Geometrical Gaze Following in Common Marmosets (Callithrix jacchus) Judith Burkart University of Zurich Adolf Heschl Konrad Lorenz Institute for Evolution and Cognition Research A series of experiments investigating the degree of gaze understanding in common marmosets (Callithrix jacchus) is reported. Results show that marmosets follow the gaze of a human experimenter readily and also use the gaze to locate food in a modified version of the object choice task if influences of chance probabilities and prepotent response tendencies are controlled for. In addition, this new version of the task allows the assessment of the accuracy of gaze following. Marmosets precisely extrapolate gaze direction, past distracting objects and from considerable distances, thereby meeting the criteria of so-called geometrical gaze following. The presence of this ability in common marmosets suggests that higher forms of gaze following might be more widely distributed among nonhuman primates than previously thought. Keywords: gaze following, object choice task, nonhuman primate, Callithrix jacchus The ability to follow the gaze of a conspecific has obvious adaptive value because it allows an individual to direct its own focus of attention to those stimuli in its environment that have attracted the attention of other individuals of its group. Although benefits from gaze following accrue even to very simple mecha- nisms such as reflex-like coorientation, more flexible, cognitive forms of this ability enhance the amount of information that can be acquired. Moreover, gaze-following abilities also serve as impor- tant precursors to other forms of social-cognitive abilities that in their fully fledged form are exclusively human (Baron-Cohen, 1995; Calder et al., 2002; Charman et al., 2000). As a conse- quence, the investigation of gaze following is of special interest for developmental as well as comparative psychology and has pro- duced a large number of findings over the last several decades. Human newborns already show a special interest in eyes (Batki, Baron-Cohen, Wheelwright, Connellan, & Ahluwalia, 2000; Ca- ron, Caron, Caldwell, & Weiss, 1973; Maurer & Salapatek, 1976). First reports of simple gaze following, that is, “looking where someone else is looking” (Butterworth, 1991), stem from 2- to 4-month-olds (D’Entremont, Hains, & Muir, 1997; Scaife & Bruner, 1975). Under facilitating conditions, by 6 months of age, children become able to localize the target object of the gaze of the caregiver if the target object is nearby and no distractor objects in the immediate visual field are present (Butterworth & Cochran, 1980; Butterworth & Grover, 1989; Butterworth & Jarrett, 1991; Morales, Mundy, & Rojas, 1998). However, more reliable gaze following to a specific object, past distractor objects, usually only develops later and is manifest between 9 and 12 months (Carpen- ter, Nagell, & Tomasello, 1998; Corkum & Moore, 1995, 1998; Morissette, Ricard, & Gouin-De ´carie, 1995). Butterworth (1995) proposed three mechanisms of gaze follow- ing to explain its developmental pattern. The ecological mecha- nism is thought to operate in children younger than 1 year. When this mechanism is at work, the infant’s visual search is initiated or triggered by the adult’s gaze but is then governed by interesting objects in the environment. As a consequence, the child follows the gaze only to targets within its immediate visual field and ceases visual search as soon as any interesting object is encountered. The more advanced geometric mechanism is thought to emerge around the 1st birthday and allows infants to extrapolate an adult’s gaze vector more accurately. They will ignore closer distracting objects and try to identify the specific target of the gaze but still search only within their visual field. The representational mechanism finally results in nonegocentric knowledge, namely that two indi- viduals can see different things in different regions of space. Even if the transition from nonrepresentational mechanisms to represen- tational mechanisms might be more fine grained (e.g., Flavell, 1999), and the exact age of the described transitions is still being specified (e.g., Moll & Tomasello, 2004), Butterworth’s model of the development of gaze following provides a powerful framework for studying gaze-following abilities not only in human infants but in nonhuman primates as well. In nonhuman primates, sensitivity to gaze is obvious in species showing gaze aversion. Staring eyes are a signal for immediate danger (predator) for many species, including fish and iguanas (Emery, 2000). However, gaze aversion responses merely require the ability to discriminate between a directed and an averted gaze, with a directed gaze causing arousal in the receiver; they do not require any understanding of gaze direction. Studies on the under- standing of gaze direction typically investigate whether an indi- vidual visually coorients with another individual (simple gaze following) and whether it is able to use gaze as a cue within an Judith Burkart, Anthropological Institute, University of Zurich, Zurich, Switzerland; Adolf Heschl, Konrad Lorenz Institute for Evolution and Cognition Research, Altenberg/Vienna, Austria. Adolf Heschl is now at the Institute of Zoology, Karl Franzens Univer- sity, Graz, Austria. This research was supported by a grant of the Konrad Lorenz Institute for Evolution and Cognition Research, Altenberg, Austria, to Judith Burkart, and portions of this work were part of her dissertation. We thank Elsa Addessi for insightful comments. Correspondence concerning this article should be addressed to Judith Burkart, Anthropological Institute, University of Zurich, Winter- thurerstrasse 190, 8057, Zurich/CH, Switzerland. E-mail: judith.burkart@ aim.unizh.ch Journal of Comparative Psychology Copyright 2006 by the American Psychological Association 2006, Vol. 120, No. 2, 120 –130 0735-7036/06/$12.00 DOI: 10.1037/0735-7036.120.2.120 120