2.1 Are descriptions of emotion sharing proto-conversations inclusive across cultures?

In the 1970s, developmental psychologists were documenting that both emotion sharing (i.e., positive, high-intensity emotion) and proto-conversations (a turn-taking exchange of vocal behavior) were within the capacities of human infants (e.g., Bullowa, Reference Bullowa1979). In this section, we evaluate the extent to which this characterization of infant interactive skills is WEIRD or WILD.

2.2 Is joint attention universal and unique to humans?

Joint attention is described in SIT as the ability of infants to coordinate their visual attention to a social partner (usually the mother) with their visual attention to an object (usually a toy). A list of 24 different (WEIRD) operational definitions of joint attention found in the literature (with 26 representative references) can be found in Table 1 in Bard et al. (Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021). In SIT, joint attention definitions include (a) infant attends to an object while the infant also visually attends to the mother (assuming that the mother already is visually attentive to the infant); (b) infants directing maternal attention to an object (e.g., with a “show” or a point); (c) infant and mother sharing a smile while jointly engaged with an object; and (d) the most theoretically laden view – both infant and mother are aware that they each are visually attending to each other, at the same time that they are aware that the other is also sharing attention to the object. Joint Attention is portrayed as a cognitive ability, emerging between 9 and 12 months of age, and is testable, for example, with evidence that infants initiate bids for shared attention, such as by pointing, infants’ expressing a “sharing” smile, or other ostensive behavior by the infant (e.g., showing an object: Tomasello, Reference Tomasello2019).

We can investigate the extent to which contemporary descriptions of the core phenomenon of joint attention are inclusive across cultural settings by adopting a WILD view. The core description of joint attention seems to be applicable for infants and children living in middle-income, two-parent families, with highly educated parents, in Western and often urban societies (i.e., 5–10% of the extant human population: Arnett, Reference Arnett2008; Henrich et al, Reference Henrich, Heine and Norenzayan2010; Nielsen et al., Reference Nielsen, Haun, Kärtner and Legare2017). But it does not fit with all extant humans worldwide. For example, almost all of the current definitions rely on visual attention. For those communities that practice more proximal caregiving, joint engagements are more tactile (e.g., infants sitting on the lap of the partner, while facing outward to others: Aka – Hewlett et al., Reference Hewlett, Lamb, Leyendecker, Scholmerich and Chagnon2017; Kaluli – from Papua New Guinea: Ochs & Schieffelin, Reference Ochs, Schieffelin, Shweder and LeVine1984). As shown in our brief overview of emotion sharing (above), when we take a Worldwide, Local and In situ view, smiles are not nurtured in infants in all cultures, especially in the context of engagement (e.g., Navajo communities in Arizona: Callaghan, Reference Callaghan, Field, Sostek, Vietze and Leiderman2000). Therefore, defining a skill purported to be universal yet requiring one specific type of emotional expression, known not to be universally valued, is problematic.

Infants taking the lead in social interactions is not universal, thus, requiring the infant to be the initiator of joint attention bids would exclude infants living in many societies in which adults direct infant actions (see review in Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021; Morelli et al., Reference Morelli, Bard, Chaudhary, Gottlieb, Keller, Murray and Vicedo2018a). Adult-centered communities expect infants to be obedient, and thus infants do not direct either adult attention (Lancy, Reference Lancy2015) or adult movement (Soman communities: Ochs & Schieffelin, Reference Ochs, Schieffelin, Shweder and LeVine1984). Additionally, infants across many cultures may have relatively few, or no toys (e.g., using instead natural materials or adult tools for play), mothers may not be play partners (Abels et al., Reference Abels, Papaligoura, Lamm and Yovsi2017), and play may not be intended to be educational (Lancy, Reference Lancy2022; Gottlieb & DeLoache, Reference Gottlieb and DeLoache2016). Thus, the SIT description of joint attention fits only with Western, middle-income caregiving practices and artifacts, e.g., distal mother-infant interactions, in an education setting, with toys. It is plausible that joint attention does not exist in infants from non-WEIRD cultural settings, but instead we think that the way in which joint attention has been defined by SIT overlooks the diversity inherent in the capacity of human infants to triadically engage (Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021).

It may also be the case that different views of who takes the lead in interactions are related to more general developmental goals. Chapin (Reference Chapin, Quinn and Mageo2013), in talking about rural Sri Lankans’ caregiving pattern of anticipating infant needs, and comparing this pattern with that of Western, middle-class parents responding to infant cues, states “It seems to me that this way of interacting with young children so that their needs are met quickly and satisfactorily, but without the children themselves having to recognize, articulate, or advocate for ways to meet those needs may lead to a kind of autonomy that is more about self-control than about self-assertion” (p. 159). Indeed, 4-year-old children from rural Nso communities, where infant needs are anticipated, exhibit significantly more self-control in the delay of gratification task than 4-year-old German middle-class children, where infants are expected to signal their needs, and these effects are causally related to socialization practices assessed at 9-months of age (Lamm et al., Reference Lamm, Keller, Teiser, Gudi, Yovsi, Freitag and Lohaus2018).

Bard et al. (Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021) adopt a culturally inclusive definition, which changes our view of early social cognition from WEIRD to WILD, by proposing the concept of triadic connectedness to replace the concept of joint attention. One of the most important aspects of this more inclusive view is that it allows adults to initiate and direct instances of joint engagement about things or events, and is inclusive of engagements in communities in which caregivers anticipate infant needs rather than requiring infants to signal their needs (e.g., Chapin, Reference Chapin, Quinn and Mageo2013; see details in Morelli et al., Reference Morelli, Quinn, Chaudhary, Vicedo, Rosabal–Coto, Murray, Gottlieb, Scheidecker and Takada2018b). This more inclusive view allows for attention and engagement to be in nonvisual modalities (e.g., tactile communication as is found in communities practicing more proximal caregiving), for variety in partners’ emotional tone (e.g., for those valuing calmness), for varieties of shared topics (e.g., for those living in socio-ecologies with a primary focus on social events), and for inclusion of either partner taking the lead (e.g., for those contexts that are relatively more hierarchical and/or relatively more egalitarian than those described by SIT).

Bard et al. (Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021) found that the description of the most typical form of joint engagement (JE) differed across socio-ecological settings for human infants (Figure 1). Although the infant triadically engaged with an adult female and an object was the most typical form in university towns in the UK, this was not the pattern found in the rural Nso and the Aka foraging communities. In these communities, non-females were partners as often as adult females, and joint engagements were equally often about social events as objects. Although the partner most often expressed positive emotion in both the UK and the Nso samples, significantly more JE events in the Aka sample contained no discernible emotion from the partner during joint engagements. Surprisingly, the partner was the predominant initiator in the UK sample, counter to the usual focus on infants as initiators of joint attention (e.g., Tomasello, Reference Tomasello2019).

