Communicative complexity and development in chimpanzees (Pan troglodytes) and bonobos (Pan paniscus) in the wild

(1)

Communicative complexity and development in chimpanzees (Pan troglodytes) and bonobos

(Pan paniscus) in the wild

Dissertation submitted for the degree of Doctor of Natural Sciences

Presented by Marlen Fröhlich

at the

Faculty of Sciences

Department of Biology

Date of the oral examination: 7 July 2016 First supervisor: Prof. Dr. Michaela Hau

Second supervisor: Dr. Simone Pika

(2)

(3)

„Die Natur muß gefühlt werden, wer sie nur sieht und abstrahiert, kann … Pflanzen und Tiere zergliedern, er wird die Natur zu beschreiben wissen, ihr aber selbst ewig fremd sein.”

Alexander von Humboldt

(4)

(5)

Summary

In my doctoral thesis I examined the diversity and ontogeny of gestural communication in chimpanzees (Pan troglodytes) and bonobos (Pan paniscus) living in their natural environments. Great apes use gestures as intentional and flexible communicative strategies in a wide range of social contexts, which has led many researchers to emphasize the role of gestures in the evolution of language. However, little is known about the ontogeny of these communicative means, especially in wild ape communities that are not influenced by a human-modified environment. The developmental approach yet comprises an indispensable tool to gain insights into the cognitive complexity underlying gestural communication. Since we can only understand the communicative abilities of a species if within-species variability is considered, I conducted the first study on ape gestural communication in multiple social groups using a consistent methodology across study sites and subsequent years. I observed chimpanzees living in two communities of different subspecies (P. t. schweinfurthii at Kanyawara, Kibale National Park, Uganda, and P. t. verus at Taï South, Taï National Park, Côte d’Ivoire).

Additionally, I included corresponding data on mother-infant dyads in two bonobo communities: LuiKotale at the fringe of Salonga National Park, and Wamba in the Luo Scientific Reserve, Democratic Republic of the Congo. I thus had the exceptional opportunity to compare social groups of different communities and sub-species, and for mother-infant interactions also different ape species.

To better understand the communicative complexity of these species, I combined the comparative with the developmental approach, which included cross-sectional comparisons between individuals of different groups, as well as longitudinal comparisons of the same individual between subsequent years. To assess general developmental transitions of the infants and to ensure comparability across species, this thesis was built upon methods developed by researchers investigating communicative development of human and non-human primates. Moreover, I adopted a conversation-analytic, multimodal framework derived from linguistics to gain a deeper understanding of chimpanzee communication. I specifically focused on three communicative contexts—

mother-infant joint travel, play solicitation and food sharing—to assure that the examined signals actually carried the ‘meaning’ perceived by the observer (i.e. ‘goal-outcome matching’). Thus, my thesis differed substantially from the majority of gestural research

(8)

on great apes that included entire gestural repertoires of ape species across all age groups.

The major results of my thesis are summarized below:

First, based on my findings on the development and inter-individual variability of gesture production, my co-authors and I constructed a revised theory of gestural acquisition termed Social Negotiation. According to this theory gestures represent the output of social interactions, shared understanding and mutual construction in real time by both interactants. This study thus provided further support that apes acquire their gestures via learning mechanisms, as opposed to genetic predisposition (Chapter II).

Second, by adopting a conversation-analytic framework for mother-infant interactions in chimpanzees and bonobos in Chapter III I demonstrated that both species are capable of engaging in cooperative, sequential turn-taking interactions to achieve a joint goal.

Moreover, I found profound differences related to structural and temporal parameters of communicative interactions between the two species. In Chapter IV I demonstrated that chimpanzees possess the cognitive flexibility to adjust their gestural signalling to attributes of the recipient, and that age and relationships of the interactants strongly affected gestural usage. Lastly, I examined and compared gestural performance of infant chimpanzees in three social contexts and showed that communicative development crucially relies on interactive experience with social partners outside the mother-infant dyad (Chapter V).

Overall, my thesis provides hitherto undocumented evidence to which extent communicative abilities in chimpanzees and bonobos are linked to input from their complex social environment. With my work I hope to have stimulated future studies to also adopt a more holistic view of ape communication in natural environments. Further in-depth studies of ape gestural production are necessary to test the Social Negotiation Hypothesis in detail and to examine the usage of gestural signals in relation to developmental phase, social context and interaction partner. Hence, future comparative research should implement a developmental approach and incorporate different signal modalities and social settings. This enables us to test how the ecology and social structure of a species shapes its communicative complexity. In turn, this might shed further light on the extent to which communicative and cognitive abilities of our primate ancestors informed the evolution of the unique human communication system that we termed

‘language’.

(9)

Zusammenfassung

German summary

In meiner Doktorarbeit untersuchte ich die Diversität und Ontogenese gestischer Kommunikation in freilebenden Schimpansen (Pan troglodytes) und Bonobos (Pan paniscus). Menschenaffen verwenden Gesten als zielgerichtete und flexible Kommunikationsstrategien in verschiedensten sozialen Kontexten, weswegen ihre Bedeutung für die Sprachevolution wiederholt hervorgehoben wurde. Insbesondere in freilebenden, vom Menschen unbeeinflussten Menschenaffen ist bisher jedoch sehr wenig über die Ontogenese von Gesten bekannt. Dabei ermöglicht der entwicklungsorientierte Ansatz Einblicke in die kognitive Komplexität, die gestischer Kommunikation zugrunde liegt. Die kommunikativen Fähigkeiten einer Art lassen sich allerdings nur dann vollständig erfassen, wenn auch die intraspezifische Variabilität berücksichtigt wird.

Daher nutzte ich standardisierte Methoden um die gestische Kommunikation in mehreren sozialen Gruppen in verschiedenen Forschungsgebieten und Studienjahren zu klassifizieren. Ich untersuchte je zwei Schimpansen-Gruppen aus verschiedenen Unterarten (P. t. schweinfurthii in Kanyawara im Kibale-Nationalpark, Uganda, und P. t.

verus in Taï South im Taï-Nationalpark, Elfenbeinküste). Zusätzlich analysierte ich Daten von Mutter-Kind-Paaren aus zwei Bonobo-Gruppen: LuiKotale nahe des Salonga- Nationalparks und Wamba im Luo Scientific Reserve, Demokratische Republik Kongo.

Ich konnte also die sozialen Gemeinschaften verschiedener Unterarten und —hinsichtlich Mutter-Kind-Interaktionen— Arten von Menschenaffen vergleichen.

