The Origin of Representational Drawing: a Comparison of Human Children and Chimpanzees

Child Development, xxxx 2014, Volume 00, Number 0, Pages 1–15

The Origin of Representational Drawing: A Comparison of Human Children and Chimpanzees

Aya Saito Misato Hayashi Chubu Gakuin University Kyoto University

Hideko Takeshita Tetsuro Matsuzawa The University of Shiga Prefecture Kyoto University

To examine the evolutional origin of representational drawing, two experiments directly compared the drawing behavior of human children and chimpanzees. The first experiment observed free drawing after model presentation, using imitation task. From longitudinal observation of humans (N = 32, 11–31 months), the developmental process of drawing until the emergence of shape imitation was clarified. Adult chimpanzees showed the ability to trace a model, which was difficult for humans who had just started imitation. The second experiment, free drawing on incomplete facial stimuli, revealed the remarkable difference between two species. Humans (N = 57, 6–38 months) tend to complete the missing parts even with immature motor control, whereas chimpanzees never completed the missing parts and instead marked the existing parts or traced the outlines. Cognitive characteristics may affect the emergence of representational drawings.

The oldest representational drawings in existence are examining the drawing behavior of chimpanzees the upper Paleolithic cave drawings of Homo sapiens, (Pan troglodytes), humans’ closest living relatives. who drew animals with a variety of materials and Chimpanzees and humans share about 98.8% of the reﬁned techniques (Beltran, 2000; Chauvet, Des- genome (Chimpanzee Sequencing and Analysis champs, & Hillaire, 1996). A recent study using Consortium, 2005) and share a common ancestor uranium-thorium dating methods estimated that that existed until about 6 Ma. Chimpanzees show some of these drawings are more than 40,000 years marked similarities with humans in some aspect of old (Pike et al., 2012). Since that time, humans have tool using and social behavior. By comparing created art by drawing or painting in every period in behavior between the two species, we can infer every culture. Archeological ﬁndings, such as older cognitive traits shared with our common engraved pieces of ochre and shell beads, indicate that ancestor, and differences, the divergent newer our use of symbols emerged at least 100,000 years ago traits, achieved separately by humans and chimpan- (Henshilwood, d’Errico, & Watts, 2009; Henshilwood zees following evolutionary separation. Although et al., 2011). Thus, it is reasonable to suggest that there are no reports of drawing behavior in wild humans had the cognitive capacity for producing chimpanzees, it is common for captive chimpanzees representational drawing when Homo sapiens emerged to learn to draw or paint by manipulating a pen or in Africa about 200,000 years ago. At the very least, a brush on paper. In the early attempts at studying it is likely that they had this capacity when they chimpanzee drawing, Kellogg and Kellogg (1933) spread out of Africa approximately 100,000 years ago. and Ladygina-Kohts (1935/2002) individually cared However, the underlying cognitive mechanisms for for chimpanzees along with their human children drawing behavior are yet unknown. and compared the two species’ development of many behaviors, including drawing. Kellogg and Kellogg reported that a chimpanzee scribbled after Drawing Behavior in Chimpanzees observing a model drawing, but she did not imitate The present study aimed to assess the cognitive the human’s drawing. In contrast, human children capacity that led humans to begin drawing by