Figure 1.

(from Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021): Triadic engagements in 1-year-old human infants take different forms in different socio-ecologies, although the inclusive form of joint engagement (JE) occurred significantly more frequently than the WEIRD form of joint attention (JA) in all three settings (Type). The entries reflect the extent to which the partner is more often an adult female (e.g., the mother) or not, the shared topic is more often an object or a social event, the emotion of the partner is positive or not, and whether the infant or the partner initiated the triadic engagement. An entry of “Neither” means that there was no significant difference in the rates of the dichotomized alternatives.

We found that although there was variation in phenotypes (Figures 1 and 2), human infants evidenced triadic connectedness at similarly high levels across the three diverse groups of human 1-year-olds (medians of 65% of observation time in Aka foraging communities, Nso farming communities, and UK middle-class communities). Taking a WILD view allowed us to re-define joint attention to be worldwide and inclusive of diversity.

Figure 2.

(from Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021): Triadic engagements in 1-year-old chimpanzee infants take different forms in different socio-ecologies, although the inclusive form of joint engagement (JE) occurred significantly more frequently than the WEIRD form of joint attention (JA) in chimpanzees living in all three settings (Type). The entries reflect the extent to which the partner is more often an adult female (e.g., the mother) or not, the shared topic is more often an object or a social event, the emotion of the partner is positive or not, and whether the infant or the partner initiated the triadic engagement. An entry of “Neither” means that there was no significant difference in the rates of the dichotomized alternatives.

The cross-species data challenge the view that joint attention is unique to the human species. Bard et al. (Reference Bard, Bakeman, Boysen and Leavens2014) found young chimpanzees were successful in standardized tasks requiring joint attention (with a WEIRD definition), although human infants from the USA were significantly more successful than the laboratory-raised chimpanzees by 12 months of age.

When we took a more WILD perspective and sampled chimpanzees, using naturalistic observations of everyday life, while living in their diverse settings (Chester Zoo in the UK, Primate Research Institute, an enriched laboratory in Japan, home-raised in the USA, and wild chimpanzees at Gombe National Park), and used our inclusive definition of triadic connectedness, we found diversity in the phenotypes (as in the humans), but the chimpanzee infants exhibited triadic connectedness at the same high levels as found in the human groups (i.e., ∼65% of observation time). Figure 2 shows that there is not a “typical” chimpanzee form of joint attention. Although adult females are the most frequent partner in the captive setting, in the wild setting other partners were as frequent. Although social topics were more common in the chimpanzees living in social groups (in the wild or in captivity), the home-raised chimpanzees expressed joint attention with objects as much as with social events. Partner emotion varied, as in the human infants, with chimpanzee partners in the Home-raised group showing significantly more positive emotion but the chimpanzee partners in the other groups showing equal frequencies of positive and non-positive emotion. Wild chimpanzee infants initiated joint engagement significantly more often, but home-raised chimpanzees followed the pattern of the UK human infants in JE more often being partner-initiated. These findings contradict the dual claims of SIT that joint attention is unique to humans, based on evolutionary history, and that joint attention is of a single form in either humans or chimpanzees.

Often differences between a group of apes and a group of humans are misinterpreted as showing species differences when, in fact, there are more proximal factors that can explain the group differences (e.g., different task-relevant experiences: Leavens et al., Reference Leavens, Bard and Hopkins2019). It is usually not possible to match captive apes with humans on pre-experimental experience relevant to SIT. Consider, in a thought experiment, how many pointing episodes children will witness, or participate in, from the age of six months (Figure 3). In most Western, middle-income settings, from 6 months, children develop within a referential cultural environment that focuses on distal signals, and they deploy a range of nonverbal, deictic tools to bring interactants into attentional alignment on a locus of interest to the signaler. A conservative estimate is 20 joint attention episodes per day (during 4 meals a day, during 8 changes of diapers, and 8 episodes of social play – some involving objects, some just face-to-face). If this amount of joint engagement starts from six months of age, by the time they attain 18 months of age, they will have 7,200 such experiences. We note that some non-western, non-middle-class human infants may have substantially fewer experiences (see, e.g., Salomo & Liszkowski, Reference Salomo and Liszkowski2013, for cultural differences in exposure to pointing gestures), but it is usual that great apes are compared with Western, middle-class human infants in SIT. Great apes raised in many captive circumstances will not experience anything close to this number of these types of joint attention, that is, the type of joint attention defined by SIT. We might generously give an institutionalized chimpanzee infant one joint attention episode per day or one joint attention experience every other day (Leavens, Reference Leavens, Cappuccio and Froese2014; Leavens et al., Reference Leavens, Bard and Hopkins2010, Reference Leavens, Bard and Hopkins2019), while acknowledging that non-institutionalized chimpanzees might have many more (e.g., an ape enculturated into Western middle-class activities, such as language-training: Fouts, Reference Fouts1997). At 18 months, institutionally raised chimpanzees would have ∼360 joint attention experiences. These experiences matter, as they are involved in SIT assessments. For example, Leavens and Hopkins (Reference Leavens and Hopkins1998) found that 54% of mother-reared chimpanzees at Yerkes (3–56 years) displayed gaze alternation between a banana and a human experimenter, compared to 80% of nursery-reared chimpanzees. Therefore, if we compare the number of joint attention experiences at 36 months, the western, middle-class human infant would have 18,000 experiences but the institutionalized chimpanzee would have 900. It would take the institutionally reared chimpanzee 50 years to attain the same level of experience in joint attention contexts as the 36-month-old human child from Western middle-class settings.

Figure 3.

(Top) Hypothetical illustration of different intensities of environmental exposure to joint attention episodes expressed as the number of joint attention episodes per day (JA/day) in captive chimpanzees (top matrix, sparse exposure) and WEIRD children (bottom matrix, dense exposure), assuming that learning begins at six months of age. (Bottom) Hypothetical learning curves for measures of receptive joint attention skill as a function of the dense (solid line) and sparse (dashed line) learning regimes depicted in the top panel (figure elaborated from Leavens, Reference Leavens, Cappuccio and Froese2014, Figure 4.4).

