes to children experiencing “autonomy,” whereas failure contributes to children experiencing “shame.” It should be noted that children experiencing self-evaluative shame from failing in relation to internal standards occurs in Phase 3 rather than Phase 2. However, toddlers in Phase 2 can be teased and shamed by others as indirect forms of social control when they fail to act in the appropriate way (see Rogoff, 2003).

When learning these social routines, toddlers want to be independent, do things right, and be in control—the classic “terrible twos” where the exclamations “Me do it!” and “No!” rule. In fact, this childhood phase is not always so terrible, but independence and control do matter. Yes, toddlers are being asked to control themselves; they are being asked to “stand on their own two feet.” But caregivers structure the learning settings and activities to support and scaffold what the children need to learn and deal with (see Rogoff, 1990; Saarni, 2000).

What is important to recognize here is that what each child is being asked to do depends on what his or her family (or community) says should be done. Because this varies considerably across cultures (Rogoff, 2003; Rogoff, Paradise, Mejia Arauz, Correa-Chavez, & Angelillo, 2003), what exactly needs to be done is basically arbitrary. It is arbitrary like the communication practice of which particular word sound category is used to refer to the four-legged animal that barks and tries to lick you (the word sound category “dog” if the child happens to be in an English-speaking community). Children learn their family’s eating practices, dressing practices, and so on, like they learn the practice of symbolic word usage when communicating, “This is how we do it.”

In Phase 2, then, children create a shared reality with significant others, especially with adult authorities, regarding how to carry out particular activities—shared procedural knowledge, shared practices. These include shared practices about helping another person who is having difficulty completing a goal, such as helping an experimenter retrieve an object that was accidentally dropped on the floor and is now out of reach (see Warneken & Tomasello, 2006). It also includes participating in new shared practices necessary for coordinated play, such as working together with an experimenter on a play activity where success requires that each person performs a different function (e.g., one person pushes up a cylinder with an object and holds it in place until the other person retrieves the object; Warneken, Chen, & Tomasello, 2006; Warneken & Tomasello, 2007). It is also during Phase 2 that children engage with others in a more coordinated turn-taking manner, such as waiting their turn to vocalize and timing what they do to be responsive to what’s happening with their partner (e.g., Dix, Cheng, & Day, 2009; Rutter & Durkin, 1987; see also Eckerman, Davis, & Didow, 1989). From a shared-reality perspective, then, the cognitive-developmental, social-cognitive developmental, and socioemotional development changes that occur in Phase 2 are describing the same basic developmental change—the emergence of shared practices.

One additional activity that emerges during this period should also be mentioned—pretend play (see, e.g., Harris, 2000; Harris & Kavanaugh, 1993; Piaget 1951/1962). Around age 2 years, children begin to use props to play out an imaginary activity, such as imagining that a teddy bear is wet and using a piece of paper as a make-believe towel (Harris, 2000). The props function like symbolic words—arbitrary symbolic vehicles that refer to a mental category. Children use these symbolic props to carry out shared practices, such as drying an object that is wet. The use of such props allows children to practice carrying out shared routines without needing the normal conditions associated with that routine, which has benefits for children (see Harris, 2000).

Once again, this developmental change in shared reality has both benefits and potential costs for children with respect to both epistemic and social relational motivation. When children use symbolic word utterances, for example, the sound pattern category (the lexical category), the object category, and the symbolic relation between them are all shared realities. It is the epistemic significance of this change that has received the major emphasis in the literature. This includes the classic “Whorfian hypothesis” that thought is shaped by language (Whorf, 1956). The social-relational significance has received much less attention, although as Vygotsky (1978, p. 28) pointed out, “Signs and words serve children first and foremost as a means of social contact with other people.” When we use symbolic utterances we are relating to others, experiencing shared realities with others, and experiencing connection to others. Thus, Phase 2 reflects an important change in how children are connected to others. Ironically, this change from the emergence of symbolic single-word utterances is in some ways more profound than Whorf’s hypothesis because the shared reality of symbolic words impacts with whom you feel connected and share the world—whose knowledge about the world is relevant—rather than only affecting thought.

The epistemic and social-relational effects of symbolic single-word utterances are highlighted by the specifically human behavior of incorporating words when using tools (Vygotsky, 1978). Children in Phase 1 will use a spoon to eat and also to bang to make noise or get attention. A spoon is just a particular object in the world that can be used in multiple ways. But when a spoon is called a “spoon” in Phase 2, there begins the encoding and decoding of “spoon” as shared communication practices. Such use of names has the benefit of adults communicating to children “how we do it” and children learning which practice is “the right way to do it.” On the other hand, this could inhibit creative use of an object such as a spoon because the name for the object, “spoon,” activates a particular practice or use for that object, that is, eating, rather than alternative uses such as taking the cap off a jar or making noise. It is interesting in this regard that young bilingual children who learn early on that there is more than one right way to name the same object are also more creative (Leikin, 2012).

From the perspective of epistemic knowledge of oneself, there is also the potential downside of language use—as a “double-edged sword” (Harter, 1999, p. 34)—where children use language in a way that distorts their actual experiences (see, e.g., Bretherton, 1987, 1991; Harter, 1999; D. N. Stern, 1985). In addition, as Harter (1999) pointed out, children’s new language skills can be used to withhold experiences with others as well as to share them.

