Original ArticleA structure-based repertoire of manual gestures in wild chimpanzees: statistical analyses of a graded communication system
Introduction
Insight into language evolution can be gained from the communicative systems of nonhuman primates (Slocombe, Waller, & Liebal, 2011), especially those of chimpanzees, our closest living relative (McGrew, 2010). Although not a ‘missing link,’ chimpanzees display a number of features considered characteristic of early human populations, such as fission–fusion social organisation and life at the forest–savannah interface (van Lawick-Goodall, 1968). The study of communication in chimpanzees contributes to our understanding of the probable communicative abilities present in our last common ancestor, helping to identify the evolutionary pressures that have shaped human communicative abilities (Tinbergen, 1963).
Studies of chimpanzee communication have primarily focused on compiling the vocal repertoire (Mitani, 1996). Signal repertoires are species-specific collections of ritualised actions or cues, deployed to change the behaviour of recipients (Bradbury & Vehrencamp, 1998). More recently, attention has shifted to gestures as a potential evolutionary precursor to human language (Slocombe et al., 2011). Manual gestures, defined as movements of the hands without the use of objects or a substrate, have attracted considerable attention because of the possibility of being an ancestral trait that humans share with their primate relatives (de Waal, 2003). It has been argued that manual gestures are governed by specific neurological structures homologous to the ones responsible for human language (Perrett et al., 1985). Only humans and other apes habitually use their hands to communicate (de Waal, 2003), and gestural communication in chimpanzees and bonobos shows greater flexibility than either facial or vocal signals (Pollick & de Waal, 2007).
Chimpanzee gestural behaviour has been studied in a few wild (Goodall, 1986, Hobaiter and Byrne, 2011, van Lawick-Goodall, 1968, McGrew et al., 2001, Nishida, 1970, Nishida et al., 2010, Plooij, 1978, Plooij, 1979, Reynolds, 1963, Sugiyama, 1969) and captive populations (van Hooff, 1971, Tomasello et al., 1984, Liebal et al., 2004, Pollick and de Waal, 2007). Gestures are used flexibly across a diverse range of contexts, including agonism, mating, grooming, and play (Pollick and de Waal, 2007, Robert et al., 2012); in mother–offspring interactions, gestures are pivotal in negotiating nursing and food sharing, and coordinating travel (Plooij, 1978). While these studies provide important insights into overall repertoire, manual gestures have been mostly studied within the broader framework of all communicative bodily movements. Moreover, not all studies investigated whether the observed signals are intentional, as determined from the signaller directing gestures to recipients in flexible, goal-directed ways (see e.g. Roberts et al., 2012). However, it is important to determine whether the observed behaviours are voluntary because the distinction between simple behavioural actions, which may be used by others to infer intentions and meaningful gestural communication, lies in determining whether the action is used intentionally (Bates, Benigni, Bretherton, Camaioni, & Volterra, 1979). For example, activities that regularly precede a particular event, such as those that indicate changes in activity state (i.e., between resting and locomotion), can become communicative to the receiver even though this signal is unintentional from the sender's perspective (Roberts, 2010, Robert et al., 2012).
Previous research has identified broad consensus on the chimpanzee repertoire of manual gestures across captive populations (Pollick & de Waal, 2007). However, comparisons are constrained by the lack of relevant or comprehensive illustration and the subsequent difficulty of accurately matching written descriptions of visual signals across studies (but see Arbib et al., 2008 and Nishida et al., 2010). There are few direct comparisons using the same methodology, but between-group variation seems to primarily reflect the relative frequency of different gestures, which determines the probability of observing a specific gesture at any one site (Pollick & de Waal, 2007; Hobaiter & Byrne, 2011). However, identifying differences between wild and captive repertoires is important because it is unclear to what degree human contact and a captive setting might influence the ontogeny of gestural signalling (Tomasello & Call, 2004). For example, captive apes point to distal objects, an important capacity in prelinguistic human development (e.g., Liszkowski, Carpenter, & Tomasello, 2008) and a behaviour thought to be absent in wild great apes (but see Veà & Sabater-Pi, 1998).