In viewing this core foundational concept of SIT, we can conclude that there is a demonstrable WEIRD bias. When we take a WILD perspective, defining joint attention inclusively with the concept of triadic connectedness, we find diversity in phenotypes for both human infants and chimpanzee infants, but consistently high levels of triadic connectedness. We conclude that joint attention must be viewed with a more WILD perspective, and that current WEIRD conceptualizations limit generalizability and hinder appropriate evolutionary explanations of the development of shared intentionality. In Section 4, we describe our proposed approach to increasing inclusivity in SIT, urging psychological theorists to pay greater attention to the diversity of cultural socio-emotional foundations within the social cognition of human and chimpanzee species.

2.3 Evolutionary aspects of SIT

The lack of inclusivity in the human database has implications for evolutionarily grounded developmental theories (Leavens et al., Reference Leavens, Bard and Hopkins2019), such as SIT, in two dimensions. As noted above, patterns of human development typical of Western, middle-class populations have been prematurely taken to be representative of the human species. Similarly, response patterns of captive apes living in institutional settings (e.g., biomedical settings or zoos) have been prematurely taken to be typical of the species, despite clear evidence that rearing history is crucial for the development of task-relevant skills in many assays of social cognition in apes (see Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021, for review). With respect to the comprehension of pointing gestures, for example, great apes who have been enculturated by humans show (a) no apparent deficit relative to human children; and (b) strong performance advantages over institutionalized apes (Lyn et al., Reference Lyn, Russell and Hopkins2010; Russell et al., Reference Russell, Lyn, Schaeffer and Hopkins2011). Like humans, great apes adapt ontogenetically to the signaling conventions relevant to their experiences, whether wild, rehabilitant, institutionalized, or enculturated (reviewed in Krause et al., Reference Krause, Udell, Leavens and Skopos2018). We can see in the performance of captive chimpanzees on the Social Cognition test battery (Herrmann et al., Reference Herrmann, Call, Hernandez-Lloreda, Hare and Tomasello2007), that when chimpanzees are raised in WEIRD-like socio-communicative settings (e.g., given training in linguistic symbols and/or turn-taking exchanges, i.e., WEIRDly enculturated), their performance is comparable to that of human children, aged 2.5 years and raised in middle-class families in a German city (Russell et al., Reference Russell, Lyn, Schaeffer and Hopkins2011). Given that enculturated chimpanzees perform significantly better than chimpanzees raised under standard laboratory conditions, we can argue that a WILD approach is essential for both sides of any species comparisons.

Therefore, species comparisons are compromised when cross-species comparisons primarily rely on one sample to represent each species. Lack of sample generalizability affects the evolutionary aspect of theory development in two primary ways. First, direct ape-human comparisons, as currently conducted in the scientific literature, have demonstrated little more than the unsurprising finding that captive (or wild) great apes do not act very much like children raised in WEIRD environments. To the extent that developmental patterns in WEIRD populations have been over-generalized to the entire human species, this bias blinds researchers to potentially informative similarities in how human and ape juveniles respond to their specific ecological circumstances. An example is the ubiquity of pointing in chimpanzees who have been cross-fostered with humans, considered against the relative paucity of pointing in wild chimpanzees (Hobaiter et al., Reference Hobaiter, Leavens and Byrne2014), which demonstrates the great variability in phenotypic expressions found in apes from different rearing environments (Leavens et al., Reference Leavens, Bard and Hopkins2010). Thus, the evolutionary picture against which human development is understood has been significantly distorted by adherence to the developmental patterns and contexts of unrepresentative samples of both humans and apes.

Prima facie, it might seem to be the case that great apes are advantaged in tests of their sociocognitive abilities when they are considerably older than the humans to whom they are being compared. We know of two published studies that have directly tested this assumption: Horowitz (Reference Horowitz2003) and Thomas et al. (Reference Thomas, Murphy, Pitt, Rivers and Leavens2008). A study by Whiten et al. (Reference Whiten, Custance, Gomez, Teixidor and Bard1996) tested imitative fidelity in children and slightly older juvenile chimpanzees, finding that the children displayed higher fidelity in copying manual actions than did the apes. Although the apes were sometimes much more efficient in opening the puzzle box, they were less likely to open the box in the manner demonstrated to them. Tomasello (Reference Tomasello, Galef and Heyes1996) interpreted this difference in response profiles to indicate that the children displayed a higher-order cognitive capacity (“true imitation”) than did the chimpanzees, who merely “emulated” the actions of a demonstrator on the puzzle boxes. Horowitz (Reference Horowitz2003) administered the same puzzle boxes to human adults, finding that, like the chimpanzees in Whiten et al. (Reference Whiten, Custance, Gomez, Teixidor and Bard1996), they displayed “emulation” much more than “true imitation.” This shows that the chimpanzees in Whiten et al. (Reference Whiten, Custance, Gomez, Teixidor and Bard1996) displayed a response profile more representative of human adults than of children (see Leavens et al., Reference Leavens, Elsherif and Clark2023, for a review). Similarly, and of particular relevance to SIT, Povinelli and his colleagues (Reference Povinelli, Bierschwale and Čech1999) tested 3-year-old children on their abilities to follow experimenter cues to the location of rewards in one of two opaque containers, comparing their performance with 6-year-old chimpanzees. In two of three cue conditions, there were no performance differences between the children and the older chimpanzees, but in a third condition, where a cue-giver looked at the same side as, but far above a baited container, the children performed at chance, whereas the chimpanzees performed significantly above chance. Notwithstanding the superior performance of the apes in this third condition, Povinelli and his colleagues interpreted this group difference in response profiles as evidence for a higher-order understanding of visual attention by the children. When Thomas and her colleagues (Reference Thomas, Murphy, Pitt, Rivers and Leavens2008) administered a similar task to young human adults, they responded more like the chimpanzees in the original study by Povinelli and colleagues. Thus, in both of these cases, the response profiles of the children in Whiten et al. (Reference Whiten, Custance, Gomez, Teixidor and Bard1996) and in Povinelli et al. (Reference Povinelli, Bierschwale and Čech1999) were not representative of the response profiles of human adults, but the performances of the older chimpanzees were strikingly similar to those of human adults, who presumably brought to these tasks the highest-order cognitive apparatus of all of the participants. Therefore, when any systematic group difference is found between children and older apes, there is a logically dubious tendency to interpret the difference to the advantage of the children, despite their unrepresentativeness of human adults. In other words, without an age-matched control group, children’s performance profiles in direct ape-human comparisons are not clearly interpretable (Bard & Leavens, Reference Bard and Leavens2014; Leavens et al., Reference Leavens, Bard and Hopkins2019, Reference Leavens, Elsherif and Clark2023).

