Review articleMeasuring the evolution of facial ‘expression’ using multi-species FACS
Introduction
The evolution of modern Homo sapiens is thought to have been accompanied by rapid cognitive and behavioural change equipping modern humans with many complex and unique traits (Tomasello, 2008). Some of the most complex and interesting aspects of human mind and behaviour are therefore unique to humans. Human language, for example, is not found in other species and is thus thought to have evolved fairly recently in the hominid lineage, possibly as recently as 50,000 years ago (e.g. Klein, 2017). Such uniqueness renders a comparative approach to some aspects of cognition and behaviour challenging, as scientists need to investigate the precursors to these traits without the option of examining clear, unambiguous counterparts in other extant primates. Human facial behaviour (observable changes in the appearance of the face, often termed facial ‘expression’), in stark contrast, has an abundant array of similar phenomena (in both form and function) throughout the primate order and in other mammals (Waller and Micheletta, 2013). This broad continuity across species suggests that facial expression evolved long before the arrival of modern humans, and that human facial behaviours are rooted in evolutionarily ancient displays. In comparison to some other human traits, scientists are therefore presented with a much easier task when trying to understand the evolutionary trajectory and function of facial behaviour.
Despite the excellent scientific opportunity presented by the existence of similar facial behaviour across species, there are still considerable theoretical and methodological challenges. The field of comparative facial communication research has attracted (and continues to attract) divergent theoretical approaches. First, scientists disagree on which criteria are needed to identify behaviours of shared descent, and thus how to identify unambiguous counterparts. Second, facial behaviours form part of a complex system of production (in the sender) and perception (in the receiver), in which we (the scientists) take part when we make observations. We argue that precise and objective methodology is therefore essential when studying facial behaviours in any species, to avoid biasing observations with our own categorical and emotional interpretation. For example, chimpanzee bared-teeth faces are perceived as more similar to human smiles when the underlying emotion is judged as similar, suggesting observers find it hard to distinguish meaning from form (Waller et al., 2007). Here, we review the development of objective anatomically based systems for the measurement of facial behaviour across species. The Facial Action Coding System (FACS) was originally developed for humans (Hjortsjo, 1969; Ekman et al., 2002; Ekman and Friesen, 1978) and has since been modified for use with several other animal species. We strongly advocate the use of these systems for comparative facial behaviour analysis and discuss how these can be used to better understand the evolution of facial behaviour.
Section snippets
How can we identify homologous facial behaviour?
“As scientists we want to know how justified our feelings of familiarity and understanding are and to what extent our impressions of oddity are based simply on anthropocentrism. To a biologist these questions translate into the question of whether a common heritage disposes us to understand some of the primate facial displays but not others, or whether the perceived similarities and dissimilarities are only superficial and disappear on closer examination.” (Preuschoft and van Hooff, 1995, p.
What is FACS?
The investigation of human facial non-verbal communication has been greatly facilitated and standardised by the development of the Facial Action Coding System (FACS: Ekman et al., 2002; Ekman and Friesen, 1978). Prior to this, the human facial behaviour field was reliant on more subjective methods and did not have a systematic way to assess the muscular components of facial behaviour (and thus help determine homology, see above). Duchenne de Boulogne (1862), however, was the first scientist to
The development of multi-species FACS
FACS has been modified for use with non-human species to facilitate objective facial behaviour measurement. However, there are important assumptions in this endeavour that should be taken into account. The human FACS is based on the assumption that what can be observed by a scientist is similar to what is perceived during human-human social interaction. Thus, the units of FACS (AUs) are assumed to be reasonable units of human perception. This makes sense when investigating human-human
Applications of multi-species FACS
Despite the increase in the number of FACS systems available, there are relatively few studies using this tool to answer questions about the evolution and function of facial behaviours. However, the studies that are available demonstrate how FACS can be a powerful tool to inform us about 1) the phylogenetic link between the facial behaviours of humans and other animals, 2) the cognitive mechanisms underpinning the production and perception of facial behaviours, and 3) the social and ecological
Conclusion and future directions
Comparative analysis of facial behaviour across species attracts different methods and theoretical approaches. Following on from the seminal work of Preuschoft and van Hooff (1995), we argue that to be considered homologous across species, and therefore similar through common descent, facial behaviours need to demonstrate the following: a stereotyped and recognisable form (as defined by the recognition of conspecifics), similarity of multiple elements, homology of underlying facial musculature
Acknowledgements
We would like to thank Mariska Kret, Eliza Bliss-Moreau and Jorg Massen for inviting us to be part of this special issue. We would also like to thank everyone who has contributed to the development and management of the different animal FACS systems. We also thank the funders for the various FACS systems (to us and others): Leverhulme Trust (ChimpFACS and NetFACS), NIH (MaqFACS), British Academy (modification of MaqFACS for crested macaques), Waltham Foundation (DogFACS), DFG (GibbonFACS) and
References (104)
- et al.
Development and application of CatFACS: Are human cat adopters influenced by cat facial expressions?
Appl. Anim. Behav. Sci.
(2017) - et al.
Reconstructing the evolution of laughter in great apes and humans
Curr. Biol.
(2009) - et al.
Receiver psychology and the evolution of animal signals
Anim. Behav.
(1991) Language and human evolution
Journal of Neurolinguistics, Language Evolution: On the Origin of Lexical and Syntactic Structures
(2017)- et al.
Detection, tracking, and classification of action units in facial expression
Rob. Auton. Syst.
(2000) - et al.
Emotional expressiveness of 5–6 month-old infants born very premature versus full-term at initial exposure to weaning foods
Appetite
(2016) - et al.
Facial expression recognition in rhesus monkeys, Macaca mulatta
Anim. Behav.
(2009) - et al.
Facial expressions as honest signals of cooperative intent in a one-shot anonymous Prisoner’s Dilemma game
Evol. Hum. Behav.
(2012) - et al.
Holistic processing of faces happens at a glance
Vision Res.
(2009) - et al.
The language void: the need for multimodality in primate communication research
Anim. Behav.
(2011)