Recently, Levinson (2006) hypothesized that the evolution of language was possible because humans possessed a set of skills that he calls the “interaction engine.” These skills involve both the cognitive (e.g., intentionality, prosocial motivation) and the communicative (e.g., sequence organization, turn-taking organization, communication through multiple sensory modalities) domains. This interaction engine hypothesis does not represent distinct brain modules but rather should be viewed metaphorically as describing distinct principles of human interaction (Levinson, 2006; Heesen and Frohlich, 2022). While language considerably transformed human interactions (e.g., enhanced intersubjectivity; Enfield and Sidnell, 2022), the interaction engine hypothesis suggests that this distinctive set of skills involved in social interactions not only preceded the evolution of language but made language evolution possible in the first place. The evolution of these skills might have been driven by the challenges presented in group-living (Dunbar, 1993). While the entire set of skills presented in the interaction engine hypothesis might be unique to humans, single parts may have different evolutionary origins and onsets (Levinson, 2006, 2016).

One skill is the ability to act and communicate intentionally (Levinson, 2006; Heesen and Frohlich, 2022; Roberts et al., 2022). Intentionality, defined as the voluntary and goal-directed use of actions (Dennett, 1983; Townsend et al., 2017), has been a major focus in comparative research of animal communicative interactions (Graham et al., 2020; Rodrigues and Fröhlich, 2021; Schel et al., 2022) intentionality can function by switching recipient’s understanding of behavior to understanding of intention when the cognitive distraction inhibits it (Roberts and Roberts, 2022). This emphasis on intentionality, specifically in nonhuman primate gesture research, stems from studies on prelinguistic human infants (Piaget, 1952; Bruner, 1972) and early comparative studies tackling intentional use of communicative signals outside human communication (Plooij, 1978; Leavens and Hopkins, 1998; Tomasello, 2008). Indeed, intentionality has often been considered a prerequisite for communication to happen (e.g., Hobaiter and Byrne, 2011; Fröhlich and Hobaiter, 2018; but see: Roberts et al., 2014; Fröhlich et al., 2019). Because intentionality is not directly measurable (Scott-Phillips, 2015; Townsend et al., 2017), studies rely on behavioral proxies and parameters used in studies on prelinguistic human infants (Piaget, 1952; Bates et al., 1979; Bruner, 1981). These parameters involve, for instance, the gaze of the signaler and/or the recipient, adjustment to potential audiences and goal persistence (Leavens et al., 2005; Call and Tomasello, 2007; Townsend et al., 2017; Ben Mocha and Pika, 2019; Ben Mocha and Burkart, 2021; Rodrigues and Fröhlich, 2021). Using this method, studies mainly focused on signals that were accompanied by a specific set of intentionality criteria (e.g., Hobaiter and Byrne, 2011; Graham et al., 2018; but see: Roberts et al., 2014; Fröhlich et al., 2019; Aychet et al., 2021). In addition, these intentionality criteria often differed between studies (Bourjade et al., 2020; Graham et al., 2020), and are not generalizable across modalities (Rodrigues and Fröhlich, 2021). For example, the intentionality criterion sensitivity to the attentional state of recipient can be measured for silent-visual gestures and facial expressions but not for vocalizations or tactile gestures (Rodrigues and Fröhlich, 2021). This leads to studies a priori selecting a few elements of the communicative interaction for their analysis based on proxies for intentionality. Additionally, there are other elements such as gaze and relative body position that make up the communicative interaction (Wilkinson et al., 2012; Rossano, 2013a,b). While intentionality is an important component of the interaction engine (Levinson, 2006; Heesen and Frohlich, 2022), it is not the only one. For other components of the interaction engine (e.g., sequence organization or turn-taking organization), one needs to investigate all elements which are part of the interaction independent of whether they may have been produced intentionally or not (e.g., Rossano, 2013b; Mondada and Meguerditchian, 2022).

One field that specifically focuses on how participants achieve coordinated social interaction, while at the same time remaining relatively agnostic about a participant’s intentionality with regard to a particular element of talk or non-verbal behavior in that interaction, is conversation analysis (CA). CA has been described as the dominant approach to the study of human social interaction (Sidnell and Stivers, 2013), and has also been successfully applied to nonhuman social interactions, highlighting some components of the interaction engine (e.g., Wilkinson et al., 2012; Rossano, 2013b; Genty et al., 2020; Mondada and Meguerditchian, 2022).