The second way that evolution has been problematically invoked in SIT concerns an unwarranted commitment to the assumption that human children in WEIRD cultures display their developing communicative skills because of the effects of natural selection over evolutionary time – rather than attending to the culture-specific practices that children directly experience. This leads to the fallacious attribution of virtually any developmental phenomenon to be the consequence of ancient selection pressure (albeit poorly specified). “On the contrary, if anything applies universally to all human beings, it is the fact that we differ from each other in many ways because our developmental trajectories differ from one another in many ways, and this is a fact of culture as much as of nature because in the end the two are not distinct” (Sanches de Oliveira & Baggs, Reference Sanches de Oliveira and Baggs2023, p. 25). SIT invokes evolutionary selective pressures to support and promote theoretical commitments to human prime cognitive movers (Tomasello, Reference Tomasello2019). This fallacious approach to evolutionary psychology is corrected when it is acknowledged that culture is entwined with development at every stage (Sanches de Oliveira & Baggs, Reference Sanches de Oliveira and Baggs2023) and that evolution has acted on culture and development (Jablonka & Lamb, Reference Jablonka and Lamb2007). We note that all developing organisms, including captive apes, are ecologically situated, and that systematic differences in rearing, learning, and socialization environments can illuminate those parameters which influence developmental change across and within species. “Elaborating on the relationship between culture and development first requires recognizing that evolution and development are not mutually exclusive, then building on that insight to explore how evolved developmental mechanisms that are sensitive to cultural influence can create psychological and behavioral variation across and within societies” (Kline et al., Reference Kline, Shamsudheen and Broesch2018, p. 1–2).

One of the central claims of SIT is that humans have evolved cognitive specializations for social cognition (Tomasello, Reference Tomasello2019). The idea of human species-specific cognitive specializations is not, in itself, implausible. For example, there are many human behavioral competencies that have no obvious parallel in the behavioral repertoires of our nearest living relatives, the great apes (Hominidae), such as infantile vocal mimicry, the articulation of speech, and the formation of vast lexicons comprising tens of thousands of elements. These language-related behavioral competencies can be plausibly viewed as indices of uniquely human cognitive (information-processing) adaptations. Here, however, we are concerned with particular research evidence in wide contemporary use, that forms the evidential basis for many claims of human cognitive exceptionalism: species differences in social cognition (for additional examples, see Sauciuc & Persson, Reference Sauciuc and Persson2023).

Thus, with respect to pointing, for example, which was long held to be an evolutionarily given, species-specific human gesture (Butterworth, Reference Butterworth and Kita2003), more recent evidence shows that apes and other animals readily adopt pointing in some developmental circumstances (Krause et al., Reference Krause, Udell, Leavens and Skopos2018). The fact of animal pointing has forced a re-examination of the universality of manual pointing in humans. For example, apes do not evince a single canonical deictic gesture, and re-examination of human developmental and ethnographic findings shows that there is not one canonical deictic gesture for humans, either (Krause, Bard, & Leavens, Reference Krause, Bard and Leavensin press). Because the frequency and forms of pointing differ systematically across ape groups in accordance with the intensity of their exposure to human interactions (Leavens et al., Reference Leavens, Bard and Hopkins2010), this demonstrates that pointing gestures in our nearest living relatives have low heritability and high environmentability. This finding authorizes a search for environmental influences on gesture development in our own species. Today, after decades of scientific theory that viewed pointing with the index finger as an evolved communication adaptation for fostering the acquisition of speech, we can now see that steeping apes in referential communities fosters their development of deictic gestures, reflecting the apes’ experiences with human cultural nonverbal deictic conventions. This opens the door to further exploration of similar constructionist mechanisms in the development of human nonverbal deictic signals (as suggested by Call and Tomasello, Reference Call and Tomasello1994). The potential scientific value of the great apes, in particular, to developmental psychologists, and researchers interested in the evolution of language, will increase when we take more seriously within-species differences for both humans and great apes (Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021).

SIT has relied on the two-group cross-species design, which is inherently inadequate to illuminate the influence of biological adaptation (Leavens et al., Reference Leavens, Bard and Hopkins2019). Moreover, SIT relies on the argument that if adult apes (raised in BIZARRE settings) differ from human children (raised in WEIRD settings) then one can conclude any observed difference is a result of a biological adaptation that is not due to incomplete cognitive maturation in the apes, or limited lifespan experience in the apes. While we acknowledge that adult apes are cognitively mature and have lifetime experience, we argue that it may be exactly those years of differential experiences (as we illustrate in Figure 3), coupled with different ages at testing, that can explain the observed group differences in these types of studies (Bard & Leavens, Reference Bard and Leavens2014; Leavens et al., Reference Leavens, Bard and Hopkins2019; Thomas et al., Reference Thomas, Murphy, Pitt, Rivers and Leavens2008). The general method we advocate for, here, acknowledges that no one sample of a hominid species will be representative of the range of environments in which those organisms develop. Therefore, scientific resources would be better spent sampling response profiles across as wide a range of rearing environments and cultures as possible (Worldwide), then looking for plausible parameters in the (Local) environments of these different groups that vary systematically with performance differences (Diversity) both within and between groups (i.e., use a multiple group design, investigating In situ behavior, for each species: Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021; Leavens et al., Reference Leavens, Bard and Hopkins2019; Newson et al., Reference Newson, Buhrmester, Xygalatas and Whitehouse2021).

This enduring and pervasive bias towards evolutionary explanations of what are, in many cases, developmental adaptations has significantly skewed the starting assumptions of philosophers, paleoanthropologists, developmental psychologists, psycholinguists, cognitive scientists, and many others, particularly those interested in developing models of the evolution of language. One strong advantage to reconsidering the foundations of theory in this area is that, to the extent that humans and their nearest living relatives, the great apes, are ontogenetically plastic, is the extent to which apes can serve as animal models for the last common ancestor of apes and humans, which was, presumably, a non-speaking ape (de Waal, Reference de Waal2022; Leavens et al. Reference Leavens, Elsherif and Clark2023).

2.4 Is the Shared Intentionality theory inclusive?

The initial conceptualization of Shared Intentionality (in hypothesis form) was 20 years ago (Tomasello et al., Reference Tomasello, Carpenter, Call, Behne and Moll2005). It is possible that since the initial position, the theory has incorporated cross-cultural information (and possibly interpreted it differently than the way we do). Therefore, here we consider the extent to which cultural diversity is considered in more recent conceptualizations of the theory. We apply the technique of Cheon et al. (Reference Cheon, Melani and Hong2020) to evaluate the extent to which non-WEIRD samples are included in the database. We consider two recently published presentations of SIT, specifically, Tomasello’s (Reference Tomasello2022) article claiming to specify the unique characteristics of social cognition in the human species and Tomasello’s (Reference Tomasello2018) review of the literature to support the claim that SIT makes sense of the purported human-unique ability of Theory of Mind. Note that these are only two examples, from many articles by many authors that endorse and/or apply SIT (e.g., Krupenye & Call, Reference Krupenye and Call2019; Mundy, Reference Mundy2018). These two articles were chosen as they are relatively recent, and likely represent the working database of SIT, because they were written by the original SI theorist.