Ich kombinierte den vergleichenden mit dem ontogenetischen Ansatz, um die kommunikative Komplexität dieser Arten besser verstehen zu können. Um den generellen Entwicklungsverlauf der Jungtiere zu erfassen und die Vergleichbarkeit zwischen den Arten zu gewährleisten, ist diese Dissertation auf Methoden der vergleichenden Kommunikationsforschung aufgebaut, die für Studien zur kommunikativen Entwicklung von menschlichen und nicht-menschlichen Primaten erarbeitet wurden. Außerdem setzte ich konversationsanalytische, multimodale Methoden ein, die aus der linguistischen Forschung abgeleitet sind, um ein tieferes Verständnis über die Kommunikation von Schimpansen zu gewinnen. Meine Arbeit zielte speziell auf nur drei kommunikative Funktionen —die Initiierung des gemeinsamen Ortswechsels bei Mutter-Kind-Paaren, des sozialen Spiels und des Teilens von Nahrung— um sicher zu stellen, dass die analysierten

(10)

Signale tatsächlich jene „Bedeutung“ trugen, die vom Beobachter wahrgenommen wurde („goal-outcome matching"). Insofern unterscheidet sich meine Dissertation wesentlich von früheren Arbeiten zur gestischen Forschung an Menschenaffen, welche größtenteils das gesamte gestische Repertoire über alle Altersklassen hinweg erfassten. Die wichtigsten Ergebnisse meiner Dissertation fasse ich hier zusammen:

Zum Einen entwickelte ich gemeinsam mit meinen Koautoren, basierend auf Ergebnissen zur Ontogenese und inter-individuellen Variabilität gestischer Kommunikation, eine revidierte Theorie zur gestischen Entwicklung in Menschenaffen.

Dieser Social Negotiation Hypothesis zufolge resultieren gestische Signale aus Sozialinteraktionen, gegenseitigem Verständnis und werden von beiden Interaktionspartnern in Echtzeit konstruiert. Die Studie unterstützt die Ansicht, dass Menschenaffen ihre Gesten hauptsächlich erlernen (Kapitel II). In einer zweiten Studie nutzte ich einen konversationsanalytischen Ansatz, um die Mutter-Kinder-Interaktionen von Schimpansen und Bonobos systematisch zu vergleichen. Ich konnte zeigen, dass beide Arten zu kooperativer Kommunikation und interaktiven Rollenwechseln fähig sind, um ein gemeinsames Ziel zu erreichen. Zudem habe ich ausgeprägte Unterschiede hinsichtlich struktureller und zeitlicher Parameter zwischen den Interaktionen beider Arten festgestellt (Kapitel III). Weiterhin konnte ich zeigen, dass Schimpansen über die kognitive Flexibilität verfügen, ihre gestischen Signale auf Attribute des jeweiligen Rezipienten abzustimmen, und dass das Alter und die Beziehung der Interaktionspartner zueinander die gestische Kommunikation beeinflussen (Kapitel IV). Schließlich habe ich die Entwicklung der gestischen Kommunikation in den drei kommunikativen Kontexten verglichen und fand, dass die kommunikative Entwicklung von Schimpansenkindern entscheidend auf interaktiver Erfahrung mit Sozialpartnern außerhalb der Mutter-Kind- Beziehung beruht (Kapitel V). Zusammengefasst liefert meine Dissertation neue Erkenntnisse darüber, inwieweit kommunikative Fähigkeiten von Schimpansen und Bonobos von ihrem komplexen Sozialgefüge beeinflusst werden. Insofern hoffe ich, dass meine Arbeit künftige Studien dazu anregt, die kommunikativen Fähigkeiten von Menschenaffen in ihrem natürlichen Lebensraum mit Hilfe eines vergleichenden, multimodalen Ansatzes zu untersuchen. Weitere umfangreiche Studien zum Gesteneinsatz sind notwendig, um die Social Negotiation Hypothesis im Detail zu prüfen und den Einfluss von Entwicklungsphase, sozialem Kontext und Interaktionspartner zu untersuchen. Zukünftige vergleichende Studien sollten daher einen ontogenetischen Ansatz, Signalmodalitäten und die soziale Umgebung berücksichtigen. Diese können

(11)

Rückschlüsse darauf ermöglichen, wie die Ökologie und das Sozialgefüge einer Art deren kommunikative Komplexität beeinflusst. So könnten wir schließlich wichtige Erkenntnisse gewinnen, in welchem Ausmaß die kommunikativen und kognitiven Fähigkeiten unserer menschenaffen-ähnlichen Vorfahren als Voraussetzung für menschliche Sprache dienten.

(12)

(13)

Chapter I

General introduction

(14)

1.1 Evolutionary perspectives on the origins of language

The evolutionary origin of human language remains an intriguing issue for a large and interdisciplinary community of scholars (Christiansen & Kirby, 2003). Central to the debate is the extent to which socio-cognitive prerequisites of human language, such as intentionality, the ability to learn new signals and create new meanings, and cultural transmission were built upon existing cognitive competencies in our ape-like ancestors (Armstrong et al., 1994; Enard et al., 2002). Understanding language evolution has been hampered by the fact that the brain and the vocal production apparatus, crucial elements for speech production, do not fossilize (Ghazanfar & Rendall, 2008). Moreover, language leaves no traces in the archaeological record, unlike other human behaviours, such as art and tool making. For many decades most scientists agreed with the first views of linguist Noam Chomsky of language building on specific cognitive skills that are unique to humans, with little or no ape-like precursors (Chomsky, 1959, 1966). Since animal communication lacks grammatical items and the infinite productivity of human language, animal communication has been dissociated from human language (Bickerton, 1992). By now many scientists, including Chomsky and colleagues (2002), no longer hold strictly to the human-uniqueness claim. They have urged research on the aspects of human language that are unique to us or shared with extant nonhuman species. Indeed, the comparative approach might present the most powerful tool we have to draw conclusions about the communicative abilities of our extinct ancestors (Hauser et al., 2002). In the following I will provide a brief overview of comparative research related to language evolution to date, focusing on the three modalities of vocalisation, facial expression and gesture.

To gain insights into the evolutionary roots of language, researchers first looked at primate vocalisations (Marler, 1980; Seyfarth, 1987; Snowdon et al., 1982). This initial focus seemed to be justified by the initial findings on the informational content of referential alarm calls in monkeys (Seyfarth et al., 1980). A vocalisation is classified as

‘functionally referential’ if it (1) is correlated with the occurrence of objects or events in the external world of the signaler, and (2) induces the receiver to respond adaptively in the absence of direct cues from the eliciting stimulus (Evans, 1997; Marler et al., 1992).