From a relational point of view, calling a spoon a “spoon” in Phase 2 now connects to others, which is beneficial to children and to society. However, by connecting more closely to those who share reality about the symbolic utterances, distance can also be created from those who do not share this reality (see Spelke, 2013). In an interesting study that suggests how important it is to younger children that others use language in the same way that they do, accent versus accuracy were pitted against each other (Corriveau, Kinzler, & Harris, 2013). Children 3, 4, and 5 years old who were native English speakers observed two adults speaking English suggest novel names for different novel objects, with one speaker having a native accent and the other having a foreign accent. All of the children preferred the novel name suggested by the native-accented speaker. Then the two speakers gave names for familiar objects: In one condition the native speaker always gave the incorrect name and the foreign-accented speaker always gave the correct name. Following this “accuracy” manipulation, the 4- and 5-year-old children now chose the novel name for a new novel object suggested by the accurate foreign-accented speaker rather than the inaccurate native speaker. The younger children did not do so, suggesting that for them epistemic accuracy did not trump relational connectedness.

More generally, sharing practices—shared knowledge of how things should be done—has the relational benefit to the in-group community of accepting one another when they perform the shared practice, but it also has the potential cost of rejecting out-group members who do things differently. Moreover, this potential for creating in-group/out-group differences in Phase 2 can include what will develop later into moral ideological differences. Shared eating practices, for example, can relate to becoming for or against vegetarianism or becoming for or against Hindu eating practices or Kosher Jewish eating practices. Shared practices are shared beliefs to which children become committed and moral ideologies are shared beliefs to which people are committed. The emergence of shared practices in Phase 2, therefore, has the potential downside of initiating the development of the type of in-group/out-group differences that develop into moral ideological differences.

There is another potential downside from Phase 2 children wanting to share with others a practice for how to do an activity. This motivation can be strong enough that children will give up a better benefit/cost ratio for the sake of sharing a practice with another person—the case of overimitation. In a study by Nielsen (2006), for example, 12-, 18-, and 24-month-old children saw a model open three different boxes, each containing a desirable toy, by using a particular object to open each box. The children could then try to attain the toy by copying how the model opened each box with the object, or they could ignore how the model opened each box and simply open it without using the object, which was more straightforward. The 12-month-olds cared more about the outcome than the modeled means or practice—they simply used their hands to try to open the box rather than the modeled object (see also Gergely et al., 2002). In contrast, the 18-month-olds and especially the 24-month-olds used the modeled object to try to open the box even though it was more straightforward to just use their hands. Nielsen (2006) suggested that the older toddlers copied the means used by the model to sustain or initiate interaction with the model. It is important to note that the 24-month-olds persisted in copying the model’s practice even when doing so failed to open the box.

This study demonstrates a potential downside of this phase of development where a shared reality for a practice can trump a more efficient means for attaining a desired end-state. Of course, as Nielsen (2006) suggested, the shared reality could be worth the attendant inefficiency for social connectedness benefits. In addition, beyond relational motives, there could be subsequent epistemic consequences from children trying to understand the meaning of what the adult is doing. It has been suggested that by trying to treat all of what the adult is doing as being meaningful, the children can revise their causal understanding of the object or activity, which subsequently can have both benefits and costs (see Lyons, Young, & Keil, 2007). Consistent with a shared-reality perspective, it has also been suggested that such overimitation in children derives from humans being such a thoroughly cultural species that copying others is generally functional for children as a kind of default strategy (see Whiten, McGuigan, Marshall-Pescini, & Hopper, 2009).

Phase 3 (3–5 years) shared self-guides

From a historical perspective, it is interesting that the classic cognitive-developmental and socioemotional developmental literatures differed in the weight that they gave to what happens in the 18- to 24-month period versus what happens in the 3- to 5-year period. In the cognitive-developmental literature, the emergence of symbolic processes in the 18- to 24-month period has historically been given the greater weight. In contrast, what happens in the 3- to 5-year period has historically been given greater weight in the classic socioemotional literature—the emergence of children adopting significant others’ goals and standards for their self-regulation and self-evaluation. This latter emphasis began with S. Freud’s (1923/1961) discussion of the development of the superego (or conscience) in the 3- to 5-year period that changes children’s self-regulation forever by their identifying with significant others’ standards and objectives for them. What is the superego that underlies the development of a conscience? It is, in its most general form, the shared reality between a child and a significant other (or others) regarding who the child should be or become—a shared reality about the child’s objectives of goal pursuit and standards of self-evaluation.

The change in Phase 3 from children anticipating others’ external responses to their behavior to children regulating themselves in terms of others’ goals and standards for them (without surveillance) has been referred to as “identification” (e.g., Harter, 1999, 2006; Moretti & Higgins, 1999) and as “internalization” (e.g., Kochanska, Murray, Jacques, Koenig, & Vandegeest, 1996; Stipek, Recchia, & McClintic, 1992) in the literature. It does involve internalization in the sense that self-regulation develops from children taking into account others’ external responses to their behavior to children regulating in relation to internal representations of others’ goals and standards for them. But this should not be confused with internalization in the sense of children self-regulating in relation to their own personal standpoint that is experienced as distinct from the standpoint of others. Self-evaluation that distinguishes the child’s own personal perspective from the perspective of another person does not occur until after Phase 3 (Harter, 1999, 2006). Thus, to avoid confusion, it is preferable to refer to what is happening in Phase 3 as identification rather than as internalization, although consistent with the literature I will also distinguish children’s use of internal versus external standards of self-regulation and self-evaluation. (See also the discussion of Deci & Ryan, 1985, 2000 about internalization being a developmental continuum and the proposal by Lewis, 2007, for using the term “incorporation” for the identification process).