CA focuses on the vocal (e.g., talk, grunts) or embodied (e.g., gesture or body movement) social practices which a participant can draw upon to produce a meaningful interactional contribution. These practices are social in that they are resources which are available to, and understood by, the members of the culture or group (Atkinson and Heritage, 1984). As such, another participant in the interaction can use their knowledge of these practices in order to make sense of a participant’s contribution and respond to it appropriately (Heritage and Atkinson, 1984; Iedema, 2003). A goal of CA is to explicate these social practices and how participants make use of them to co-create a coherent interaction on a moment-by-moment, turn-by-turn basis. Knowledge and use of these practices have been shown to underlie various features of human interaction, such as: action formation, i.e., how a contribution is designed such that it can be understood to be performing a particular communicative action (such as a request or a greeting); turn-taking organization, including how participants are able to coordinate their contributions in a timely manner; and sequence organization, which includes how sequences of actions are structured across different participants, such as how an initiating action (such as a question) produced by one participant can provide an opportunity for, or make expectable, a responding action (such as an answer) from another participant (Schegloff, 2007).

A methodological consequence of this analytic stance is that rather than the researcher speculating on what a participant intended or whether that intention was really understood by a recipient, the focus of the analysis is on observing and describing the social practices used by the participants to produce meaningful contributions and to respond to others’ contributions (Atkinson and Heritage, 1984).

While early work in CA drew predominantly on audio-recordings of telephone calls and, as such, focused predominantly on talk, more recent work in CA and other areas of interaction research has also increasingly examined embodied aspects of social interaction (e.g., Heath and Luff, 2013; Mondada, 2016). Embodied conduct includes not only gestures, specific body movements and eye gaze behavior but also the whole bodies of participants as they engage with other participants, often in ways which involve close coordination and rapid co-adjustments as seen in activities such as sports, dancing or physical play (Goodwin, 2017; Meyer and von Wedelstaedt, 2017).

This work draws on concepts such as intercorporeality, i.e., a notion of embodied conduct that highlights how it is not possible to understand the ways in which an individual body acts or perceives within joint activities without taking into account its inter-relationship with the bodies of these others which are simultaneously acting and perceiving (Merleau-Ponty, 1962; Meyer et al., 2017). As Meyer et al. (2017, p. xviii) put it in relation to the human body, from an intercorporeality perspective, the human body is “constituted by its corporeal relations and interactions with other human or animal bodies” (emphasis in original).

In understanding how such joint activities can be successfully achieved, the notion of bodies as instruments or ‘signalling devices’ (Meyer et al., 2017) expressing in physical form the products of previously conceived intentions appears to have limited application. Rather, from an intercorporeal perspective, intentionality can be conceptualized as ‘interactionally emergent rather than causally prior to action’ (Meyer and von Wedelstaedt, 2017, p. 12; see also Heesen et al., 2017; Bangerter et al., 2022).

In summary, research on intercorporeality focuses on the inter-relations between bodies and their actions rather than on individual bodies and actions. While work drawing on the notion of intercorporeality would appear to have much to offer studies of animal communication and animal social behavior, the application so far has been limited (but see, e.g., Due, 2023, on intercorporeality in an interspecies interaction).

Examining intercorporeality can be one way of investigating how participants participate together in a social activity. The notion of participation has been heavily influenced by the work of Goffman, including his concept of a participation framework (Goffman, 1981). Goffman’s concept primarily concerned the relationship between human participants in an encounter which includes talk, such as a multiparty conversation, and focuses on the different types of roles that speakers and hearers can be inhabiting at a particular point in time within the interaction. More recently, work on participation and participation frameworks has broadened in various ways, including moving away from quite a static picture of participant roles to a focus on more dynamic features of how participants collaborate together in the co-construction of activities through shifting embodied participation frameworks (Goodwin and Goodwin, 2004; Goodwin, 2007) as well as being extended to nonhuman participants, such as chimpanzees engaged in social play (Heesen et al., 2017) or bonobos (Pan paniscus) in joint-travel initiations (Rossano, 2013b). It is this broader conception of participation and participation frameworks that we will draw on here.