However, until now there is little evidence that great apes have similar referential vocalisations (however, see Slocombe & Zuberbühler, 2005; Watson et al., 2015). In addition, Wheeler and Fischer (2012) recently suggested that the perception of

(15)

functionally referential signals is less cognitively demanding as previously thought (Slocombe & Zuberbühler, 2005; Zuberbühler, 2003), since they are closely tied to particular contexts, making the integration of contextual cues less critical for the generation of meaning. Therefore, they argue neither the production nor the perception of functionally referential signals is more relevant to the origin of language than any other animal signal (Wheeler & Fischer, 2012). In addition, studies showed that, unlike speech, the production of vocalisations of nonhuman primates seems to be genetically fixed, mostly unlearned, tightly linked to affective states, and broadcasted indiscriminately in the immediate vicinity (Tomasello, 2008; Tomasello & Call, 1997). Taken together, the majority of research on nonhuman primate vocalisations revealed that call morphology and call usage seem to have only limited flexibility (Corballis, 2002), which renders it unlikely that human language evolved from a system based upon these types of communicative means alone (Arbib et al., 2008).

Similar to primate vocalisations, facial expressions have been commonly assumed to be expressions of emotional states (Ekman, 1992) that are largely innate and involuntarily produced (Tomasello, 2008). Therefore, this modality has only been discussed in relation to language origins when considered as part of a gesture or vocalisation (Liebal et al., 2013). However, recent studies have highlighted ‘lip- smacking’ facial expressions for the understanding of language precursors, proposing that rhythmic facial movements required for speech production evolved from facial movement patterns in a primate ancestor (Bergman, 2013; Ghazanfar et al., 2012). This theory is based on evidence that the temporal pattern and developmental trajectory of ‘lip- smacking’ (starting with a variable ‘babbling’ phase) in rhesus macaques resembles the timing and ontogeny of speech (Ghazanfar et al., 2010; Morrill et al., 2012).

In the past couple of decades, advances in behavioural and neural research on nonhuman primates has led many scientists to emphasize the role gestures might have played in the evolution of language (e.g., Arbib, 2005; Corballis, 2002; Hewes, 1973;

Vauclair, 2004). To date there is strong evidence that great apes have considerably more intentional control, flexibility, and interactional awareness of their productive gestural communication than of their vocalisations or facial expressions (Call & Tomasello, 2007;

Pollick & De Waal, 2007; Tomasello, 2008, see also 1.3.). However, since both apes and humans do not gesture or vocalise in isolation, scientists have proposed to study different component signals in concert (Levinson & Holler, 2014; Slocombe et al., 2011).

(16)

Adopting an integrated multimodal approach might be crucial to avoid methodological discontinuities between the single modalities and to better understand the precursors to human language (Liebal et al., 2013). For example, in an attempt to reconcile vocal and gestural models of language evolution, Lameira and colleagues (2012) recently proposed that ‘gesture-calls’ (i.e. kiss-squeaks) of great apes may have constituted a speech exaptation (a change in function of a pre-existing characteristic during evolution),

‘providing functional advantages in a human ancestor […] allowing the emergence of the neural and communicative basis for subsequent selection favouring basic abilities for speech’ (Lameira et al., 2012; p. 417).

1.2 Gestures in human communication

Human communication is characterised by a tight integration between gesture and speech (Goldin-Meadow & McNeill, 1999; McNeill, 2000). Children use gestures for communication before their first spoken words (Iverson & Goldin-Meadow, 1998), and adult speakers normally accompany all their speech with expressive manual gestures (co- speech gestures; McNeill, 1992). The emergence and development of communicative abilities in from pre-linguistic gesture has revealed many insights important to the understanding of human communicative abilities (Bruner, 1981b). Intentionality and referentiality, both crucial features of language, develop in children around the age of nine to twelve months (Bates et al., 1975). For instance, the ARM RAISE gesture used by human children to be picked up by the caretaker has been described as one of the first intentional signals (Caselli et al., 2012; Clements & Chawarska, 2010; Lock, 1978).

These transitions have been suggested to have far-reaching implications for the children’s cognitive and social development and have been associated with abilities such as understanding of others as social agents and the onset of true symbolic communication (Bates et al., 1975). Furthermore, studies on word-gesture combinations of children in the one-word stage showed that gestures and vocalisations are interlinked in children from an early age (Bretherton & Bates, 1979; Wetherby et al., 1988) and that they thus master more complex combinatorial meanings than evident in their one-word utterances alone (Greenfield & Smith, 1976).

(17)

Moreover, deaf cultures develop full-fledged sign languages which function without any use of speech at all (e.g., Klima & Bellugi, 1979). Goldin-Meadow and colleagues showed that deaf children raised without sign language will create and use idiosyncratic signs containing reference and displacement, which are considered essential properties of human language (Goldin-Meadow, 2005; Goldin-Meadow et al., 2005; Morford &

Goldin-Meadow, 1997). Thus, for both deaf and hearing individuals, structures of human language express themselves visually depending on circumstances.

Finally, spontaneous gestures employed during speech (‘co-speech gestures’) have been shown to provide both usable information to the recipient and insights into the thought process of the speaker (Goldin-Meadow et al., 1993; McNeill, 1992). It is still debated to which extent these gestures function as communicative aids for the recipient (e.g., Kendon, 2004), cognitive aids for the signaller (e.g., Krauss et al., 1995), or both (e.g., Bavelas, 1994; Özyürek, 2002). While spoken language breaks up meaning into words and uses combinations thereof to encode meaning, gestures allow the transmission of holistic ideas, often transmitting information about the quality or nature of an event not encoded in the accompanying speech (McNeill, 2000). The mutual interaction of gesture and speech suggests that human language, though overwhelmingly verbal, is usually a multimodal communication system.

1.3 Gestural research in great apes

In a pioneering comparative study, Plooij (1978) demonstrated that wild chimpanzee infants go through a sequence of communicative development similar to that of human infants, with a gradual shift to intentional communication around the age of eight to twelve and a half months. In this developmental shift from non-intentional to intentional acts, the chimpanzee infants gradually begin to direct gestures as intentional acts to influence the behaviour of others, and to initiate interactions such as play and grooming.