Gestures are often categorised by context or perceived function rather than precise morphology because it can be difficult to disentangle gestures from noncommunicative actions. Gestures can also be more broadly defined as including locomotor activities (e.g., bipedal jump), orofacial movements (e.g., lip-lock), or object-directed actions (e.g., throw object) (e.g., Liebal et al., 2004, Hobaiter and Byrne, 2011), with the result that repertoire sizes can differ considerably across studies at the same location (up to threefold: Liebal et al., 2004, Pollick and de Waal, 2007). Previous studies have also been heavily biased towards the play context (Slocombe et al., 2011), which may not reflect the form and function of these gestures more broadly. Play also overrepresents younger individuals, whose gesture repertoire may differ from adults, at least in chimpanzees (Tomasello et al., 1984), although this is confounded by age differences in overall gesture production rates (Hobaiter & Byrne, 2011).
Detailed morphology can be informative but is largely neglected or only selectively considered, for example, in examining the potential ritualisation of taking actions into reach gestures. However, perceived morphological variation is often considered to lack communicative function, and subsequently, gesture variants are lumped together (Hobaiter & Byrne, 2011; but see Roberts et al., 2012). The underlying assumptions are that a human observer can identify and classify meaningful units (in a similar manner to the animals themselves) and also that similar morphological properties reflect similarities in function. Although analysing gesture function requires accurate identification of morphological characteristics of different gestures, there is no systematic method for analysing gestural signals based on their morphology (but see Forrester, 2008). This hinders comparison of repertoires within and across species, and contrasts with more standardised, bottom-up approaches for categorising human posture (e.g., see Coulson, 2008). Morphological classification provides objective criteria to classify graded signals and addresses subjectivity in the lumping/splitting of behaviours. Such an approach also separates form and function during classification of gestures, avoiding potential biases in the interpretation of meaning and circularity in the study of gesture function.
In this study, chimpanzee gestures are analysed quantitatively based solely on their morphological features and without a priori assumptions about context or function. The component features of gestures are classified statistically and clustered into groups (Bortz, 1993). For example, if a chimpanzee extends its hand towards a receiver, this action can differ in both intensity and form of hand and arm shape, with fingers either flexed or stretched, a smooth, sweeping movement or a forceful, stretched, and linear action, in either the vertical or horizontal plane. A bottom-up analysis provides a more rigorous, quantitative description for each gesture type in terms of morphology and movement configurations. For example, human posture can be quantified in terms of the degree of rotation of major joints that underlie the relative position of head, torso, and arms, and also the tempo, plane, and direction of these movements (Coulson, 2008, Gross et al., 2010).
Importantly, only structural descriptions of gestures can also determine the extent to which the gestural repertoire of chimpanzees is discrete or graded in order to relate overall structure to social and ecological factors (Marler, 1976). A repertoire may consist of only discrete signals (with no intermediates between adjacent elements) or fluid signals that change continuously and grade from one prototypical form to another. Graded repertoires are considered advantageous in open habitats with close range social interaction, while discrete repertoires better suit poorer visual conditions and less frequent direct social interactions (Marler, 1976). Moreover, the structural nature of a repertoire can also be informative about the cognitive processes underlying gesture production and the amount of genetic control (McGrew et al., 2001).
Here, we provide first systematic analyses of manual gestures of wild chimpanzees using a quantitative morphological approach. We establish an inventory of intentional manual gestures using standard clustering techniques and subsequent validation with discriminate function analysis. The resulting clusters are compared with previously reported repertoires to explore potential differences between and within groups. Finally, morphological variation is examined in relation to repertoire size in two other communication modalities, vocalisations and facial expressions, to explore social and ecological influences on chimpanzee gestural communication.
Section snippets
Study site and subjects
Manual gestures of one community of habituated East African chimpanzees were examined over an 8-month period (September 2006, April–July 2007, and March–May 2008) at the Budongo Conservation Field Station, Budongo Forest Reserve. Budongo is in western Uganda on the edge of the western Rift Valley (1˚37′–2˚00′N; 31˚22′–31˚46′E) at a mean altitude of 1100 m. The reserve covers an area of 793 km2 composed of grassland and semideciduous tropical forest with predominantly continuous forest cover of
Determining manual gesture types and contextual usage in wild chimpanzees
The number of cases contributing to each gesture type is presented in Table 2. Mean±S.D. number of gesture types contributed by each focal individual was 7.26±2.59, N= 197 cases, 12 individuals; ad libitum data N= 21 cases, 6 individuals; with a range of 5–12 gesture types per focal subject. For the cluster analyses, we were able to include N= 218 gesture events. Hierarchical cluster analysis produced a tree representing 30 gesture types (Fig. 1). Overall, the morphological differences between
Discussion
While previous research has focused on examining the morphological complexity of vocal behaviour and facial expressions (Parr, Cohen, & de Waal, 2005), this is the first systematic quantitative demonstration of such complexity in gestural behaviour beyond qualitative descriptions (Reynolds, 1963, van Lawick-Goodall, 1968, Sugiyama, 1969, Plooij, 1978, Plooij, 1979, Nishida et al., 2010, Hobaiter and Byrne, 2011). Our study demonstrates that adult chimpanzees have a multifaceted and complex
Supplementary materials
The following is the Supplementary data to this article.