Tomasello’s Reference Tomasello2022 article contains 65 references, and 35 references indicate human participants. All these references were from WEIRD cultural contexts. In other words, all foundational data pertaining to human development in this article were from infants or children living in Western, middle-class settings. This article, published in a high-impact journal, describes a “uniquely human” type of social cognition, but there is no evidence of inclusion nor any notice about the limitations of generality (as recommended by Simons et al., Reference Simons, Shoda and Lindsay2017).

In Tomasello’s Reference Tomasello2018 review article, there were 86 references. Five titles indicated studies with non-WEIRD samples (< 6% of the total). However, this review article relied on databases from 16 references that were opinions, comments, reviews, or meta-analyses. It was plausible that viewing these primary sources would reveal more cultural inclusion, so we additionally searched all the citation lists of these 16 references. Although this increased the number of articles considerably (adding 1265 articles), we found only 43 titles indicating the use of samples from non-WEIRD settings (3%). Therefore, considering the two articles (Tomasello, Reference Tomasello2018, Reference Tomasello2022), and the 16 references providing databases (from many additional authors), we find that over 96% of the database (1322 of 1370 titles) is based on WEIRD samples. Therefore, it appears that SIT, and the previous incarnation, SIH, neglect consideration of the 90% of human infants and children who live in a wide variety of non-WEIRD settings (Arnett, Reference Arnett2008; Nielsen et al., Reference Nielsen, Haun, Kärtner and Legare2017). The SIT monocultural narrative is neither universal nor inclusive.

Although our major focus here is on the lack of inclusivity in the database of early social cognition, we note that there is also a lack of inclusivity in purported characteristics of the social cognition of older individuals. A recent review of the ethnographic evidence (Kramer, Reference Kramer2022), for example, makes it clear that collaboration, as conceptualized in SIT, is not inclusive of collaboration and cooperation as practiced in rural, indigenous, foraging, farming, and small-scale communities. Moreover, the highly cognitive explanation of cooperation advocated by SIT does not mesh with the way in which cooperation is manifest in the daily activities of some cultural communities (e.g., Nso: Ndzenyuiy & Keller, Reference Ndzenyuiy, Keller and Verschure2025; Savannah Pume’ hunter-gatherers: Kramer, Reference Kramer2022), or in cooperative activities (e.g., childcare, foraging, food processing, resource process, trading, daily labor, rituals, coalitions: see Table 1 in Kramer for review of specific cultural practices). In the Nso, for example, cooperation in farming, occurs organically, as a person walks to their field, and chats with neighbors along the way, and some join in with the person in tilling the one field. Cooperation in childcare occurs as a result of unspoken cultural practices, for example, if there is someone close to an infant, that person will engage in caregiving activities, as needed. The SIT conception of cooperation as cognitive plans, with iterative acknowledgement of others’ intentions, may not fit the way that cooperative activities are undertaken on the ground. The Nso conception of cooperation is as a way of life. It is similar to the Ubuntu philosophy “ubuntu ngumuntu ngabantu” (Zulu language) emphasizing “reliance on others to achieve ones’ potential,” a philosophy that has shaped living patterns of many African communities (Ndzenyuiy & Keller, Reference Ndzenyuiy, Keller and Verschure2025, p. 134).

2.5 Does the Shared Intentionality theory generalize?

For a theory to claim universality, the minimal requirement is that the database upon which it is based is inclusive and the results generalizable, that is, Worldwide, both in the sense of an inclusive database and in the sense of generalizing to humans living in all cultural settings. There is no evidence that SIT generalizes beyond WEIRD settings, despite the very common assertion that SIT explains what “the human species” does. We have provided evidence that the core foundations of SIT do not generalize even to the few cultural settings that we discussed, thus, SIT outcomes do not represent outcomes in the entire human species. It is especially problematic that without evidence of generalization, the studies that use SIT should, but rarely do acknowledge that their database is neither inclusive nor likely representative of global humanity.

Furthermore, the evolutionary argument used by SIT to support its claims of unique human developmental pathways is flawed in its logic and its application (e.g., Leavens et al., Reference Leavens, Bard and Hopkins2019). For one example, virtually no direct ape-human comparison has matched the apes and the humans on life-history stage; thus, essentially all claims of evolutionarily-based differences in social cognition are confounded with age differences – it is not clear whether apes and humans sometimes differ in these comparisons because they are tested at different stages of life or because of their different biologically adaptive histories (e.g., Bard & Leavens, Reference Bard and Leavens2014; Leavens et al., Reference Leavens, Elsherif and Clark2023). Moreover, few direct ape-human comparisons match the groups on pre-experimental or pre-observational life experiences (Leavens et al., Reference Leavens, Bard and Hopkins2019). When researchers are faced with differences between groups in ape-human comparisons, these numerous confounded factors prohibit any unambiguous attribution of the cause of the group difference to evolutionary history. In other words, with these kinds of research designs, it is impossible to know whether any group differences in response profiles are because the apes were apes and the humans were humans, or differences were caused by confounded non-evolutionary factors. Thus, the most commonly used method for illuminating the putative evolutionary foundations of human social cognition, the direct ape-human comparison, fails to control for the influence of individual learning experiences and local developmental context. This systematic weakness permeates the contemporary scientific literature, and results in large part from over-generalizing from both the human samples (taking WEIRD norms as paradigmatic for the human species) and the ape samples (taking captive apes that are convenient to sample as representative of their respective species).

In summary, we find that the Shared Intentionality Theory (Tomasello, Reference Tomasello2019) is based on WEIRD and not WILD perspectives. We find that there is a WEIRD bias in many steps of the research process, including selecting a topic, framing the research questions, selecting participants, developing assessment procedures, devising the context of assessment, to finally interpreting the data (Broesch et al., Reference Broesch, Lew-Levy, Kärtner, Kanngiesser and Kline2022; Sanches de Oliverira & Baggs, Reference Sanches de Oliveira and Baggs2023; Singh, Reference Singh2024). We have touched on these issues throughout this section. We conclude that this developmental theory is based on a relatively monocultural database, espouses relatively monocultural constructions of developmental pathways, and does not explain human-specific social cognition. This bottom-up theory is neither inclusive nor generalizable. We conclude that this theory does not meet the requirements of a WILD perspective (see section 4 for how to adopt a WILD perspective for Shared Intentionality Theory).