One feature of intercorporeality concerns how bodies engaged in a collaborative activity (e.g., acrobatics: Brümmer and Alkemeyer, 2017) provide affordances for other bodies to produce certain actions. The notion of affordances comes from Gibson (1986) who discussed the affordances provided by both inanimate and animate objects. Our focus here will be primarily on the affordances provided by animate objects (people and animals), and specifically on how the affordances provided by chimpanzees’ bodies during contact social play provide opportunities for responsive action and thus facilitate the coordination of that play.

The nature and use of such bodily affordances have been explored in human interaction. For example, Katila (2018), in a study of intercorporeal formations between mothers and their (2 to 3 year old) toddlers, showed how different corporeal formations between mother and child provided different affordances. For example, a ‘nested tactile arrangement’ where, for instance, the child nestled in the mother’s lap, afforded continual tactile communication between mother and child, in that the mother could at any moment rock, stroke or tickle the child. A ‘distal tactile arrangement’, on the other hand, where the mother and child were physically close but not in constant physical contact, afforded both bodies greater independence, such as the mother being able to more easily interact with others.

A focus of our analysis will be how a bodily affordance provided by one chimpanzee during contact social play can provide an opportunity for responsive action by another chimpanzee, thus facilitating the fluent moment-by-moment co-construction of this play activity. This sequential organization of (1) what the bodily affordance offers and (2) how another chimpanzee responds to that affordance is somewhat analogous to sequence organization in talk-in-interaction (Schegloff, 2007) where a verbal action such as a question by one participant sets up a ‘next move slot’ for a responsive action (an answer) by another (see also Hutchby, 2001; Withagen et al., 2012, on different ways in which affordances can provide opportunities for action).

Bodily affordances differ from more “traditional” signals used in the comparative study of communication (King, 2004). Within the study of animal communication, signals are commonly divided into three categories: vocalizations, gestures and facial expressions (Slocombe et al., 2011; Liebal et al., 2013). Signals are seen as distinct units of information which are send from a signaler and affect the behavior of the receiver. They have been traditionally studied as part of the information-processing paradigm (Shannon, 1948; Seyfarth et al., 2010; Wheeler et al., 2011). They are classified based on their form, and different signals might lead to the same behavioral change or the other way around (e.g., Call and Tomasello, 2007; Hobaiter and Byrne, 2014). While some bodily movements, such as gestures, are studied in this paradigm, bodily affordances better fit a different paradigm, that of dynamic system theory (Shanker and King, 2003; King, 2004; King and Shanker, 2016). However, this does not mean that signals and bodily affordances cannot be studied simultaneously. Bodily affordances are part of any social conduct involving bodies. On top of this bottom layer analysis of interacting bodies, signals (intentionally produced or not) can be added for increased complexity. Here, we want to show how bodily affordances can add to the comparative study of communication in conjunction with “traditional” signals.

In this paper, we wish to draw on the concepts of intercorporeality and bodily affordances to explicate how chimpanzees can collaboratively and fluently produce contact social play activities together. Contact social play activities, such as play biting, can be contrasted with non-contact social play activities, such as chasing (Caro, 1995). Joint activities such as these can present a puzzle in terms of how animals can co-produce them with such apparent ease and fluency, given that they involve improvisation, rapid coordination and fast adjustments (Pellis and Pellis, 1996; Palagi, 2006). Play contexts have long been an object of interest for comparative psychologists for various reasons. For instance, in regard to human infants, playing is crucial for the development of metacognitive skills such as planning, reflection, and self-regulatory behaviors (Whitebread et al., 2009), and is linked to language acquisition (Tomasello and Rakoczy, 2003; Orr, 2021). Concerning other animals, playing is common across several taxa (Bekoff and Byers, 1998). Additionally, this behavioral domain has also been suggested to be a useful platform to study both the evolution of shared intentionality (Heesen et al., 2017) and joint commitment as a process (Heesen et al., 2021a; Bangerter et al., 2022; Rossano et al., 2022). As Heesen et al. (2017, p. 1) note about social play: ‘One aspect that has recurrently fascinated scholars is its complex, cooperative nature that requires substantial on-the-fly coordination and improvisation (Bekoff and Allen, 1998; Bekoff, 2001; Palagi, 2006), in comparison to other social activities like grooming or sex that involve more stereotyped activity-specific patterns’. Being a dynamic, as well as a very common, form of interaction, social play forms an ideal context to study the different components of the interaction engine (Heesen et al., 2017).