Self-regulating one’s goal pursuit while taking others’ desires into account is clearly related to S. Freud’s concept of the superego because the superego represents children’s understanding of what significant others want them to do. But children’s understanding of what their parents (or other significant others) want or expect them to do is not restricted to what their parents believe they ought to do. It also includes an understanding of their parents’ hopes for them (Higgins, 1991). It should be noted that children acquire these self-guides not only from feedback and direct instruction from their caregivers but also from observing how a significant other responds to others, including a positive or negative response or comment about the behavior of an older sibling or a playmate of the child (i.e., observational learning; see Bandura, 1986; Lewis, 2003).

What is central to Phase 3 shared self-guides is that there are standards of good behavior or performance that significant others want the child to follow or achieve and the child accepts and self-regulates in relation to these standards without the need of there being external control by others. Specifically, children follow the standard even when there is no one observing what they do who will reward or punish them for doing it, and there is no one in the immediate situation who is asking or telling them to do it. The use of self-guides in self-regulation is also revealed when children have self-evaluative feelings about failing to meet the standard (e.g., shame) in the absence of observation or feedback from an adult.

An illustration of this paradigmatic case is provided in a longitudinal study by Kochanska and Knaack (2003). A composite measure of “internalization” or “conscience” was created by combining the results of two tasks given to children who were 56 months old. In one task, the child’s mother told the child that playing with attractive toys on a shelf was off limits. Then the mother left the child alone in the room doing a dull task. Whether the child followed the prohibition standard was observed. In the other task, the child played a throwing game with velcro-covered balls and was told that he or she would receive a prize for each ball that stuck to the target. The experimenter’s rules for the game (e.g., face away from the target; stay far away from the target) made it practically impossible to win. The experimenter left the child alone, allowing the child to cheat unobserved. What the study found was that children’s effortful control (performance on tasks requiring the child to suppress a dominant response and perform instead a given subdominant response) was highly longitudinally stable and coherent across tasks at 44 months old and was a source (predictor) of children’s “internalization” as measured at age 56 months.

The Kochanska and Knaack (2003) study illustrates self-regulation in relation to internal standards by examining whether children follow rules they have been given even when they are alone. A study by Wu and Su (2014) illustrates whether children follow the rule of sharing with others without direct social pressure from others. They found that 4-year-old children were more likely than younger children to share their toys with an experimenter-manipulated puppet without the puppet needing to express their desire for the toys or make an explicit request for the toys (see also Svetlova, Nichols, & Brownell, 2010).

In Phase 3, then, children self-regulate in terms of standards or guides they have learned from others without needing external social pressure or control (Harter, 2006, 2015; Higgins, 1991). These guides influence what they choose to do or not to do. In addition, by identifying with these guides, they experience positive and negative feelings, respectively, when they succeed or fail to satisfy or meet them. Stipek et al. (1992), for example, described how there is a shift between ages 2 and 4 years in children’s self-evaluative responses, such as a dramatic increase in pouts or frowns after failure and smiling after winning, which they interpret as a gradual internalizing of external reactions and self-evaluations that no longer depends simply on how they expect adults to react to their performance.

These positive and negative feelings from self-evaluation include feelings such as pride and shame or guilt (Kochanska, Casey, & Fukumoto, 1995; Lewis, 1995; Lewis, Alessandri, & Sullivan, 1992; Mascolo & Fischer, 1995; Stipek, 1995). Indeed, it has been argued that children’s experiences of these “moral” emotions stem from children evaluating themselves in relation to standards of what kind of person they want to be or should be (e.g., Lagattuta & Thompson, 2007), and around age 3 years these standards are internally represented social standards (e.g., Hart & Matsuba, 2007; Stipek, 1995). There are earlier self-conscious emotions, but around age 3 years children begin to evaluate their behavior in relation to internal social standards, which produces pride, shame, and guilt (Lewis, 2003, 2007).

Given its implications for self-regulation and self-evaluation, this general socioemotional part of the story of Phase 3 is very important. However, there is also an important social-cognitive developmental change in Phase 3 that contributes to these socioemotional changes and needs to be highlighted. Again using Case’s (1985) neo-Piagetian model of cognitive development, children from ages 3 to 5 years can for the first time coordinate two separate systems of interrelations, such as coordinating observable responses with unobservable responses. This cognitive capacity underlies their ability to make inferences about the expectations and preferences that their parents have in relation to their own behaviors (see Shantz, 1983). In particular, children can now understand that another person wants them and expects them to behave in particular ways and that how this person responds to them depends on what this person wants and expects of them (Fischer & Watson, 1981)—this person’s standard or guide for them. In Phase 3, children accept this standard or guide, identify with it, and use it for their own self-evaluation—transforming it to a self-guide.

Notably, this change in inferring and responding to the inner states of others is also evident in changes in how children play with their peers. Cooperative play with peers emerges between ages 20 and 24 months and increases between ages 24 and 28 months (see, e.g., Eckerman et al., 1989; Eckerman & Whitehead, 1999), but only around age 3 years does children’s cooperative play involve sharing desires, intentions, and goals (see, e.g., Brownell, Nichols, & Svetlova, 2005; Howes, 1988). The interaction between children’s new cognitive capacities in Phase 3 and the socialization input from their significant others form a basis for children to identify with their significant others and use new self-guides for their self-evaluation (Harter, 1999, 2006; Higgins, 1991).