Acknowledgments
Dr Anna Roberts was funded by the Economic and Social Research Council 3 + fellowship and the University of Stirling. The authors wish to thank the Royal Zoological Society of Scotland for providing core funding for the Budongo Conservation Field Station, Uganda, as well as the Ugandan Wildlife Authority and the Uganda National Council for Science and Technology for support and permission to carry out research. We thank Dr Quentin Atkinson, University of Auckland, for use of his programme for
References (52)
- et al.
Twelve-month-olds communicate helpfully and appropriately for knowledgeable and ignorant partners
Cognition
(2008) - et al.
Visual analysis of body movements by neurons in the temporal cortex of the Macaque monkey—a preliminary-report
Behavioural Brain Research
(1985) - et al.
Referential gestural communication in wild chimpanzees (Pan troglodytes)
Current Biology
(2006) - et al.
The language void: The need for multimodality in primate communication research
Animal Behaviour
(2011) - et al.
Primate vocalisation, gesture and the evolution of human language
Current Anthropology
(2008) - et al.
Observing interaction: an introduction to sequential analysis
(1997) - et al.
The emergence of symbols
(1979) Statistic fuer Sozialwissenschaftler
(1993)- et al.
Principles of animal communication
(1998) - et al.
Different social motives in the gestural communication of chimpanzees and human children
Developmental Science
(2011)
Primate calls, human language, and nonverbal-communication
Current Anthropology
Expressing emotion through body movement
Emotions inside out: 130 years after Darwin's. The expression of the emotion in man and animals
The social brain hypothesis
Evolutionary Anthropology
Acoustic features of female Chacma baboon barks
Ethology
Statistical methods for rates and proportions
A multidimensional approach to investigations of behaviour: revealing structure in animal communication signals
Animal Behaviour
Gestural communication of the gorilla (Gorilla gorilla): repertoire, intentionality and possible origins
Animal Cognition
The chimpanzees of Gombe: Patterns of behaviour
Methodology for assessing bodily expression of emotion
Journal of Nonverbal Behavior
The gestural repertoire of the wild chimpanzee
Animal Cognition
Neuronal control of mammalian vocalization, with special reference to the squirrel monkey
Naturwissenschaften
Gesture: visible action as utterance
Genetic and ‘cultural’ similarity in wild chimpanzees
Proceeding of the Royal Society B
Analyzing multivariate data
Use of gesture sequences in chimpanzees
American Journal of Primatology
Cited by (47)
Using networks to visualize, analyse and interpret multimodal communication
2024, Animal BehaviourSequential and network analyses to describe multiple signal use in captive mangabeys
2021, Animal BehaviourBidding evidence for primate vocal learning and the cultural substrates for speech evolution
2017, Neuroscience and Biobehavioral ReviewsWhat is a gesture? A meaning-based approach to defining gestural repertoires
2017, Neuroscience and Biobehavioral ReviewsCitation Excerpt :In doing so great apes meet the criteria for 1st order intentional communication (Dennett, 1987). There is evidence for the 1st order intentional (hereafter intentional) use of one or two signal types in a very few non-ape species (e.g. grouper: Vail et al., 2013; macaque: Gupta and Sinha, 2016), but compare this with the extensive body of evidence for the intentional use of a large repertoire of gestures within all ape species in both captivity (chimpanzee: Tomasello et al., 1985, 1989, 1994; Halina et al., 2013; bonobo: Pika et al., 2005; gorilla: Tanner and Byrne, 1996; Pika et al., 2003; Genty et al., 2009; and orang-utan: Liebal et al., 2006; Cartmill and Byrne, 2007) and the wild (chimpanzee: Hobaiter and Byrne, 2011a,b, 2012, 2014; Roberts et al., 2012, 2014; bonobo: Graham et al., 2016). This large data set of intentional non-human signal use provides us with a unique opportunity: we are able to ask what a great ape gesture ‘means’ in a human language-like sense (Grice, 1957; Hobaiter and Byrne, 2014; Moore, 2016; although c.f. Scott-Phillips, 2015, 2016).
Protolanguage and mechanisms of meaning construal in interaction
2017, Language Sciences