Since Plooij’s study research on cognitive skills underlying gestural communication in great apes has expanded remarkably (Cartmill & Byrne, 2007; for orangutans: Liebal et al., 2006; for gorillas: Pika et al., 2003; for bonobos: Pika et al., 2005b; for chimpanzees:

Tomasello et al., 1994). These studies have firmly established that great apes frequently use gestures in their natural communication and that they have elaborated gestural

(18)

repertoires (for review see: Call & Tomasello, 2007). Great ape gestures qualify as intentional signals (Leavens et al., 2005) due to the overwhelming of evidence that apes produce gestures flexibly in social contexts (Arbib et al., 2008; Call & Tomasello, 2007) and are influenced by the attentional state of the recipient (e.g., Cartmill & Byrne, 2007;

Roberts et al., 2014b). In addition, research in this field has also demonstrated that apes use gestures as sophisticated communicative means (1) characterized by their flexible relation between signal and outcome (means-ends dissociation), and (2) used to attract and redirect attention. Means-ends dissociation implies that individuals are able to use different signals/gestures to achieve the same outcome/goal, and a single gesture for several outcomes (Bates et al., 1975; Bruner, 1981b; Pika & Liebal, 2012a; Plooij, 1978)..

A crucial limitation of the majority of previous studies on gestural communication is their restriction to captive settings (Cartmill & Byrne, 2007; Liebal et al., 2006; Pika et al., 2003, 2005b). Although captive studies provide excellent data for detailed analyses, they inevitably affected from contact with humans or through modification of the environment (Bard, 1992; Boesch, 2007; Tanner & Byrne, 1996). To understand the role gestures may have played in the dawn of human language it is crucial to complement findings on captive groups with studies on populations living in their natural environment, where active selection pressures are at work (Boesch, 2007). Although recently researchers started to investigate great ape gesturing in populations in the wild (Genty et al., 2009;

Hobaiter & Byrne, 2011a; Pika & Mitani, 2006; Roberts et al., 2012), these few studies have so far focused on only a single group each. In addition, only few studies have examined the effects of age, sex, and relationships between interactants (e.g., Pika &

Mitani, 2006).

Box 1. Definition of gesture in primate research

Gestures are movements of the extremities or the body, or postures that are directed towards a specific recipient, mechanically ineffective and potentially receive a voluntary response (Scott & Pika, 2012)

(19)

1.4 The ontogeny of great ape gestural communication

The flexible and intentional production of gestural signals in great apes is well understood. Much less is yet known about the development and cognitive prerequisites of these signals. So far only few studies have applied a developmental approach investigating ape gestural communication in great apes (Halina et al., 2013; Schneider et al., 2012a; Tomasello et al., 1994; Tomasello et al., 1997). Nonetheless, this method is crucial for understanding the processes underlying communicative skills and their cognitive prerequisites (Bard, 2010; Bruner, 1972). Moreover, given the large body of work on ape gestural signalling and its cognitive prerequisites, surprisingly little is known about the ‘cradle’ of communication: mother-infant coordination as co-regulated social interaction (King, 2004). A large body of research has emphasized the benefit of conceptualising the mother-infant dyad as a model system (for a review see, van de Rijt- Plooij & Plooij, 1987). This system assumes that the mother-infant dyad behaves as an organised unit characterised by mutual modification of each other's behaviour in response to feedback (Watzlawick et al., 1967). Pioneering work has been carried out by Plooij (1978, 1979) forty years ago, who investigated gestural ontogeny in mother-infant communication in chimpanzees at Gombe, Tanzania. He showed that communicative development between chimpanzee infants and their mothers progresses similarly to humans, with a shift around an infant’s age of 9 to 12 months from acts without social- communicative intention to intentional acts. At this age, the infant is not only able to maintain an interaction, e.g. ‘play-tickling’, but also to initiate it by using behaviours whose values have been established in earlier sessions (Plooij, 1978). In addition, Goodall (1967) described several visual signals used by wild chimpanzee infants to re-establish physical contact with their mothers (e.g., REACH ARM accompanied by pout face and HOO WHIMPER). These examples point towards the importance of intentional visual gestures as the physical distance between mother and offspring increases with development (Bard et al., 2005).

Currently, there is a debate about whether ape gestures are acquired through learning or whether they are genetically predisposed (Hobaiter & Byrne, 2011a; Liebal &

Call, 2012). Most researchers argue that the majority of ape gestures are learned (Pollick

& De Waal, 2007; Tomasello et al., 1997) and thus more similar to human language than other modes of communication used by apes. In terms of how nonhuman primates acquire

(20)

their gestures during ontogeny, researchers have differentiated between individual and social learning processes (Pika & Liebal, 2012a; Pika et al., 2005a). If individuals acquire the same behaviour independently due to ‘similar learning environments’, it is called individual learning. In contrast, social learning describes a mechanism in which observation of other individuals leads to the acquisition of a novel behaviour (Call, 1999).

For a better understanding of how particular gestures are learned, has been proposed to look at between-individual and between-group differences in gestural usage.

Tomasello and colleagues (1994) studied the acquisition of gestures in captive chimpanzees and suggested that chimpanzees acquire their gestures through an individual learning process termed Ontogenetic Ritualization (OR). In Ontogenetic Ritualization, an individual A first performs a physically effective behavioural sequence to achieve its goal of influencing individual B; over the course of many dyadic interactions, B learns to anticipate A’s likely forthcoming behavioural sequence on the basis of its initial step and responds ‘early’. Subsequently, A relies on B’s anticipation, producing only the initial movement to achieve its goal in a ritualized form: A’s behaviour has become a communicative signal over time (Tomasello & Call, 1997). Consequently, each dyad, even within a social group, might use different gestures for the same purposes. Thus, similarities between the gestural repertoires within a group, and the occurrence of group- specific gestures would provide evidence for the existence of a social learning process, whereas individual differences that overshadow group differences (i.e. a lack of systematic group differences, idiosyncratic gestures) would imply that an individual learning process is involved (see Box 2). Some group-specific gestures have been found in captive groups of chimpanzees (Tomasello et al., 1997), gorillas (Pika et al., 2003), and orangutans (Liebal et al., 2006) but also in wild populations. For instance, the GROOMING

HANDCLASP has been observed at the chimpanzee communities Mahale, Kanyawara and

Ngogo (Ghiglieri, 1984; McGrew & Tutin, 1978) but not at Budongo and Gombe, indicating social (‘cultural’) learning of this behaviour (Whiten et al., 1999).

Nevertheless, little is known about how new gestures spread between individuals of a group (Liebal & Call, 2012).