However, due to its dynamic nature playing is difficult to analyze using a “traditional” approach, as possible underlying intentionality is hard to assess. Additionally, playing often comprises many bodily movements that do not readily fall under “traditional” signal categories. Therefore, while playing is of interest for many reasons, it has received relatively little research attention in studies of animal communication (but see: Pika et al., 2003, 2005; Palagi, 2008; Heesen et al., 2017). Existing studies into playing interactions have focused on specific signals (Pika et al., 2003, 2005; Palagi, 2008; Genty and Byrne, 2010; Mancini et al., 2013), specific phases (Flack et al., 2004; Fröhlich et al., 2016a), or described the whole playing interaction (Palagi, 2006; Heesen et al., 2017, 2021a,b). Most studies have investigated species living in captivity (but see: Fröhlich et al., 2016a).

In our analysis, we specifically focus on data of spontaneous interactions of Eastern chimpanzees (Pan troglodytes schweinfurthii) living in a habituated community in their natural environment, the Kibale National Park in Uganda. The communicative repertoire and usage of chimpanzees have been extensively studied (e.g., Goodall, 1986; Hobaiter and Byrne, 2011; Tomasello and Call, 2018; Crockford, 2019). However, to better understand quick, dynamic joint activities such as contact social play, traditional approaches focusing on communicative signals are not expedient. For such activities, we suggest to adopt and expand analyses drawing on concepts and methodologies from work into intercorporeality, such as that of bodily affordances. Such an approach can provide insights into how chimpanzees and possibly other animal species co-produce these rapidly changing joint activities through particular forms of communication and co-ordinated action.

A total of 22 mother-infant dyads (9 females, 13 males) were followed on 155 days between 7:00 AM and 17:30 PM. This method resulted in a total of 139 h of video recordings with 2,723 interactions of which 825 interactions involving playing (i.e., two or more individuals engaging in playing without a break of more than 30 s; Koski et al., 2007; Newton-Fisher and Lee, 2011; Roberts et al., 2012). The recordings were made using a digital camera (Sony AX100E 4 K) and a directional microphone (Sony shotgun ECM-CG60) to capture all visual and audible acts produced in the interactions. Data was collected following the sampling rule of focal-animal sampling focusing on the infant with a recording rule of continuous recording when the visibility allowed for it (Altmann, 1974; Martin and Bateson, 2007). The ages of the infants ranged from 3 to 46 months (median: 21 months).

For the purpose of the present paper, we selected four different videos from the bigger dataset based on the following rationale: They (a) were representative of a playing session in the dataset; (b) were characterized by a high level of visibility, meaning both participants, their communicative signals and movements could be clearly seen; (c) showed instances of bodily affordances; and (d) involved different physical arrangements (e.g., nested arrangement, one individual hanging from a branch in front of/above the other individual) and, in particular, play types (e.g., play biting, tickling, falling on the other individual). Each video contained one playing session. From these playing sessions, we extracted one or multiple episodes to highlight different ways interactants could, initiate, end, or maintain a joint contact play. More detailed information concerning the individuals involved in the selected videos, as well as which extracted episodes relates to which video, can be found in the Supplementary Table S1. A video, of each extracted episode can be found in the Supplementary Videos S1–S8.

The episodes described below are from videos that all involve social play interactions, specifically contact social play between an infant and an adult individual. Contact social play was defined as at least two individuals engaging in a social interaction containing playful behaviors which involve chimpanzees’ bodies touching each other in some manner (e.g., play biting, tickling) and signals associated with play (e.g., play face, slap other; Flack et al., 2004; Fröhlich et al., 2016a). For playing to commence, both participants needed to actively engage in the playing interaction. Every example contains one playing session. However, similar to grooming (Goodall, 1986), playing is often not a continuous activity and one playing session can contain multiple playing bouts interspersed with breaks. A play bout ended when one of the participants ceased its play behavior. We defined the end of a play session when play behavior stopped for more than two minutes (Heesen et al., 2021a,b). However, it should be noted that using a time criteria to distinguish between the end of a playing bout (interruption) and a playing session (true ending) is, ultimately, arbitrary.

To refer to distinct signals used, we adopted terms from the existing literature on chimpanzee communication (Nishida et al., 1999; Parr et al., 2005; Call and Tomasello, 2007; Hobaiter and Byrne, 2011) and denoted them in capitals (Table 1).