3.1 The universality assumption of Attachment theory

The core premise, in this top-down theory, is that all children form an attachment with a primary figure, if there are no extraordinary hindrances. The attachment system is assumed to function with the same purpose and the same behavioral patterns in all infants around the world. Universality is based on Bowlby’s understanding of evolutionary theory: he regarded attachment as a mechanism to increase survival of the child through protection from environmental hazards, especially predators, with its universal form evolving within the Environment of Evolutionary Adaptedness. This is based on an incorrect understanding of evolution: from an evolutionary perspective, both behavior and its development are considered as adapted to the local context, excluding universal solutions (e.g., Keller, Reference Keller2022; Jablonka & Lamb, Reference Jablonka and Lamb2007; Sanchez de Oliveira & Baggs, Reference Sanches de Oliveira and Baggs2023). Moreover, the universality assumption is extended to all other aspects of the conceptualization of attachment, implying that the emergence of attachment follows a universal pathway, the behavioral correlates of attachment and its qualities apply equally well to all children, and the consequences of early attachment qualities are the same everywhere. However, it is well-known that many aspects of early social-emotional relationships differ across cultures (see Gottlieb & DeLoache, Reference Gottlieb and DeLoache2016; Keller & Bard, Reference Keller and Bard2017; Otto & Keller, Reference Otto and Keller2018; Quinn & Mageo, Reference Quinn and Mageo2013), including the specificity of the attachment figure, the type of emotional atmosphere surrounding infants during their early development, and perhaps most importantly, the core function of the attachment system (i.e., to support infants’ developing independence: Gaskins et al., Reference Gaskins, Beeghly, Bard, Gernhardt, Liu, Teti, Thompson, Weisner, Yovsi, Keller and Bard2017; Keller, Reference Keller2022).

Although relationships can be regarded as important worldwide for infant survival, attachment-defined solutions can differ across cultures. In nuclear families living in WEIRD settings, secure attachments with one or both parents may be the optimal solution. However, most infants in the world live in small-scale or subsistence settings, where having an extended relational network is a more optimal solution. There are numerous benefits to having many attachment figures, including providing more safeguards, also in the absence of the mother; infant survival is often dependent on communal-level food provisioning; infants learn how to sustain and develop different social relationships with partners of different age/sex classes. Consideration of attachment networks, moreover, reveals another core function of the attachment system, that is, integration into an interdependent social network via introduction to social learning (Gaskins et al, Reference Gaskins, Beeghly, Bard, Gernhardt, Liu, Teti, Thompson, Weisner, Yovsi, Keller and Bard2017). Developing a communal identity is fostered by trusting, cooperative relationships with the cultural community as a whole, which starts in infancy. Thus, the core motivating force proposed by AT of developing independence is in direct contradiction to the core motivation of many human communities, which is developing interdependence.

Attachment researchers have claimed to prove empirically that attachment is universal with evidence consisting primarily of applying the Strange Situation Procedure (SSP), or its equivalent, – mainly the attachment Q sort (E. Waters & Deane, Reference Waters and Deane1985), to infant samples from different regions or countries. This is problematic for a number of reasons. First, these studies lack a conception of culture, often making the mistake of equating country with culture (e.g., issues discussed by Chaudhary et al., Reference Chaudhary, Misra, Bansal, Valsiner and Singh2022; Keller, Reference Keller2022). Thus, context and culture are not recognized as the driving forces in the formation of behavioral patterns throughout development. Secondly, attachment is assessed with the methods that attachment researchers developed for Western, middle-income samples (see section 3.4 below for discussion of these issues), without empirical evidence that the concepts being tested generalize to other cultural contexts (Vicedo, Reference Vicedo2013; Sanches de Oliveira & Baggs, Reference Sanches de Oliveira and Baggs2023). Accordingly, cultural anthropological and psychological ethnographic research that takes a WILD methodology, including local contexts of infants’ caregiving environment and caregiving networks, is ignored. Although ethnographic research (observations and interviews in the natural environment) was the tool that Ainsworth used in her first attachment study in Uganda, before her death, she wrote about her regret that ethnographic research was abandoned by AT (Ainsworth & Marvin, Reference Ainsworth and Marvin1994, p. 12).

The universality assumption is like the meta-assumption of AT, since it is the implicit, but pivotal foundation for all other assumptions, including: (a) The assumption that the major goal of infancy is to balance exploration of the physical world with safety given by a primary attachment figure; (b) the assumption that independence, as observed in Western, middle-class settings, is the normative developmental goal during infancy, rather than development goals following a path aimed more at interdependence; and (c) the assumption that infants’ psychobiological system is tuned to regulating the distress of being separated from the primary attachment figure, rather than having an attachment network with regulation provided by more proximal contact and rarely, if ever, experiencing being alone (e.g., Gaskins et al., Reference Gaskins, Beeghly, Bard, Gernhardt, Liu, Teti, Thompson, Weisner, Yovsi, Keller and Bard2017; Morelli et al., Reference Morelli, Quinn, Chaudhary, Vicedo, Rosabal–Coto, Murray, Gottlieb, Scheidecker and Takada2018b). The universality assumption has far-reaching consequences because the hidden assumptions revolve around this claim that attachment is not highly culture-specific. Moreover, Attachment Theorists continue to assert that the form of attachment manifest within WEIRD families is both normative and the gold standard (two major flawed assumptions of WEIRD psychology: Kline et al., Reference Kline, Shamsudheen and Broesch2018; but see new acceptance of attachment networks by approximately 70 AT practitioners, Forslund et al., Reference Forslund, Granqvist, van IJzendoorn, Sagi-Schwartz, Glaser, Steele and Duschinsky2022).

3.2 The assumption of normativity in Attachment theory

Another core assumption of AT is that secure attachment is the most optimal attachment quality universally and is most prominent, if adversity and other life-threatening circumstances do not interfere. This assumption is based in Bowlby’s misguided evolutionary ideas of adaptive behavioral patterns (see above section 3.1), which ignore the vital role of local context as crucial for behavioral adaptations. Evolution does not have a goal; therefore, the claim that secure attachment is the best strategy worldwide is based on an incorrect understanding of evolution. Although infant survival may be a common ultimate function of caregiving, it is the case that cultural practices, social systems, and indigenous beliefs create proximate solutions. The proximate solutions of nuclear families living in recently emergent post-industrial settings (e.g., WEIRD settings) cannot be considered the norm for the human species.