Recent studies on wild chimpanzees and gorillas (Genty et al., 2009; Hobaiter &

Byrne, 2011a) challenge the idea of OR and proposed a genetically predisposed, species- specific gestural repertoire similar to the relatively fixed behavioural patterns described by van Hooff (1973). This was largely due to the finding that gestures differed

(21)

systematically from the non-intentional, physical actions in how they were performed and the absence of idiosyncratic gestures in the wild (Hobaiter & Byrne, 2011a). Hence, the production of at least some gestures, such as the CHEST-BEAT of the gorilla, is likely more due to genetic predisposition. In addition to these three mechanisms (OR, social learning and genetic endowment), Perlman and colleagues (2012) suggested the on-line (‘real- time’) adaptation of action as another process that is involved in the acquisition of ape gestures. As recently proposed by Bard and colleagues (2014), it is likely that there are different modes of acquisition for different gesture types, while the bulk of gestures are yet still co-constructed as a result from social interactions.

In an attempt to solve this debate, one of the aims of my PhD-thesis (Chapter I) was to test the predictions for gestural acquisition generated by the three major processes (Box 2): If gestural usage was mainly due to genetic factors, one would expect high rates of concordance both between and within non-neighbouring groups, as well as the absence of idiosyncratic (individual-specific) or group-specific gestures. OR on the other hand predicts low rates of concordance both between and within groups, presence of idiosyncratic but absence of group-specific gestures. In case of social learning being the major mechanism for gesture acquisition, high rates of concordance within but not between groups, presence of group-specific but absence of idiosyncratic signals are expected.

Box 1.2. Predictions for routes of gestural acquisition Mode of gesture

acquisition

Concordance within groups

Concordance between groups

Idiosyncratic gestures

Group-specific gestures I. Ontogenetic

Ritualization low low present absent

II. Social learning high low absent present

III. Genetic pre-

disposition high high absent absent

(22)

1.5 The model system Pan

To unravel the evolution of human language, the majority of comparative studies has been focusing on the communicative abilities of our closest living relatives: chimpanzees (Pan troglodytes) and bonobos (Pan paniscus). Thus far, systematic comparisons of the communicative abilities between the two species, using a consistent methodology and taking into account within-species variability, are lacking (Chapter II). Comparisons of the social behaviour between these sister species have revealed a number of striking differences that have been related to species-specific social structures: female dominance (Parish, 1994), low aggression levels within and between communities (de Waal, 1989;

Kuroda, 1980), frequent male-female and female-female associations, and a frequent and varied sexuality (de Waal, 1987; Kano, 1992) had been considered as bonobo- characteristic and markedly different from chimpanzees. However, accumulating data from field researchers has questioned the behavioural dichotomy between the two Pan species (Doran et al., 2002; Stanford, 1998). Firstly, considerable inter-site variability has been reported regarding chimpanzee social behaviour in the wild. Thus, the question arises whether some of those stressed differences are rather related to different ecological conditions than to true interspecies variability (Boesch & Boesch-Achermann, 2000).

Secondly, the majority of studies on bonobo behaviour has been conducted in captive environments, and thus far profoundly more data from the wild is available for chimpanzees than for bonobos (Boesch et al., 2002).

1.6 Thesis outline

In my PhD-thesis I examine the complexity and ontogeny of gestural signalling in chimpanzees and bonobos in their natural environments. To target this aim, I combined the comparative with the developmental approach, which included horizontal (i.e. cross- sectional) comparisons between individuals of different groups, as well as vertical (i.e.

longitudinal) comparisons of the same individual between subsequent years. To fully understand the of communicative abilities of a species we also need to consider within- species variability (Boesch, 2007). Thus, I studied two distinct populations of eastern and

(23)

western chimpanzees (P. t. schweinfurthii at Kanyawara, Kibale National Park, Uganda, and P. t. verus at Taï South, Taï National Park, Côte d’Ivoire; Fig. 1.1). Moreover, for Chapter III I included corresponding mother-infant data from two communities of free- living bonobos (Pan paniscus): LuiKotale in Salonga National Park, and Wamba in the Luo Scientific Reserve, both in the Democratic Republic of the Congo (DRC).

Figure 1.1 Study sites for chimpanzee (Pan troglodytes verus/schweinfurthii) and bonobo communities (Pan paniscus) observed over the course of the PhD project (1: Taï South, 2: Kanyawara, 3: LuiKotale, 4:

Wamba). Species distributions for chimpanzees and bonobos were derived from Oates et al. (2009) and Fruth et al. (2008), respectively.

Studying the behaviour of chimpanzees and bonobos each in two different communities offered the rare opportunity to analyse behaviours of interest in a sample that was more likely representative of the species. Moreover, this dissertation differs considerably from previous gestural research that mainly focused on gestural interactions among dyads of all ages. First, I focused explicitly on communicative interactions between infants and mothers (Chapters II and III) or between infants and conspecifics, with conspecifics including mothers, maternal siblings, peers and other individuals (Chapters IV and V).

Second, I focused on three of the most frequently occurring social contexts – joint travel (Chapters II, III and V), social play (Chapter IV and V) and food sharing (Chapter V) – to examine communicative interactions between chimpanzee infants and their social

!"#$%&'()*+,-

!.&%&'()*+,-

(24)

partners. I focused on specific communicative functions, since keeping the behavioural outcome (i.e. joint travel, social play or food sharing) constant facilitated to investigate the study of gestures with the same ‘meaning’ (Smith, 1965). In recent years, this approach has been widely used to address questions related to cognitive abilities underlying gestural communication in great apes (Genty & Zuberbühler, 2014; Pika &

Mitani, 2006; Roberts et al., 2014b). To also assess general developmental transitions of the infants and to ensure comparability across species, this study was built on methods developed by researchers investigating communicative development of human children (Bates et al., 1975; Keller et al., 2011) and chimpanzees (Bard, 2010; van Lawick- Goodall, 1968). In the following paragraphs, I will provide a brief overview of each chapter of my PhD thesis.

In Chapter II, I examined communicative interactions of mother-infant dyads by focusing on a unique communicative function: the initiation of carries for joint travel. The aim was to gain a better understanding of the complexity and variability of communicative exchanges of mother-infant dyads, which is considered the core niche in which communication is learned (King, 2004). Additionally, I aimed to shed light on gestural acquisition in great apes, which has recently been subject to a lively debate (Liebal & Call, 2012). Since previous studies (Halina et al., 2013; Plooij, 1978) had suggested that the communicatory context of joint travel represents a promising candidate for frequent communicative exchanges between mother-infant dyads about a distinct goal (leaving a location), I focused my research efforts on this single communicative function.