There is another important social-cognitive change that occurs in Phase 3 that also influences children’s self-regulation and self-evaluation. As Nelson (2005) noted, 2-year-old children have a limited sense of their self in the past or the future. Even their play concerns “what is or can be, not with what was and what may yet be” (Nelson, 2005, p. 130). They are not Tulving-like “mental time travelers” (see Tulving, 2005). However, by age 4 or 5 years, children become aware of themselves as individuals with a past and a future (Gopnik & Graf, 1988; Nelson, 1992). The development of a narrative self contributes to the emergence of a cultural self (Nelson, 2003, 2005). During Phase 3, children also begin to consider the possibility that things as they are might instead be some other way (Nelson, 2005).

All of this matters because it means that children can now self-regulate in relation to shared goals and standards not in terms of who they are in the present but who they could become in the future (Higgins, 2005). Thus, children cannot just rest on their laurels. In Phase 2 it is enough to carry out the shared practices, but in Phase 3 the child needs to continue to improve, especially for self-guides that are aspirations for the future. As J. M. Baldwin (1897, p. 36) put it, “It is not I, but I am to become it. Here is my ideal self, my final pattern, my ‘ought’ set before me.” In his book, Becoming, Gordon Allport (1955) suggested that because children in this period are self-regulating in relation to what others want them to become, they develop a stable sense of self-identity and long-range goals that provide directions for striving.

There is yet another well-known change that occurs during Phase 3 that is typically discussed under the phrase, “Theory of Mind” (TOM). Some of what I have already discussed is related to this literature. However, the emphasis in this literature is not on the development of shared realities, per se. Indeed, perhaps the most famous example of the development of TOM from 3 to 5 years old would appear to involve the ability to recognize a nonshared reality—the false belief test. It is often described as the ability to understand that others have beliefs, desires, and intentions that are different from one’s own. In the classic Wimmer and Perner (1983) study, a character Maxi had put his chocolate somewhere and then left, and while he was away his mother moved it to a different location. The child participants in the study are asked where Maxi will look for the chocolate when he returns. Generally speaking, children under age 3 years state that Maxi will look where his mother put the chocolate. Children around age 4 years usually state that Maxi will look where he originally put his chocolate, even though they know that this is not where the chocolate actually is now. This phenomenon is described as the older children passing the false belief test. It is called this because Maxi’s belief about the location of the chocolate is false and different from the children’s belief about where the chocolate actually is now.

This TOM shift as a function of age is an interesting phenomenon (see Flavell, 2004; Wellman, Cross, & Watson, 2001). As I said, it appears to illustrate a nonshared reality that develops in Phase 3, although aspects of its development may occur in Phase 2 (for thoughtful reviews, see Baillargeon, Scott, & He, 2010; Wellman, 2014). But let’s consider now a somewhat different interpretation of what is happening in studies like the Maxi study—an interpretation that involves a shared reality rather than a nonshared reality. It is true that the children participants and the character Maxi have different beliefs about the location of the chocolate. And children cannot just assume that what they know and what Maxi knows about the world is the same. It is different from the Phase 2 case of shared practices because it is fine for 2-year-olds to assume that others share the practice of using particular words for particular object categories and that they share basic eating practices, and so on. In their family and local community, others probably do share those practices.

You will fail the false belief test, however, if you simply assume that “what I know you know.” There needs to be something more. I propose that the something more could be a particular kind of shared reality. Specifically, it involves sharing reality with Maxi about where you would look for the chocolate if you were in his situation—what I years ago called “situational role-taking” (Higgins, 1981b). “If I were in Maxi’s shoes, I would look for the chocolate where Maxi put it because he doesn’t know it was moved.”

So, according to this proposal, the TOM change in Phase 3 does involve a shared-reality change after all—it involves the ability to share with another person the same course of action for getting the chocolate in the same knowledge situation. Guide in its most general sense means something that directs or influences the course of action to a particular end (Webster’s, 1989). Here, the children are assuming that in the same knowledge situation, what would guide their behavior and what would guide Maxi’s behavior is the same. This constitutes another kind of self-guide regulation because the children are guided by what they would do if they were in the target’s situation—”What would I do with the same knowledge as Maxi?” It is not the truth about where the chocolate really is, but it is the truth about what you and Maxi would do in the same knowledge situation—a shared self-guide. It is important to note that to pass this test children do not have to inhibit what they would do in Maxi’s situation because they believe that Maxi and they are different kinds of individuals who would react differently even in the same situation. That would be a much more challenging task. It is the individual role taking of seeing the world through the eyes of others in the same situation, where their eyes are different from your eyes, rather than just putting yourself in their shoes and assuming that you would both react in the same way (see Higgins, 1981b). Individual role taking is the kind of non-egocentrism that the literature has shown does not develop until later (see, e.g., Selman & Byrne, 1974). But this TOM task only requires situational role taking, and situational role taking involves shared self-guides.

I should also note that there is an intriguing positive relation between children’s engagement in make-believe play, such as role play or having an imaginary friend, and the TOM development in Phase 3 (see Harris, 2000; Schwebel, Rosen, & Singer, 1999; for the positive relation of make-believe play to creativity, see Mottweiler & Taylor, 2014). Again, make-believe play, such as pretending to be a pirate who has found a treasure that other pirates might steal away (Harris, 2000), does not require the kind of nonegocentrism that is involved in individual role taking. But it does require situational role taking, such as imagining what you would do if you were in the pirate’s situation. Like TOM, this involves shared self-guides.

Situational role taking can also be involved in children sharing their toys with others because children can put themselves in another child’s situation of not having the toy they are playing with and know that in that child’s situation they would want to be given a chance to play with the toy. If so, then one would expect that TOM would relate positively to children sharing their toy with another without needing a direct expression of desire for the toy or a direct request for it. Indeed, this is what was found in the Wu and Su (2014) study described earlier.