Correspondence concerning this article should be addressed to © 2014 The Authors Aya Saito, Faculty of Child Studies, Chubu Gakuin University, Child Development © 2014 Society for Research in Child Development, Inc. 30-1 Nakaoida-cho, Kakamigahara, Gifu, 504-0837, Japan. Elec- All rights reserved. 0009-3920/2014/xxxx-xxxx tronic mail may be sent to [email protected]. DOI: 10.1111/cdev.12319 2 Saito, Hayashi, Takeshita, and Matsuzawa preferred imitation. Ladygina-Kohts described stages Premack (1975) verified a likely cognitive factor. In of scribbling in a chimpanzee, but not representa- a composition task with fractured facial photo stim- tional drawings. uli that reduced motor demands, only a 12-year-old The first systematic study on chimpanzee repre- chimpanzee named Sarah succeeded in completing sentational drawing was conducted by Schiller an accurate configuration of the face; 3 other partic- (1951), who presented geometric figures to a chim- ipants did not succeed. panzee, Alpha. Alpha changed her scribbling pat- The ability to manipulate tools by relating one tern depending on the stimuli. For example, she object to another develops in chimpanzees just as it marked on the relatively large figures drawn in the does in human children (Hayashi & Matsuzawa, center of the sheet and scribbled in blank space 2003; Takeshita, 2001). Generally, as chimpanzees when presented with relatively small figures drawn gain experience drawing, they become better at in the periphery. She scribbled on the fractured controlling their pens and are able to produce a piece of a Pacman-like figure or arranged circles. variety of smooth lines (Gardner & Gardner, 1978; Schiller discussed those responses as balancing the Kellogg & Kellogg, 1933; Ladygina-Kohts, 1935/ composition, ascertaining that Alpha was capable 2002; Morris, 1962; Tanaka et al., 2003). However, of intuiting a human-like sense of order. Schiller’s most previous reports concern only infant or juve- studies were followed by Morris (1962), Smith nile chimpanzees as participants, as it is difficult to (1973), and later, Boysen, Berntson, and Prentice control adult chimpanzees safely in face-to-face (1987). Like Schiller, these researchers reported that situations in order to conduct standardized experi- their chimpanzees marked on the figures or scrib- ments. In this study, we conducted experiments bled on blank space; however, they did not observe with two juveniles and four adult chimpanzees, all the balancing behavior observed by Schiller, and of whom had considerable experience participating thus it appears problematic to claim that chimpan- in face-to-face experiments due to their long-term zees possess a sense of order akin to what is likely relationships with a human tester (Matsuzawa, the origin of human aesthetic sense (Lenain, 1995, 2009). 1997). In order to test our hypotheses with chimpanzee In most cases, drawing occurs spontaneously, participants who had considerable drawing experi- that is, without food rewards or special training, ence, we devised two experiments: free drawing and apes will draw or paint as self-gratifying play after presentation of a model (Experiment 1) and (Boysen et al., 1987; Lenain, 1997; Matsuzawa, 1995; free drawing on illustrations of a chimpanzee face Morris, 1962; Schiller, 1951; Smith, 1973; Tanaka, (Experiment 2). In general, after human children Tomonaga, & Matsuzawa, 2003). For this reason, begin scribbling at around 1 year of age, their scrib- drawing opportunities are sometimes proposed as bling develops from accidental markings to con- environmental enrichment for great apes in captiv- trolled lines, and it gains variation as their motor ity. Despite decades of such experiences, however, skills develop. They finally begin to draw represen- chimpanzees’ drawings have consistently been tational figures when they are around 3 years of regarded as scribbles, without clear evidence for age (Cox, 1992; Saito, Hayashi, Ueno, & Takeshita, representational figures. Gardner and Gardner 2011). Although many studies have been conducted (1978) reported that chimpanzee Moja, who learned on child drawing, the majority of systematic studies American Sign Language (ASL), signed on her focused on representational drawing of children drawings “bird” when they asked what it was. This older than 3 or 4 years of age (e.g., Arnheim, 1954; kind of labeling was also observed in other ASL- Freeman, 1972; Golomb, 1973; Goodnow, 1977). The trained chimpanzees or gorillas (Patterson, 1986). studies conducted on scribbling stages have been Nevertheless, to human eyes their works appeared limited to longitudinal observation studies of one to be no more than scribbles and it is difficult to or a few children (Eng, 1954; Luquet, 1927) or a recognize what, if anything, was represented on the cross-sectional study by collecting the drawn fig- page. ures of a large number of children (Kellog, 1969). Why is it that chimpanzees do not draw repre- We carried out the same experiments in chimpan- sentational figures? Certain mechanisms must zees and human children to establish a comparative underlie the human capacity for representational scale of development. Based on the results of the drawing. For example, we must have motor skills two experiments, we discussed potential explana- to control the lines, some cognitive function to tions for the scale of development and to thus translate perception into action, as well as the moti- explore the cognitive basis of representational vation to pick up a writing utensil and draw. drawing. The Origin of Representational Drawing 3