In addition the joint travel interactions of chimpanzee mother-infant dyads, I had the opportunity to collect the corresponding data in two bonobo communities. Our aims in Chapter III were two-fold: First, I revisited the claim that communicative interactions of our closest living relatives, bonobos and chimpanzees, lack the cooperative nature of human communication (Levinson, 1995; Tomasello, 2008). Second, I investigated whether bonobos comprise the better model species for understanding the precursors of human communication, as argued by (Pollick & De Waal, 2007). To target these aims I tested and expanded some of the parameters derived from human conversation analysis by Rossano (2013) in situ (i.e. in natural environments) by focusing on interactions of mother-infant dyads. To understand potential differences in communicative abilities between species, I took into account within-species variability by including two bonobo and two chimpanzee communities (Boesch, 2007).

(25)

Next to learning mechanisms, little is known about apes’ flexible usage of gestures in natural environments. In Chapter IV, I thus aimed to examine whether chimpanzee are capable to adjust their communicative signals to certain attributes of conspecifics, such as age, sex and relationship to the signaller. Here, I focused on the communicative function of play solicitation to investigate the influence of demographic factors and relationship between signaller and recipient in both chimpanzee communities. In terms of play- soliciting gestures I differentiated between ‘pure’, object-accompanied and self- handicapping gestures.

Chapter V addresses the effects of social and locomotor development on gestural performance in chimpanzee infants. I used a novel combination of video recordings and focal scans of chimpanzee infants living in both communities. Hence, I traced several domains of development by linking interaction rates and mother-infant proximity to gestural signalling for the initiation of mother-infant joint travel, social play and food sharing.

By applying the comparative approach on our closest living relatives, insights into the communicative abilities that have been present in our last common ancestor can be gained (Hewes, 1973). In turn, this might shed new light on the question about which role gestures played in the evolution of human language (Arbib et al., 2008; Corballis, 2002).

(26)

(27)

Should I stay or should I go?

Initiation of joint travel in mother-infant dyads of two chimpanzee communities in the wild

Marlen Fröhlich, Roman M. Wittig & Simone Pika Animal Cognition (2016): 19(3), 483–500

(28)

2.1 Abstract

It is well established that great apes communicate via intentionally produced, elaborate and flexible gestural means. Yet relatively little is known about the most fundamental steps into this communicative endeavour—communicative exchanges of mother-infant dyads and gestural acquisition; perhaps because the majority of studies concerned captive groups and single communities in the wild only. Here, we report the first systematic, quantitative comparison of communicative interactions of mother–infant dyads in two communities of wild chimpanzees by focusing on a single communicative function:

initiation of carries for joint travel. Over 156 days of observation, we recorded 442 actions, 599 cases of intentional gesture production, 51 multi-modal combinations and 80 vocalisations in the Kanyawara community, Kibale National Park, Uganda, and the Taï South community, Taï National Park, Côte d’Ivoire. Our results showed that (i) mothers and infants differed concerning the signal frequency and modality employed to initiate joint travel, (ii) concordance rates of mothers’ gestural production were relatively low within but also between communities, (iii) infant communicative development is characterised by a shift from mainly vocal to gestural means, and (iv) chimpanzee mothers adjusted their signals to the communicative level of their infants. Since neither genetic channelling nor ontogenetic ritualization explains our results satisfactorily, we propose a revised theory of gestural acquisition, social negotiation, in which gestures are the output of social shaping, shared understanding and mutual construction in real time by both interactants.

2.2 Introduction

Across cultures and languages, human children enter language hands first. It has been hypothesised that this brief period in human ontogeny recapitulates phylogeny, with gestures being the modality out of which human language may have blossomed (for an overview see Hewes, 1973). This so called gesture-first hypothesis especially inspired comparative researchers to search for evolutionary precursors to human language in nonhuman primate gesturing (Tomasello, 2008). Systematic studies in the last decades have shown that gestures indeed are used as intentionally produced, elaborate and flexible

(29)

communicative strategies and play, similar to vocalisations, a crucial role in great apes’

everyday communication (for overviews see: Call & Tomasello, 2007; Pika & Liebal, 2012b). While there is a large body of work focusing on the description of gestural repertoires in a variety of different primate species (Call & Tomasello, 2007; Genty et al., 2009; Hobaiter & Byrne, 2011a), usage of distinct gesture types (Leavens et al., 1996;

Pika & Mitani, 2006) and cognitive mechanisms underlying gestural signalling (Genty &

Zuberbühler, 2014; Liebal et al., 2004; Pika & Mitani, 2006; Roberts et al., 2014b), surprisingly little is known about the first step into this communicative endeavour:

mother-infant coordination as co-regulated social interaction (King, 2004).

A large body of research has been emphasizing the benefit of conceptualising the mother-infant dyad as a system decades ago (for a review see, van de Rijt-Plooij & Plooij, 1987). This system assumes that the mother-infant dyad behaves as an organised whole characterised by mutual modification of each other's behaviour in response to feedback (Watzlawick et al., 1967). Pioneering work has been carried out by Plooij (1978, 1979) forty years ago, who investigated gestural ontogeny in mother–infant communication in chimpanzees at Gombe, Tanzania. He showed that, similar to communicative development in human children, interactions between chimpanzee infants and their mothers slowly progress, with a shift around the ages of 9–12 months from acts without social-communicatory intention to intentional acts. At this age, the infant is not only able to maintain an interaction, e.g. ‘play-tickling’, but also to initiate it by using behaviours whose values have been established in earlier sessions (Plooij, 1978). Plooij thus concluded that gestures in chimpanzees do not represent innate signals but are acquired through a process of ‘social negotiation’ (also termed 'conventionalization'; Mead, 1910).

This idea was later developed into a formal hypothesis, ‘Ontogenetic Ritualization’ (OR), in which the forms that gestures take derive directly from repeated social interactions in which individuals participate (Tomasello et al., 1994). Thus, evidence for the process of OR would be high degrees of individual variation within dyads, groups and between communities but also concerning the means used to achieve the same goals. Halina and colleagues (2013) recently investigated mother-infant coordination for the purpose of joint travel (carries) in captive bonobo (Pan paniscus) mother-infant dyads and were able to attribute the process of OR to several carry-initiating gestures. This study, thus, supported the hypothesis of Tomasello and colleagues (1994; Call & Tomasello, 2007) that gestures are acquired via repeated social interactions. For current purposes, the term individual learning refers to a process in which two or more individuals independently

(30)

acquire the same behaviour due to ‘similar learning environments’ (Whiten & Ham, 1992). Contrarily, the term social learning is used to indicate situations in which individuals learn distinct behaviours by imitating (Bandura, 1986) but also by interacting and observing each other. Recently, Byrne and colleagues (Genty et al., 2009; Hobaiter &

Byrne, 2011a) challenged the idea that learning plays a role in great ape’s gestural production and suggested that similarly to vocal production and facial expressions, gestures appear hard-wired and can be explained as a result of genetic channelling during development alone. This hypothesis is in contrast to great apes’ high degree of manual flexibility in other behavioural domains such as food processing and tool-use, and considerable inter-site variability (Byrne et al., 2011; van Schaik et al., 2003; Whiten et al., 1999). However, since systematic quantitative comparisons of gestural signalling in wild populations are still lacking the absence of evidence might merely reflect a paucity of data, rather than a lack of gestural complexity on behalf of the apes.