It should be emphasized that although both Phase 2 and Phase 3 involve shared guidance, the nature of that guidance is different. The guidance in Phase 2 involves children learning from others how food utensils, word sounds, and so on, should be used—shared practices. The guidance in Phase 3 involves children learning from others what goals they should pursue and what standards they should use to evaluate their success or failure in goal pursuit—shared self-guides.

Children in Phase 3 understand that other people want them to behave in particular ways and that others’ responses to them depend on whether or not their behavior matches the others’ goals and standards for them (i.e., self-guide mediation). By using these viewpoints for their own self-regulation and self-evaluation (identification), children acquire shared self-guides. This acquisition has the benefit that children can plan their behaviors to match these shared self-guides, which gives them an enhanced subjective sense of self-control. In addition, to the extent that significant others’ responses to a child are based on their goals and standards for that child, children’s new ability to infer those goals and standards increases the likelihood that they will be effective in self-regulation—a major benefit from this shared-reality phase.

In terms of mechanisms of social influence, there is a major difference between social regulation based on compliance and social regulation based on identification, from behaving in certain ways because they are instrumental for attaining rewards and avoiding punishments from others versus behaving in certain ways because they match values and standards for how one should behave (see Hoffman & Saltzstein, 1967; Kelman, 1958). As Anna Freud (1937) suggested, when children reach this phase, their self-control no longer depends on the anticipation of suffering that may be inflicted by outside agents. A permanent institution has now been set up that embodies others’ wishes and requirements. This shift in social influence benefits society because children will better control how they behave in ways that match what society wants without the need for direct surveillance (see Kelman, 1958). It also benefits children’s sense of self-determination (Deci & Ryan, 1985, 2000) because they experience their choices of how to behave as being intrinsically motivated (what they want to do) rather than extrinsically motivated (what they have to do).

The self-guides that children acquire in Phase 3 change children’s motivational and emotional lives because they begin to use those self-guides as goals to pursue and as standards to evaluate their success or failure to attain those goals. If their monitoring system perceives a discrepancy between their actual self and a self-guide, they can take action to reduce it. They can also plan action and take potential discrepancies into account, thereby using prospective self-evaluation in the service of present self-regulation (see also Bandura, 1986; Carver & Scheier, 2008). They can situationally role take, which can help them understand how others would feel if they were to behave toward them in a prosocial or antisocial manner and thereby better internally control their actions toward others. Such improvement in self-regulation could contribute to the decline in aggression that occurs during this period (see Dodge, Coie, & Lynam, 2006; Eisenberg & Fabes, 1999).

These are all self-regulatory benefits for children from the Phase 3 development in shared reality. There are potential costs as well, however, including epistemic costs. When planning, you choose among the options that you know. The problem is that children’s shared goals and standards may be the only accessible options to them. Other possible options may not be considered, even when another option could be a better fit for a child, such as building things versus dressing up for a particular girl. In addition, this potential downside continues so that teenagers plan to have careers that their parents want them to have rather than what fits their interests or skills. Shared self-guides can be so strong that children don’t even know what else they might choose.

With respect to emotions related to failure to meet moral standards, there is also evidence, mentioned earlier, that Phase 3 children who violate moral standards manifest guilt reactions about their standard violation (e.g., Kochanska et al., 1995). Indeed, depression involving excessive guilt has been observed in preschoolers (see Luby, 2010). It also worth noting that children with more secure attachments are also more likely to share moral standards (Kochanska, Aksan, Knaack, & Rhines, 2004). This finding is consistent with the importance of connectedness or social-relational motives in wanting to create a shared reality with others. It also illustrates another interesting trade-off in Phase 3 because children, especially those with secure attachments, will have stronger shared moral standards, which means that they can benefit by behaving more morally but also suffer more from feeling guilty when they fail to meet the moral standards (cf. Newman, Higgins, & Vookles, 1992).

Phase 4 (9–13 years) shared coordinated roles

Admittedly, deciding the age parameters for this last phase of shared-reality development was more difficult than the previous phases. From a neuroscience development perspective, there is a transition around 10 years old that marks an important change point in amygdala–cortical functional connectivity (Gabard-Durnam et al., 2014). From a sociocultural development perspective, important changes occur as soon as children begin formal schooling, which would be around 6 years old, including that formal schooling is a new social life phase that exposes children to a greater variety of adult authorities and peers than ever before (see, e.g., Erikson, 1950/1963; Higgins & Parsons, 1983; Pianta et al., 1999; Saarni, 2000). As Erikson (1950/1963, p. 259) stated, “School seems to be a culture all by itself.” In cultures where children early on participate in adult activities with family members and learn productive roles (e.g., farm families) rather than preparing to become adults through schooling (see Rogoff, 2003; Rogoff et al., 2003), this last phase of shared roles could occur even earlier. From a traditional psychodynamic perspective, the Phase 4 age range includes the late juvenile period along with the early adolescence period, which some would prefer to separate (e.g., Sullivan, 1953). Thus, finding a precise demarcation period is difficult. Indeed, illustrating historically the difficulty of demarcating this phase, Freud’s proposed “latency period” lasts from about age 5 years to puberty, and Mead was not clear about when the “generalized other” stage actually appeared.

I am left then with a weak but I believe accurate claim: Regardless of precisely when Phase 4 begins, the shared-reality change that emerges during this phase (and perhaps a little earlier or a little later) is qualitatively distinct from the shared-reality change that emerged in Phase 3. Moreover, the shared-reality change that emerges in Phase 4 really matters for becoming a human adult. This phase spells the end of childhood.