There is substantial evidence that human psychology varies across contexts and cultures, which impacts how social relationships, including attachment, are defined and regulated. Although cross-cultural views have been posited (e.g., about Japanese or other indigenous understanding of attachment: e.g., Rothbaum et al., Reference Rothbaum, Weisz, Pott, Miyake and Morelli2000), they are disregarded often by attachment researchers who tend to look for confirmation only. The vague definition of attachment as an emotional bond is problematic since we know that there are substantial differences in emotional experience, understanding, and expression across cultures (see section 2.1.2 above; Mesquita, Reference Mesquita2022). AT posits a WEIRD form of the emotional bond, that is expressed mainly as affection in positive facial expressions and concomitant verbalizations (including praise). But across cultures, we see that the emotional bond can be expressed as more action-oriented – feeding and food play a special role in many cultural communities (Funk et al., Reference Funk, Scheidecker, Chapin, Schmidt, El Ouardani and Chaudhary2023; Schmidt et al., Reference Schmidt, Keller and Rosabal Coto2023) – or the emotional bond can be manifest mainly in physical care, without distal emotional expressions from caregivers (Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021) or infants (Keller, Reference Keller2007). The emotion that bonds infants with others may not be positive or may not be mainly expressing affection (see section 2.1.2). There are similar definitional problems with other attachment concepts that have resulted from the lack of cultural inclusivity (Keller, Reference Keller2022). Therefore, it is certain that the norms of AT are biased toward infants living in WEIRD settings, and are not based in a WILD view.

3.3 Attachment theory’s assumptions of sensitivity and competence

There are additional core assumptions of AT, for example focusing on the characteristics of the attachment figure, such as maternal sensitivity and the consequences of “optimal”/normative attachment, e.g., competence. Discussion of the lack of cultural inclusion in these concepts can be found elsewhere (e.g., Keller et al., Reference Keller, Bard, Morelli, Chaudhary, Vicedo, Rosabal-Coto, Scheidecker, Murray and Gottlieb2017; Keller, Reference Keller2022). Here we briefly present the cultural exclusion in other hidden assumptions of AT (summarized from Keller, Reference Keller2022). These conceptions are philosophical/ideological metatheoretical constructs about human nature and, therefore ought to be culture-conscious, by definition.

3.4 Lack of inclusivity in Attachment theory

In summary, AT was developed with a theoretical orientation of explaining emotional bonds within nuclear families, typical of infants living in western, middle-income settings, based in a deficit model of post-World War II childhood and positing a universal foundation in evolution. There is a lack of inclusivity in all aspects of AT, starting with the core assumption of universality (ignoring convincing evidence of diverse culture-specificities), and continuing to claims of normativity (based, e.g., on SSP). This WEIRD perspective at the theoretical level impacts the methods used for testing and, thus, for providing information about normativity: “Biased tools will produce a biased dataset for the norming” (Oppong, Reference Oppong2023, p. 2). Furthermore, AT has the core assumption of a singular function of an attachment figure (a secure base for infants’ developing independence), which lacks appropriate WILD considerations of attachment functions, for example, networks can function to integrate infants into social communities from which infants develop interdependence. The core concepts of maternal sensitivity, dyadic interactions, and distal communication are all based in WEIRD practices (see Section 2) and ignore common varieties of infant interactions with others (e.g., in people of the global majority). AT, from the top-down, is not inclusive. Moreover, there is an urgent need to revise the theory as the WEIRD bias is used in the design of many otherwise well-meaning interventions. These interventions often require participants to engage in WEIRD practices that undermine not only their own valid cultural practices of child rearing, but importantly can undermine their own valid indigenous belief systems (e.g., Keller et al., Reference Keller, Bard, Morelli, Chaudhary, Vicedo, Rosabal-Coto, Scheidecker, Murray and Gottlieb2017). These interventions are decidedly not guided by WILD principles, as they do not include diversity according to local culture, they are deficit based (because they use WEIRD norms; Kline et al., Reference Kline, Shamsudheen and Broesch2018), and, therefore, can be considered unethical (Scheidecker et al., Reference Scheidecker, Chaudhary, Keller, Mezzenzana and Lancy2023), or as preconditions for epistemological violence (Oppong, Reference Oppong2023).

4.1 The five steps

The core initial step involves gathering a cultural understanding via a cultural investigation of the phenomena, specifically, in situ observations, gathering culturally informed knowledge of their meaning (e.g., de Barbaro & Fausey, Reference de Barbaro and Fausey2022; Yahalom & Fein, Reference Yahalom and Fein2025). This cannot be obtained from conducting experiments with inherent WEIRD biases. Nor can this essential first step be successfully accomplished by observing only those behaviors that are the focus of existing WEIRD theories (e.g., independent exploration with the mother as a secure base as derived from AT). This step must also include local investigations of caregiving beliefs and cultural ethnotheories and consider these within the context of diversity across cultural settings (i.e., Worldwide). No single culture is representative of the entire human species, rather the human species must be described in a manner that respects diversity (e.g., Broesch et al., Reference Broesch, Lew-Levy, Kärtner, Kanngiesser and Kline2022; Kline et al., Reference Kline, Shamsudheen and Broesch2018; Newson et al., Reference Newson, Buhrmester, Xygalatas and Whitehouse2021; Sanchez de Oliveira & Baggs, Reference Sanches de Oliveira and Baggs2023)

Once the culturally relevant behaviors and practices associated with the phenomenon of interest have been discovered, the second step involves investigating their developmental course. These initial investigations of local developmental pathways also require inclusive consideration of other samples to obtain core knowledge of the diverse nature of the human species. We should be clear, as Chaudhary et al. (Reference Chaudhary, Misra, Bansal, Valsiner and Singh2022, p. 410) have stated “For such a task, dissolving existing binaries like East and West, North and South is vital, with an accompanying elaboration of precise cultural historic specificity.” Developmental outcomes and processes, especially those proposed to be universal, cannot be understood by studying only relatively homogenous samples, composed of urban-dwelling, middle-income, 2-parent families living in rich economies. We must acknowledge that a single narrative will not be sufficient to describe the human species, even when focusing narrowly on infancy (Chaudhary et al., Reference Chaudhary, Misra, Bansal, Valsiner and Singh2022).