The aim of the present study was to gain a better understanding of the complexity and variability of communicative exchanges of mother-infant dyads and to shed light on gestural acquisition. To do so, we enabled the first systematic quantitative comparison of gestural signalling in two chimpanzee communities of different subspecies in their natural environments (Kanyawara, Kibale National Park, Uganda, and Taï South, Taï National Park, Côte d’Ivoire). Since other studies (Halina et al., 2013; Plooij, 1978) had suggested that the communicatory context of joint travel represents a promising candidate for frequent communicative exchanges between mother-infant dyads about a distinct goal (leaving a location), we focused our research efforts on this single communicative function. To enable horizontal comparisons between individuals of different communities and vertical comparisons of the same individuals, behavioural data were collected in two consecutive years. This important methodological tool for understanding the cognitive prerequisites underlying different communicative skills had so far only been employed in captive settings (Pika et al., 2003; Schneider et al., 2012a; Tomasello et al., 1997).

We addressed the following three questions:

First, which behaviours do chimpanzees employ to initiate joint travel? Plooij (1978) for instance had noted that mothers who initiate joint travel (i) lower their bottoms, (ii) look back at their infants, (iii) reach back towards him/her and (iv) make tonal grunts. They thus employ a complex set of gestures (LOWER BACK, LOOK BACK, REACH BACK¹) and

1From now on gesture and vocalisation types are depicted in SMALL CAPITALS.

(31)

multi-modal combinations (LOOK BACK and GRUNT) to communicate the distinct goal of joint travel and also the direction to travel to. To investigate this question, we compiled individual repertoires of behaviours produced to initiate joint travel and analysed signal production in terms of gesture category (e.g., visual or tactile) and signal modality (gesture, vocalisation or combinations of the two, i.e. multi-modal signals). We expected chimpanzee mothers in the wild to be the main carry initiators thereby contributing the majority of travel-initiating behaviours (van Lawick-Goodall, 1967).

Second, are gesture types employed to initiate joint travel due to learning (including both individual and social learning) between mothers and infants or can their production simply be explained as a result of genetic channelling? Since it is impossible to observe developmental processes as they unfold over time under natural conditions, a window approach onto gesture acquisition was applied: We investigated the degree of variability in gestural production to initiate joint travel within dyads, within communities and between communities (Pika et al., 2003, 2005b). Furthermore, since the presence of idiosyncratic gestures is a key indicator of individual learning and evidence against a phylogenetic origin of gestures, we examined whether idiosyncratic gestures were employed (found to be used by only a single individual of the whole community over two subsequent years and study periods). Pronounced variability in individual gestural production across dyads and communities (e.g. low concordance rates between individuals’ repertoires and idiosyncratic gestures) would provide evidence for the impact of learning in mother-infant communication, whereas high rates of concordances in gestural variability across dyads and communities may imply genetic channelling.

Third, do chimpanzee mothers adjust their behaviour to the developmental stage of their infants, and how does infant age influence signal production in both mothers and infants? As suggested by Plooij (1978), the means mothers employ to communicate with their infants might be influenced by the developmental shift from actions to intentional communication in young chimpanzees. In addition, a mother’s accumulated experience in interactions with previous offspring might also shape the carry interaction substantially, as well as the prevailing behavioural context (i.e. varying necessity to carry). For instance, while frequent gestural interactions can often be observed in evolutionarily non- urgent, or ‘relaxed’, situations (e.g., playing and grooming; Pika, 2014; van Lawick- Goodall, 1967), they sometimes outrival vocalisations in evolutionary ‘urgent’ contexts, where silent communication transfer is an advantage (e.g., consortship; Hobaiter &

Byrne, 2012).

(32)

2.3 Methods

2.3.1 Study sites and subjects

The study investigated the communicative behaviour of mother-infant dyads in two different chimpanzee communities: Kanyawara in Kibale National Park, Uganda (P.

troglodytes schweinfurthii), and Taï South in Taï National Park, Côte d’Ivoire (Pan t.

verus). Detailed descriptions of the study areas can be found in Wrangham and colleagues (1992) and Boesch and Boesch-Achermann (2000). During the two study periods, the size of the Kanyawara group varied between 53 and 56 individuals, respectively 21 and 24 in Taï South. The Kanyawara and Taï chimpanzees are well habituated and have been studied regularly since 1987 (Wrangham et al., 1992) and 1979 (Boesch & Boesch- Achermann, 2000), respectively, enabling dawn-till-dusk follows and the collection of high-quality recordings. In addition, we had access to long-term data concerning the chimpanzees’ demography, social relationships, relatedness and ranks. We observed communicative interactions of a total of 13 mother-infant dyads (seven from Kanyawara, six from Taï South), with offspring ranging from 9 to 69 months of age (Tab. 2.1). At Taï one mother gave birth to another infant in the second field period, hence we observed twelve chimpanzee mothers and 13 infants.

2.3.2 Data collection

Observations were made on chimpanzees of the Kanyawara community in Kibale National Park and the Taï South group at Taï National Park during four periods between October 2012 and June 2014 (Kanyawara: Mar–May 2013, Mar–Jun 2014; Taï South:

Oct–Dec 2012, Oct–Dec 2013). We used a focal behaviour sampling approach (Altmann, 1974), while maintaining a record of the frequency with which a particular dyad had been observed. In situations where we could choose which of several dyads to film, we targeted those individuals previously sampled least often. Following Hobaiter and Byrne (2011a), who had suggested that approximately 15 h of active gesture time or approximately 150 days of field observation time would enable to assess the whole gestural repertoire of a given chimpanzee community (N = 82), we observed all 13 mother-infant dyads for a total of 156 days. All social interactions of mothers and infants

(33)

(i.e. mother-infant interactions as well as mother-conspecific and infant-conspecific interactions) that were judged to have any potential for communicative interactions were recorded using a digital High-Definition camera (Canon HF M41) with an external unidirectional microphone (Sennheiser K6). This method resulted in a total of 169 hours of video footage recorded during approximately 1198 hours of focal observations (see Table 2.1 for further details). However, the present paper focuses only on the communicative context of carry initiation, thus our analysis is based on a total of 410 high-quality recordings of mother-infant behaviour in this respective context (mean recordings per dyad: 33.2). In addition, every 15 minutes we conducted a focal scan by using a Personal Digital Assistant (HP iPAQ rx1959) with focal/time sampling utilized as sampling/recording rule (Altmann, 1974). This method enabled us to collect data on a variety of additional parameters such as behavioural context and party composition (see Appendix, Table 2.S1), resulting in a total of 4505 behavioural scans.