What children have in place by the end of Phase 4 is the motivation and ability to cooperate in group activities that require the coordination of multiple roles, and the group member in each role must understand what is expected of the members in the other roles. According to Bratman (1992), shared cooperative activity, or teamwork, has the essential characteristics of mutual responsiveness, the shared goal of mutual knowledge that we are doing the activity X together, and coordinated plans of action and intentions that require understanding the complementary roles. Such shared cooperative activity has also been highlighted as critical to the development of humans as a distinct species (e.g., Suddendorf, 2013; Tomasello, 2014), including our unique ethical concerns (Kitcher, 2011). According to Harari (2015), it is our ability to work together in groups of strangers (i.e., nonrelatives) that make us special as a species, and we can do so by organizing ourselves into coordinated and complementary roles. It is in Phase 4 that shared coordinated roles emerge.

Children in Phase 3 have shared goals and standards with their significant others. However, this is quite different from the sharing of teamwork and complementary roles that occurs in Phase 4. Most especially different is its impersonal nature. The Phase 3 standards and objectives apply to a child’s own personal self-regulation and self-evaluations. The shared self-guides are what the children’s own significant others want for them as particular individuals. The Phase 4 rules and norms demanded by a specific activity apply to anyone engaging in that activity and to anyone who enacts a specific activity-related role. With respect to children following rules, Piaget (1932/1965) said children’s codification of rules occurs around ages 11 to 12 years. He said that such rules “constitute a well-marked social reality” that is “independent of individuals” (p. 24). It is not about you; it is about the role. Someone else can substitute for you and take over the role you are performing. Moreover, most group activities like a sports game involve multiple roles, and each player must understand how each of the other roles work to coordinate with them effectively. Important characteristics of groups are that the members work interdependently for some shared purpose and the members develop specialized, differentiated roles within the group as they work together to pursue their collective purpose (see Hackman & Katz, 2010). This kind of group organization with coordinated roles emerges in Phase 4.

Mead (1934) precisely made this distinction when distinguishing play versus game. He says that in play a child takes the role of a particular significant other and in relation to self. “Daddy, I will be you, and you will be me.” In contrast, a game is not a particular other. It is a generalized other, and the expectations for each role are the same for all those who enact the role. Moreover, the role cannot be enacted alone but must take into account the other roles in the game. As Mead (1934, p. 154) said, “The attitudes of the other players which the participant assumes organize into a sort of unit, and it is that organization which controls the response of the individual.” And later he said (p. 161), “That which makes society possible is such common responses, such organized attitudes. . . .” In a game there is an organization of interrelated roles, and it is to the organization as a whole that the child belongs. Notably, Piaget (1951/1962) also said that although children’s play around age 4 years can involve differentiated roles, between 7 and 11 years old an ever-increasing coordination of roles and games with rules appear.

This is quite a different social world for children in this phase. In Phase 3, they pursued goals in the service of meeting shared self-guides concerning them as an individual person. They experienced the standards and objectives that they had incorporated as being what significant others wanted them, individually, to become. In Phase 4, strangers are now telling you what is expected of people occupying your role and how you must coordinate your role-relevant behavior with that of people playing other roles to make a social entity (e.g., family, work team, organization) thrive. The objectives and standards are no longer associated with significant others but, instead, as Mead (1934) said, are associated with “the generalized other.” It is not you that matters, it is the larger entity, which now controls your behavior. It is the organization that controls your response. According to Loevinger (1976, p. 73), Sullivan believes that it is in early adolescence when true collaboration originates, in the sense of collaboration implying mutuality. This represents a dramatic shift in socially regulated self-regulation. We share teamwork and coordinated roles with equals.

The paradigmatic case of coordinated roles when interacting with others is communication between two persons taking turns being in the speaker role and the listener role while participating in the “communication game,” where “game” refers to purposeful social interaction involving interdependent roles and conventional rules (see Higgins, 1981a). As I discussed earlier, children become better during Phase 4 in tailoring their message to suit the inner states (e.g., attitudes, knowledge) of others. Even 9-year-olds too often assume that what they know their audience also knows. It is during Phase 4 that there is a sharp improvement in children’s conversational skills and they begin to function in a manner more comparable to adults. Dorval and Eckerman (1984) found that compared with children in Grade 2 whose conversations were brief or aborted 55% of the time, the conversations of Grade 5 children were brief or aborted only 15% of the time. In addition, on a measure of conversational turn taking that distinguished between turns that were related to what had just been said versus unrelated, 40% were unrelated for Grade 2 children compared with only 15% for Grade 5 children. Notably, Dorval and Eckerman (1984) found that the conversations of the Grade 5 children on these measures were quite comparable to those of adults.

In their classic study, Glucksberg and Krauss (1967) illustrated the development during Phase 4 of effective coordination between speakers and listeners. The speakers had to tell a listener how to stack a set of colored blocks on a peg. After the speaker’s message, the listeners gave feedback that they did not understand to which block the speaker meant. Whereas kindergarteners tended to respond with silence or simply repeat what they had said before, third and fifth graders either provided a new description or modified their earlier description, which was like what adult speakers did. Notably, Phase 4 is more about learning how to cooperate in coordinated game roles than it is about language development, per se. In addition, it is important to note that there is a difference between children in Phase 4 enacting the communication roles of speaker and listener (see Higgins, 1981a, for a description of the normative expectations or rules associated with these roles) compared with children in Phase 2 being encoders and decoders of symbolic utterances.