The third step in the process of decentering psychology theories is to arrive at an understanding of the phenomena that is both inclusively defined and applicable to infants (and children) living across diverse, socio-eco-cultural settings (e.g., Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021; Lancy, Reference Lancy2022). With this inclusive knowledge in hand, scientists can then begin the fourth step, which is to use the proposed design to investigate developmental phenomena across multiple, diverse samples, with possible aims to understand (a) universals; (b) normative behaviors; and/or (c) links between particular social, ecological, and cultural variables and outcomes, via human-general processes. We note that some are questioning whether the goal of developmental theories ought to be to find universals (Kline et al., Reference Kline, Shamsudheen and Broesch2018; Sanchez de Oliveira & Baggs, Reference Sanches de Oliveira and Baggs2023), agreeing that there might be some outcomes, such as triadic connectedness, that can be described as a universal concept that take local forms (Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021). We also note that comparisons across cultural groups can be made in different ways, e.g., by taking a homogenous sample of working-class families in India to compare with a homogenous sample of working-class families in USA to assess the impact of culture on the working-class family. Alternatively, in-depth understanding of a single group in a specific culture can be helpful to understand developmental pathways, as anthropologists have traditionally shared (e.g., Gottlieb & DeLoache, Reference Gottlieb and DeLoache2016). In some domains of psychological research, such as in developmental psychology, significant effort is expended to match comparison groups on as many potentially influential variables as possible. This matching can take one of two forms: we can match individuals across groups, or we can identify key parameters and attempt to match the groups on these aggregate measures.

These steps are also applicable to investigate the phenomena of interest in cross-species comparisons, using multiple, diverse groups of each species (e.g., comparing infant development in a zoo-raised group to that in a lab-raised group or in a group living in the wild) with an aim to understand great ape-general outcomes and processes. A major problem with many current comparative studies is the lack of any serious attempt to tackle the issue of confounding variables (e.g., Bard & Leavens, Reference Bard and Leavens2014; Leavens et al., Reference Leavens, Bard and Hopkins2019).

An important final step in becoming more inclusive is to avoid unconscious bias. The privileging of a monocultural worldview, while excluding others (i.e., the mindset of theorists) seriously hinders the authentic internationalization/globalization of psychology (Chaudhary et al., Reference Chaudhary, Misra, Bansal, Valsiner and Singh2022, p. 399). To ensure that conclusions do not prioritize or privilege the WEIRD versions of the phenomenon, scientists can search instead for how local social, ecological, and cultural variables influence phenotypic behavioral expressions (e.g., see Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021) and other culture-specific adaptations (Keller, Reference Keller2022; Lancy, Reference Lancy2022; Scheidecker et al., Reference Scheidecker, Chaudhary, Keller, Mezzenzana and Lancy2023). Although this final step is not touted as part of the WILD agenda, it is clearly one of the intended outcomes. It is vitally important that these endeavors respect differences across cultures, and not equate differences with deficits (Kline et al., Reference Kline, Shamsudheen and Broesch2018). Increasing inclusivity, appreciating diversity, is required to make developmental theories representative and explanatory of worldwide human populations.

4.2 Applying the five steps to the Shared Intentionality theory

We illustrate the value of embracing diversity in both developmental outcomes and processes surrounding shared intentionality, as it brings new insights and highlights deficiencies in assumptions and premises of SIT. Applying the five steps toward inclusivity for joint attention, Bard et al. (Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021) suggest necessary revisions to SIT: (1) triadic connectedness, not joint attention, is the core social-cognitive capacity of 1-year-olds. Taking a WILD perspective, we could extend “triadic” to “multiparty,” although there was little evidence of this in 1-year-olds in the Bard et al. (Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021) study. This conceptualization allows for (a) cultural diversity in engagement style (e.g., engagements can be tactile as well as visual, demoting the importance of gaze); (b) cultural diversity in shared topics (engagements can be about a variety of topics, demoting the importance of toys or objects); (c) cultural diversity in affect accompanying triadic connectedness (cultures differ in their preference for high arousal and/or positive emotional expression, demoting the importance of e.g., “shared smiles”); (d) cultural differences in potential social partners (cultures differ in reliance on a single parent or multiple adults or children as partners for infants), and (e) cultural differences in initiator (in many cultures, children follow the lead of the adult and/or adults anticipate infant needs, demoting the importance of infant signaling).

The second necessary revision of SIT is to dispel the myth that triadic connectedness (or even joint attention) is uniquely human. Equally high levels of triadic connectedness are present in both human and chimpanzee 1-year-olds, and the behavioral forms found in some chimpanzee infants are the same as the behavioral forms found in some human infants (Bard et al., Reference Bard, Keller, Ross, Hewlett, Butler, Boysen and Matsuzawa2021).

The third necessary revision of SIT is to include understanding at the behavioral level that there are diverse forms of triadic connectedness (Figures 1 and 2). The nature of social cognition is to be highly contextualized, given that infants around the world have different social experiences. Developmental theories should be proposing ways to understand how these different social experiences impact developing social cognition. We are led to the conclusion that phenotypic expressions are proximal adaptations to local socio-eco-cultural conditions.

Therefore, we conclude that the critical features of the SIT are neither universal nor normative for the human species. These features range from the proposed developmental paths to definitional behavior patterns (e.g., infants look to others and smile in order to share experience triadically; infants invite others to share interest or direct adults’ attention). Either we relabel SIT as specific to Western, middle-class socio-ecological settings (WEIRD SIT), or the theory must change. Through this process of considering how to increase inclusivity, we find embracing a WILD perspective (Newson et al., Reference Newson, Buhrmester, Xygalatas and Whitehouse2021) enhances our understanding of developmental foundations of the diverse social experiences that are foundational to human social cognition.

4.3 Applying the five steps to Attachment theory

Embracing diversity in socio-emotional development brings new insights that undermine the core assumptions of Attachment Theory (AT). Specifically, following the five steps of inclusivity reveals that there are culturally diverse beliefs, caregiving practices, and socio-ecologies supporting (a) attachment networks of adults, children, and infants, demoting the importance of monotropy (primacy) and dyadic parent-infant relationships; (b) alternate/additional functions of attachment, e.g., attachment figures function to provide privileged entry into the social world and cultural learning, demoting reliance on the singular function in AT of regulation of distress from separation; (c) modifications of the regulating role of attachment figures to include, for example, cradling to minimize arousal and exploration; (d) embracing cultural variations of sensitive caregiving, demoting the role of responsiveness to infant signaling; and (e) changing the tool kit for assessing attachment to include observation and discussion of culturally specific beliefs and ethnotheories, demoting the SSP. AT currently excludes caregiving practices and beliefs of the majority of the human species, and should not constitute the singular framework for human infants’ socio-emotional development, in theory or in practice. We advocate for valuing indigenous patterns of child care (e.g., S. Waters et al., Reference Waters, Richardson, Mills, Marris, Harris and Parker2024), starting within local cultures, and “orienting practitioners to [the] value of adapting indigenous practices within [the] context of Early Childhood Education Centres (Wadende & Simatende, Reference Wadende and Simatende2022)” (Serpell, Reference Serpell2023, slide 44).