The analysis of contact social play activities in chimpanzees focused on the behavioral means through which adult and infant chimpanzees together negotiated and achieved the ‘on-the-fly coordination’ of play mentioned by Heesen et al. (2017). Specifically, we described these practices (Meyer and von Wedelstaedt, 2017; Mondada and Meguerditchian, 2022) in terms of signals and, in particular, of the inter-relationship between the two chimpanzee’s bodies (including within their material context, such as the tree sapling a chimpanzee is sitting on or hanging from) through which chimpanzees communicate with each other and collaborate together to initiate, end and maintain episodes of contact social play. We do not make a strong distinction between practical corporeal activity (such as an infant climbing into its mother’s lap) on the one hand, and communication on the other. As Katila (2018, p. 5) notes (in relation to bodies in contact with each other) “while producing practical actions in tactile intercorporeality, bodies are inevitably communicating their moment-by-moment unfolding actions and action intentions to each other.”

The embodied conduct and postural alignment of a chimpanzee can provide an affordance (Gibson, 1986) to another chimpanzee. This means that the body can be presented to another individual in such a way that it affords the other individual an opportunity to do something in relation to that body with their own body. That is, these corporeal behaviors and postures may not constitute solely individualistic conduct, but rather can have a social and communicative aspect, i.e., making the body available, or proffering it (with whatever level of conscious awareness) to the other in such a manner that it can be used by that other in a certain way. The analysis focused on the nature of these bodily affordances and the type of sequential organization they make possible through providing a ‘next move slot’. Here, another chimpanzee has the opportunity to align with the move constituted by the bodily affordance and take forward the activity which is underway.

We primarily, but not exclusively, drew on examples from three different types of adult-infant play, i.e., contact social play sessions where the play primarily consisted of:

A focus of the analysis is the interactional and sequential process by which the contact social play activities are co-created and negotiated on a moment-by-moment and move-by-move basis between the chimpanzees (see also Genty et al., 2020). This process has a number of stages which the chimpanzees must proceed through together for a bout of play to be initiated and maintained. At the same time, it highlights the contingency of the activity in that despite one chimpanzee engaging in play behavior or play communication, there is a reliance on the other chimpanzee to co-engage in order for contact social play to start or be maintained.

The stages of the process we observed and analyze below are as follows: In each case, we signpost relevant examples (presented as ‘Episode 1’ etc.) that will be discussed in the Results section.

We noted above that contact social play was initiated when an initiating move into play by one chimpanzee was responded to with an aligning move into play by another chimpanzee such that both were then simultaneously involved in that play. The end of a play bout (and thus potentially a play session if another bout is not initiated) involves the cessation of simultaneous joint play, most commonly by one chimpanzee disengaging from the play. This disengaging can take the form of still physically engaging with the other chimpanzee but now in the form of a non-play activity, or of stopping the play-related physical arrangement with the other chimpanzee without embarking on a joint non-play activity. We will examine an example of each of these in turn.

How contact social play is maintained can be examined in terms of (1) how a single bout of play is maintained after its initiation until its end, and also (2) how a session of play is maintained by moving from one bout to another. We examine both here. In the first case (Section 3.3.1), a question can be how the chimpanzees communicate/display to each other that the bout of play is ongoing and continuing. In the second case (Section 3.3.2), a question can be how one or both chimpanzees may communicate that despite a play bout having ended, one or both of them are available for another play bout, thus possibly extending the play session.