The broadening of the source of prescriptive norms from “significant others” to “the generalized other” where the normative expectations or rules for a child in a role such as the roles of speaker and listener are the same for anyone enacting the role—the same basic standards for all—is also reflected in a broadening of descriptive norms for self-evaluation during Phase 4. Although children in Phase 3 compare their current possessions with other individuals who belong to their group, such as comparing how much candy they have (self-other in-group comparisons), children in Phase 4 begin to make more complex comparisons involving, for example, self-other comparisons with individual members of out-groups and group-level comparisons involving the in-group and out-groups (see Levine & Moreland, 1986, for analyses of outcome comparisons in group contexts). Moreover, children in Phase 4 begin to rank their own performance and that of others along dimensions of competence using standards that apply to everyone (e.g., Ruble, 1983)—the same standard for everyone.

In Phase 4, children now perceive authority as a shared, consensual hierarchical relation between parties that is adopted (for varying lengths of time) because it confers benefits for the group. Rather than obeying an authority because that person is an older significant other, obedience is now experienced as being in the service of cooperation and coordination, and it is a shared consensual relation among more or less equal members of the group (Damon, 1977; see also Kohlberg, 1969). Because the coordination is for the common good, the peer authorities responsible for their roles in the group (e.g., captain of a team) also believe that they should listen to their subordinates who are responsible for other roles in the group (i.e., other team members). It is important to note that unlike adult authority, for peer groups, authority rests on the consent of the governed (Damon, 1977). Children can exercise mutual control from regarding each other as equals. Everyone in society shares an equal responsibility to obey the rules, and respect for them develops from recognizing their social regulation function, particularly their organization function such as division of labor.

These changes in how children relate to one another are also reflected in changes in friendships. From Phase 3 to Phase 4 there is a shift from friendship being about doing something with another child, like doing some play activity together, to friendship being about doing something for another child, such as trying to help them out with their problems or having them help you out with your problems (Damon, 1977; Selman, 1980; Youniss, 1980). It is like games where there is cooperation and coordinated roles to serve and assist in one another’s interests. There is a change in what is expected from a close friend. The role of a close friend is different. It includes sharing secrets and concerns with one another, allowing each person to help out the other more effectively. Because children now share their problems and their secrets, close peer friendships can become more intimate (Watson & Valtin, 1997), and their partner tries to help them with their problems (Berndt, 1983; Sullivan, 1953). Because the collaboration and coordination is between equals, it can have a unique kind of intimacy, a new kind of shared teamwork where each partner wants to help the other partner achieve his or her interests. In comparison, the prior parent-child relationships are asymmetric in this regard. Given that so much of children’s activities during Phase 4 are impersonal, having a special friend with this kind of intimacy is a blessing.

As with the other phase changes, Phase 4 also involves cognitive and social-cognitive developmental change. By Phase 4, children are capable of coordinating values along two distinct dimensions (see Case, 1985; Fischer, 1980; Piaget, 1970), as reflected in Piaget’s concept of decentration—the ability to consider and relate two separate dimensions at a time, such as both length and width. When comparing their performance to that of another child, for example, children can now consider simultaneously the difference in actors’ outcomes and the difference in actors’ effort, which makes ability inferences possible (see Higgins, 1991; Ruble, 1983). To be good at something now means to be better than others at something that is worthwhile, something that matters, as defined by a represented standard of value that is shared with others. Being “bad” or “good” at something thus acquires new significance—it relates to a shared reality about character traits and abilities as inner states of people, which are not present in Phase 3 and are present in Phase 4 or a little earlier (Rholes & Ruble, 1984; Ruble & Rholes, 1981).

The cognitive and social-cognitive changes that occur during Phase 4 are important, if not critical, for the role learning that occurs during this phase. In their classic chapter on role theory, Sarbin and Allen (1968) pointed out that role learning is difficult. Not only must one learn the expectations for a specific role but also the expectations for its complementary roles: “Learning a role adequately requires the learning of the entire role set” (p. 546). It also requires that children tune each time to the specific role set member with whom they are currently interacting. Thus, role enactment with a role set partner involves coordinating the expectations associated with the child’s role with those of the current complementary role. Such coordination would require Case’s (1985) elaborated coordination of two dimensions, which in this case would involve coordination of intentionality dimensions and their functional relation to each other (Case, 1992). According to Case (1985, 1992), this would occur in Phase 4 and not Phase 3.

By the end of Phase 4, children move out of the last stage of childhood. There will, of course, be more developments in shared reality, more developments in the complexity of adult-like shared reality, but the child has taken another major step by now having shared teamwork and shared coordinated roles. Children are more prepared now to engage in adult-related activities, including participating in organized activities and taking on leadership roles in those activities.

Erikson (1950/1963) provided an early proposal about the potential trade-offs that can arise from this Phase 4 shared-reality change when he called it the “Industry vs. Inferiority” age. As he noted, while children seem all set for “entrance into life,” the life being entered is some form of schooling (in the classroom or field or jungle) where personal goals must be “tamed and harnessed to the laws of impersonal things” (p. 258). Erikson considered this phase to be most decisive socially because “industry involves doing things beside and with others, a first sense of division of labor” (p. 260). If all goes well, the child becomes a team player who does his or her part to help the team succeed—industrious. However, if it does not, the child may become a conformist from the pressure to conform, such as conforming to traditional gender-role stereotypes (see Hill & Lynch, 1983).