Table 2.1 Information on observed mother-infant dyads with respective observation time and raw data set

Study site Dyad

(inf/mot) Infant sex Infant age P1 [months]

Infant age P2 [months]

Observ.

time [h]

Recorded interactions [h]

WZ/WA M 9–11 21–23 105 15.2

KANYAWARA

OB/OT M 13–15 25–27 119 18.4

MM/ML F 13–15 25–27 87 7.3

LL/LN F 3–5^a 15–17 60 7.2

TR/TG F 16–18 28–30 112 15

OL/OU F 48–50^b N/A 45 7.2

WC/WL M 55–57 67–69 73 10.1

TAÏSOUTH

MH/MB F 10–12 22–24 150 11.2

IN/IS M N/A 10–12 91 12

SL/SM M 15–17 27–29 147 14.7

JF/JL M 15^c N/A 20 0.4

KY/KS F 19–21 31–33 148 17.0

IT/IS M 64–66^b N/A 41 9.5

N 13 6:7 11 10 1198 145.2

The last line provides the total sample size for each column (P1/P2: first/second period of data collection)

a P1 not included since infant was too young

b Mothers gave birth to sibling in P2, thus no P2 data available

c Deceased on Nov 1, 2012

2.3.3 Video coding procedure

To establish the behavioural repertoires of mothers and infants used to initiate maternal carries and enable subsequent analyses, a total of 410 high quality video files of mother-

(34)

offspring carry initiations (i.e., carries with clear visibility of carry-initiating behaviours) were coded using the program Adobe Premiere Pro CS4 (version 4.2.1.). In addition, we included PDA recordings of five interactions, resulting in a total of 415 interactions.

Behavioural definitions were based on established ethograms of the behaviour of two long-term studies of eastern chimpanzees (Goodall, 1986; Nishida et al., 1999) and several gesture studies (Call & Tomasello, 2007; Hobaiter & Byrne, 2011a; Roberts et al., 2014a). Based on parameters used in previous work on great ape gesturing (Pika et al., 2003, 2005b; Pika & Mitani, 2006) a coding scheme was developed. For our purposes, all analysed joint travel events included maternal carries (i.e. involving mother-infant body contact). While coding all agent-initiated carries, we differentiated between carry- initiating behaviours via (i) physical actions, (ii) intentionally produced gestures, (iii) multi-modal combinations (gesture plus vocalisation), and (iv) vocalisations. A physical action was defined as any behaviour that resulted in joint travel through direct manipulation of another’s body or the movement of one’s own body into a carry position.

Carry-initiating actions included for instance grabbing, forcibly pulling, lifting or approaching another individual (see Appendix, Table 2.S2). Gestures were defined as directed, mechanically ineffective movements of the body or body postures that elicited (‘requested’) a voluntary response by the recipient (Pika, 2008). In addition, we only included those gestures in our analyses that were accompanied by one or more key characteristics of intentional communication (Bates, 1976; Bruner, 1981c; Pika et al., 2003):

Sensitivity to the attentional state of the recipient: The signaller shows signs of being aware of the recipient’s state of attention, e.g. by using visual gestures only when the recipient is looking.

Response waiting: The signaller pauses at the end of the signal and waits for at least two second for a response while maintaining visual contact.

Apparent satisfaction of signaller: The signaller’s communication ceases when the apparent goal has been met by the recipient (Hobaiter & Byrne, 2014).

Goal persistence: The signaller elaborates her signalling when thwarted, e.g. by repeating and exaggerating the signal or by using a different communicative means (Pika et al., 2005b; Pika & Mitani, 2006).

Gestures were clustered into three signal categories: audible (signals generate a sound while being performed, e.g., SLAP GROUND), tactile (signals include physical contact with the recipient, e.g., TOUCHING), and visual (signals generate a mainly graphic component,

(35)

e.g., RAISE ARM) signals (Pika et al., 2003). To identify carry initiations, the behaviour of both, the signaller and the recipient throughout the interaction, from first initiating action/gesture to start of carry, was taken into account to assess the success of communicative attempts (Smith, 1965). Idiosyncratic gestures, which are exclusive for single individuals in the whole community, had been observed at least three times to be included in the analyses (Pika et al., 2003, 2005b). Vocalisations, especially those accompanying gestures (‘multi-modal signals’), were analysed in terms of their broad categories (Crockford & Boesch, 2005; Goodall, 1986). Finally, for each signal or action case, we coded the following parameters: interaction role of the signaller: 2 levels, mother, infant; infant age: range = 9–69 months; necessity of carry: 2 levels (low; carry preceded by feeding, playing, resting, relaxed group travel; high: preceded by aggressive behaviours such as chasing and hitting, catching up with already left party/group, displaying and patrolling); mother’s parity: number of offspring reared at least until juvenility (plus present infant), range = 1–5, party composition: 3 levels (mother with dependent offspring only, adult females only, mixed group). A least fifteen per cent of all mother-infant interactions were coded for accuracy by a second observer and tested using the Cohen’s Kappa coefficient to ensure inter-observer reliability (Altmann, 1974). A

‘very good’ level of agreement was found for gesture type (κ = 0.88), signal type (κ = 0.81), signal category (κ = 0.84), and necessity of carry (κ = 0.82). The level of agreement for carry initiator (mother/infant) was ‘good’ (κ = 0.80).

2.3.4 Statistical analyses

Since Byrne and colleagues (Genty et al., 2009; Hobaiter & Byrne, 2011a) had argued that differences in gestural repertoires of captive apes were simply premature assumptions, with repertoires yet to reach asymptote, we plotted the cumulative numbers of observed gesture types over time for all individuals. If an asymptote was reached (i.e.

no further gesture types were observed), we concluded that we had observed the individual’s full repertoire for the specific communicative function of maternal carries.

We measured the relationship between an individual’s final repertoire size and the total time that individual had been observed using the Spearman R statistic. For our repertoire analyses we included only individuals observed for over 60 hours (N = 10; observation time range: 60.25–150 h, mean ± SD = 109.3 ± 32.1 h), which have reached the critical

Communicative complexity and development in chimpanzees (Pan troglodytes) and bonobos (Pan paniscus) in the wild