The examples used in this paper have shown that the concepts of intercorporeality and bodily affordance are useful for the analysis of social play interactions between infant and adult chimpanzees. However, these concepts can easily be used more broadly across interactions in general. While the notion of bodily affordances has clear applicability in capturing features of a dynamic, physical activity such as contact social play, it is by no means limited to that type of activity. A recent article of Mondada and Meguerditchian (2022) used a comparable approach to study greetings in Olive baboons (Papio anubis) showing how some formats of action create different affordances warranting different responses. Another joint action that has been suggested as a context of interest is grooming (Genty et al., 2020; Heesen et al., 2021a). A commonly used signal in grooming is the “present” gesture defined as an individual offering a body part such as the arm, armpit, back, genital region, leg, or rump to the recipient for subsequent grooming (Pika, 2014). At the same time, “present” is a perfect example of creating the affordance for the next step in the interaction: the grooming of that specific body part (Hobaiter and Byrne, 2011; Roberts et al., 2014). Numerous examples can be given in other situations as well (e.g., a mother performing the “lower back” gesture to afford for her infant to climb on in order to leave together; Fröhlich et al., 2016b). These examples also highlight a way in which bodily affordances can explain the use of specific gestures in certain contexts. While not every gesture might be explainable in terms of bodily affordances (e.g., “leaf clipping”: Nishida, 1980; Boesch, 1995; Nakamura et al., 2000; Pika and Deschner, 2019), the framework does add another dimension in which the production, usage and meaning of gestures can be studied and possibly better understood. Gestures could then be measured based on how much their meaning – derived from the ASO method (Hobaiter and Byrne, 2014) – overlaps with the affordance the gesture creates. As can be seen from Episode 4, some movements can create multiple conflicting affordances and in these cases, intentionally produced signals could help disambiguate the situation (Leavens et al., 2005; Roberts et al., 2013; Roberts and Roberts, 2022). Future studies may investigate if the production of intentionally produced signals is dependent on the existence of simultaneous produced conflicting bodily affordances.

The present paper has shown that while bodily affordances and gestures are based on different communication paradigms (dynamic system theory vs. information-processing), both can be studied simultaneously and that they reinforce each other, giving a more complete description of the whole interaction. As can be seen in the episodes, participants made use of both signals and bodily affordances to achieve coordinated joint activity. This has shown that the concepts of intercorporeality and bodily affordance add to the study of animal communication but do not replace more “traditional” measurements even though there is overlap between them, where, for example, gestures themselves create bodily affordances. Another direction in which the present framework may be taken is the study of the role played by bodily affordances in the acquisition of gestures (Hutchins and Johnson, 2009). Gestures may be shaped by the affordances that they represent. This would fit the social negotiation hypothesis which postulates that gestures emerge from a repeated exchange of social behaviors between interactants (Pika and Fröhlich, 2019). Interactants might initially make use of bodily affordances but over repeated exchanges a negotiation of certain behaviors might lead to the co-creation of gestures. If that is the case, one could expect gestures where the meaning is close to the affordance they create to appear earlier in development. Additionally, this convergence of signals negotiated by both participants could be augmented by intentionality markers on top of the basis of bodily affordances (Leavens and Hopkins, 1998; Pika and Bugnyar, 2011; Roberts and Roberts, 2019; Schel et al., 2022). However, these lines of enquiry are outside the scope of the current paper and will not be pursued further here.

In our results, we showed the role of bodily affordances in the opening, main body and exit phase of social play following the joint action framework of Heesen et al. (2017). As these phases also appear in playing interactions in humans (Clark, 1996; Clark, 2006; Rossano et al., 2022) future research could investigate the role of bodily affordances comparatively.

In addition, the notion of intercorporeality and bodily affordance is species-agnostic. Unlike more traditional approaches, bodily affordances can be studied independent from the cognitive mechanisms that underlie them. This allows for bodily affordances to be studied in any species in the animal kingdom that shows social behavior involving two bodies interacting with each other. Species that are similar in build and anatomy will create similar bodily affordances. Taking this approach will thus greatly increase possibilities for cross-species comparative research.

With extracts from only four different play interactions, this paper works with a limited dataset. Therefore, any patterns shown here might not represent general patterns in playing interactions in chimpanzees in the wild. For example, a common play signal laughter, did not occur in any of our examples (Van Hooff, 1972; Matsusaka, 2004; Ross et al., 2010). Additionally, confounding factors, such as social bond or rank and age difference that might influence the playing interaction were not analyzed (Heesen et al., 2020, 2021a,b). However, the aim of this paper was to introduce the concepts of intercorporeality and bodily affordance to the field of comparative communication research and demonstrate how these could be used to analyze dynamic chimpanzee social interactions. Therefore, this paper should be viewed as a “proof of concept” and not as evidence for general patterns of bodily affordance employed by all chimpanzees.