There is an important difference between children’s shared reality in Phase 3 and the shared reality that emerges in Phase 4. In Phase 3, the shared reality is with significant others, especially parents but other close persons as well. Shared reality is based on the interactions and close relationships with significant others. These shared realities with significant others remain in Phase 4, but what emerges is an additional shared reality with group norms and social rules, involving enacting social roles that are defined with respect to anyone carrying out the role. It is more impersonal and general. Children are evaluated by strangers in terms of general standards that apply to everyone rather than shared self-guides for them as a specific person. Indeed, because of this, there can be a conflict between personal self-guides and group norms, a conflict related to the self-regulatory problem that Goffman (1961) called “role distance.” From an epistemic perspective, enacting roles within a larger organizational unit can also limit children’s thinking to only what is relevant to the role, to the truths contained in their role and its complementary roles.

The change in Phase 4 to self-regulation in relation to goals and standards of the “generalized other” rather than just the goals and standards of the child’s “significant others” has an additional downside. The goals and standards set for children by others, and which they accept for their own self-regulation, are not always in their best interest. What if it would be better for a child to have different or additional goals and standards? Although it is difficult in Phase 3 to change a shared reality with a significant other because significant others typically have more power than the child, there is still some possibility of negotiation or coconstruction of goals and standards for the child because significant others usually have the child’s interests at heart (at least in today’s world). However, the generalized other of Phase 4 does not have a particular child’s interests at heart. It has the organizational activity’s interests at heart (or the interests of the players of that activity). If children do not like the goals and standards of the generalized other in relation to some activity, it is extremely difficult, if not impossible, for them to do anything about it. They would need to create a new consensus among the players, including authorities on the activity who have no personal relationship with them. How do you deal with a generalized other? What does it do to children’s sense of control effectiveness when they need to manage a generalized other?

Not only can the generalized other have different goals and standards for children than their own goals and standards for themselves but also different groups can have different goals and standards for them. Children can experience having different roles or selves in different relationships (e.g., group of friends vs. a close friend; see Harter, 1999). Children’s parents can have one set of goals and standards for them while their teachers have a different set and their peers have yet a different set. There are both benefits and potential costs for Phase 4 children from their exposure to a wide array of different viewpoints. By learning to understand and share a wide variety of viewpoints, children can improve their social perspective-taking skills, which have been found to relate positively to prosocial behavior (Fabes, Carlo, Kupanoff, & Laible, 1999; see also Eisenberg et al., 2006). On the other hand, self-regulatory conflicts can also occur. Early adolescent girls, for example, can experience a conflict between their parents’ wanting them to make schoolwork and household duties their top priorities, whereas their friends want them to make their social life together their top priority. This can create inner conflict for children between their shared-reality reference groups that make successful self-regulation difficult and produce uncertainty and confusion (e.g., Harter, 1986, 1999; Higgins, 1991; Loevinger, 1976). There is evidence that conflict between different self-regulatory guides produces uncertainty and confusion (Van Hook & Higgins, 1988). In addition, there is also evidence that early adolescence is a period of uncertainty and identity confusion (e.g., Blos, 1961; Erikson, 1950/1963; Fischer & Lamborn, 1989).

Because children need to take into account general roles, norms, and rules that apply to everyone, they are also likely to be concerned over being criticized by others, even strangers. There is a “pressure to conform” (see Hill & Lynch, 1983). Trade-offs during this phase can be seen between the benefit of children being able to balance their own needs with the needs of others when participating in group activities, while also being concerned with how others evaluate them in relation to group norms and standards such as becoming very concerned about what they wear and how they look.

There is another kind of trade-off in Phase 4 that relates to the change mentioned earlier from conceptualizing people in terms of physical features or stable behaviors to conceptualizing them in terms of dispositional abilities and traits. This trade-off includes self-evaluation. On the one hand, children can now represent their own repeated successes as reflecting their ability to meet others’ expectations for them, creating a new kind of feeling of self-confidence. On the flip side, they can also represent their repeated failures as reflecting their inability to meet others’ expectations for them, which makes them vulnerable to a new kind of helplessness feeling (see Dweck & Elliott, 1983; Rholes, Blackman, Jordan, & Walters, 1980; Ruble & Rholes, 1981). Combining this challenge with increased concerns with how others evaluate individuals in relation to group norms and standards, this can be a difficult time for children. Indeed, feelings of social anxiety and experiences of peer rejection typically intensify during this phase. In addition, the increased use in self-evaluation of normative social comparisons (i.e., shared standards) contributes to a decline in positive self-evaluations during this phase (see Harter, 2006).

The motivational significance of Phase 4–related shared worldviews is well illustrated by the recent work of Heiphetz, Spelke, and Banaji (2014). Heiphetz et al. (2014) found that 6- to 11-year-old children will attribute prosocial behaviors to those who share their ideological beliefs but not those who share just their shared preferences (like shared interests) or shared facts (like shared names for things). Why this effect of shared ideological beliefs on prosocial judgments? Heiphetz et al. (2014) suggested that religious beliefs have a stronger connection to good and bad behaviors than other beliefs. From a shared-reality perspective, it is also possible that ideologies, such as religious beliefs, relate to objectives and standards for the community that extend beyond individual goal pursuits. Shared ideological beliefs concern what a good member of the ideological group should do and believe and how they should interact with one another. They coordinate the members in terms of a collective interest and not just personal interests. If a person shares your ideological beliefs, that person is a good member of the group. It would make sense that a good member of your group would be a good person; hence the attribution of prosocial behaviors. Such in-group positive judgments of fellow group members would benefit both the individual members and the group as a whole. But, once again, it would also have the potential downside of yielding less positive judgments about those who do not share the ideological beliefs—yet another source of in-group/out-group bias.