Following previous literature, we defined the end of a playing session as the cessation of play behavior for more than two minutes (Heesen et al., 2021a,b). However, ultimately, using time criteria to classify between an interruption and a true ending is arbitrary. Stronger evidence of whether the participants experienced the joint activity as being ended could be found in the follow-up behavior of the participants. An interruption might be characterized by at least one participant attempting to actively reinitiate a new playing bout or passively keeping up the affordance for play (e.g., Episode 8), whereas a true ending might be characterized by both participants breaking the participation framework and “moving on” to another (joint) activity (Episode 5). In three out of four of the extracts analyzed in this paper, the joint activity of social contact play was ended by one or both participants engaging in a joint-travel interaction (e.g., Episode 5), thus physically removing themselves from the place of play and severely diminishing the optionality of the previously established type of play restarting. In the fourth extract (Episode 4) the option for re-engagement of play remained as the physical arrangement did not significantly change after the episode. However, there was no successful initiation of social contact play to start from and, additionally, following this unsuccessful initiation, both participants did not engage with each other for at least three minutes after the analyzed episode.

While the concepts of intercorporeality and bodily affordance can be used very flexibly, they also have their limitations. As the names suggest, intercorporeality and bodily affordance deal with physical movements and physical arrangements of bodies. Therefore, they will not add a new dimension to the study of purely vocal interactions such as long-distance vocal exchanges (Arcadi, 2000; Geissmann, 2002; Schamberg et al., 2016; Southern et al., in preperation). However, most communicative interactions are inherently multimodal (but see for behavioral ecologists: Partan and Marler, 1999; Hebets and Papaj, 2005), with multimodal referring to the simultaneous or sequential integration of signals from at least two ‘modalities’ (Liebal et al., 2013; Luef and Pika, 2017). As can be seen in the present paper, bodily affordances can be analyzed in conjunction with other signals of different modalities.

Levinson’s interaction engine hypothesis states that certain characteristics of social communicative interactions may have preceded the evolution of language and the different components of this “interaction engine” might have different evolutionary origins (Levinson, 2006). A crucial window onto these evolutionary origins and different antiquities is the comparative approach, enabling a comparison of communicative toolkits across closely-related species and beyond (e.g., Pika, 2008; Pika and Bugnyar, 2011). Comparative studies into communicative interactions may then focus on all components involved and not just on a subset of elements based on implied cognitive mechanisms such as intentionality. By a priori selecting specific parts based on intentionality criteria only a part of the communicative interaction is analyzed. Communicative interactions can contain many semiotic resources that are not captured by “traditional” measurements such as intentional signals and gaze. By drawing on the concepts of intercorporeality and bodily affordances, we showed that the common dynamic interaction of social play can be better understood through focusing on aspects of the activity which are typically glossed over, thus offering a new more complete window onto how joint activities in the animal kingdom are coordinated. We have aimed to demonstrate how chimpanzees coordinate contact social play through a form of sequential organization, using the intercorporeal relations between their bodies. This includes making use of the affordances these bodies provide, to coordinate activity on a move-by-move basis, where one move (intentionally produced or not) can set up a sequential ‘slot’ for the next move potentially to occur, and move the activity forward. This echoes findings about how human participants shift postures together, apparently almost simultaneously, where “close inspection of these moments usually shows that they involve sequential interaction in which one party proposes or preenacts a shift and is then, in a second step, joined by the other when he completes the change in posture” (Meyer et al., 2017, p. xxiii).

The sequential organization of bodily action is one aspect of the social organization of interaction which is evident in both human and nonhuman animals. It provides an apparent continuity between them in relation to how social activities are cooperatively co-constructed, with one participant monitoring the other, and producing conduct which can be seen by those present as responding to, and potentially aligning with, the other’s prior conduct. As Streeck (2017, p. 352), noted, to understand the evolution of human language and talk-in-interaction, it would appear to be important “to search for the primordial intercorporeal capacities from which the distinctly human form of sociality, language use in interaction, has emerged.” The concepts of intercorporeality and bodily affordance may therefore allow the start of broader comparisons involving carefully chosen representatives of primates and beyond. It allows us to study every type of interaction that involves bodies. From this bottom layer of interacting bodies, we can stack evidence of more “intentional” signals and gaze, to make comparisons between human interactions and those of other animals in terms of complexity. In this paper, we have shown that animals can communicate by doing instead of just by signaling and moved away from a too narrow focus on intentionality only. Therefore, the approach that we have highlighted here might bring us one step closer to finally solving the puzzle that is the evolution of language and may in addition open up other pathways in the field of comparative psychology.