Experiment 1 Table 1 Chimpanzee Participants in Experiments 1 and 2 In order to evaluate the motor skills necessary for figure drawing, a prerequisite for the emergence of Name Age in years Sex representational drawing, we used imitation- of-model drawing. Although the imitation task is a Akira 29 Adult Male Ai 28 Adult Female commonly used developmental assessment tool, Popo 23 Adult Female there is a lack of research on the developmental Pan 21 Adult Female processes underlying the change from simple scrib- Ayumu 5 Juvenile Male bling to successful imitation. Previous studies of Pal 5 Juvenile Female drawing on premarked stimuli indicated that many chimpanzees change their scribbling position to mark the stimuli or blank space (Boysen et al., childhood (Hayashi & Matsuzawa, 2003; Hayashi 1987; Morris, 1962; Schiller, 1951; Smith, 1973). et al., 2009; Matsuzawa, 2009), and all had some Matsuzawa (1990, 2000) reported that an adult degree of drawing experience before this study chimpanzee named Chloe spontaneously traced a (Hayashi & Matsuzawa, 2003; Tanaka et al., 2003). drawn circle. In this study, we modeled the draw- Two mother–infant pairs (Ai and Ayumu, Pan and ing of simple figures and observed the drawing Pal) had also participated in a free-drawing task behavior of participants on the same sheets. The using a touch-screen computer (Tanaka et al., 2003). materials and procedures were kept the same for Prior to the study, all chimpanzee participants prac- chimpanzees and humans in order to directly com- ticed free drawing on blank paper to allow observa- pare their motor-control ability. The longitudinal tion of their freestyle drawing and to familiarize observation of human children permitted analysis them with the materials and the experiment face-to- of not only when children succeeded in imitating, face situation. but also how their scribbles developed until successful imitation emerged. Procedure We used an imitation-of-model drawing, which Method was standardized as a developmental scale for human children (Kyoto Scale of Psychological Participants Development; Ikuzawa, Matsushita, & Nakase, Human participants were 32 Japanese children 1985) with a simplified procedure. The experiments (13 boys and 19 girls) who ranged in age from were carried out individually and face-to-face for 11 months to 31 months at the time of the first test- both human and chimpanzee participants. Children ing session. They were members of the Umikaze were also tested in other object-manipulation tasks Infant Laboratory at the University of Shiga Prefec- on the same day (Hayashi, 2007; Hayashi & Takesh- ture, Japan, recruited from the surrounding area. ita, 2009; Hayashi et al., 2009). Our experiments We tested each participant once every 2–3 months with chimpanzees were conducted directly after the from September 2005 to April 2009, although the computer experiments at the same experimental entry date and the duration varied for each child. booth. Human children sat beside their parent or on All experiments complied with the laws of Japan their lap while a tester sat across from them at a and were approved by the Human Ethics Commit- low table in a quiet room at the laboratory. Chim- tees of the Primate Research Institute of Kyoto Uni- panzee participants sat on a wooden board on the versity and the University of Shiga Prefecture. floor in front of a human tester in the booth. The Informed consent was obtained from the parents of procedure was nearly identical for human and all participants. chimpanzee participants and used the same materi- Chimpanzee participants were 4 adult and 2 als, namely, B4-sized paper (257 mm 9 364 mm) juvenile chimpanzees (Table 1), living in a group of and water paint markers. A session started with a 14 chimpanzees at the Primate Research Institute of free drawing trial, which allowed the participant to Kyoto University, Aichi Prefecture, Japan. All had draw freely on a blank paper. The tester said, “Let’s previously participated in cognitive experiments draw on this,” and observed the participant’s draw- using a touch-screen computer. The chimpanzees ing behavior for at least 1 min. We used the first were familiar with the experimental setting and five figures from the Kyoto Scale of Psychological with face-to-face situations with human testers since Development (Ikuzawa et al., 1985) for a session: (1) 4 Saito, Hayashi, Takeshita, and Matsuzawa horizontal lines, (2) vertical lines, (3) a circle, (4) a figure: moved scribbles to mark the model figure; cross, and (5) a square. In each trial, the tester drew (c) similar lines: change in scribbling touches a model figure with a pale orange marker in front of depending on the model figure (e.g., horizontal up the participants while saying, “Can you draw like and down strokes increased after the presentation this?” Then, the participant drew on the same paper of horizontal lines or spiral scribbles increased after with a marker of a different color so as to identify a circle presentation); (d) imperfect imitation: the lines in later analysis. If the participant did not attempted to draw the model figure but the criteria succeed in imitating within 30 s, the tester demon- for success were not met; and (e) traced the model strated the model drawing again by tracing the lines: traced a part of the model lines. From the model figure and then observed for another 30 s or video recordings, a main rater (A. S.) checked more. The drawing behavior of the participants was whether each categorized behavior was observed in recorded using two digital video cameras set at dif- a trial. To assess interrater reliability, an additional ferent angles. rater watched video recordings of 214 of 355 trials In a given experimental day, human children in humans and 87 of 87 trials in chimpanzees and participated in one session consisting of a free checked the participants’ behavior. The main rater drawing trial and two to five model figure trials, and the second rater agreed on each categorized depending on their concentration and their previ- behavior on 97.8% of the trials. From the longitudi- ous success in imitation. For very young children nal data of individual children, we considered the who could not imitate the first figures and had dif- first occurrence of each of the four categorized ficulty in keeping concentration for many trials, we behaviors and the first success of imitation and randomly chose two figures, one from line figures compared the occurrence ages by a one-way (1) or (2), and one from geometric figures (3) or (5) repeated measures of analysis of variance (ANOVA). in order to counterbalance comparisons of their The frequencies of the categorized behaviors were scribbling patterns with chimpanzees. calculated in each of five age ranges in humans Chimpanzees participated in three sessions at (11 months to 1 year 5 months: 65 trials analyzed; intervals of 1½ months. We used fruit to maintain 1 year 6 months to 1 year 11 months: 96 trials; their motivation but not to reward any specific 2 years to 2 years 5 months: 72 trials; 2 years drawing behavior. 6 months to 2 years 11 months: 57 trials; 3 years: 65 trials) and two groups of chimpanzees (juveniles: 27 trials; adults: 60 trials). Data Analysis We analyzed data from 679 trials from 286 Results sessions with human infants, and 87 trials from 19 sessions with chimpanzees. The mean number of From longitudinal human observations, the aver- sessions for a child was 9.0 (SD = 4.7) during 31.7 age age of the first successful imitation is shown in (SD = 13.6) months of observation period with 3.2- Table 2. On average, human children succeeded in (SD = 0.8) month intervals. The age of the first suc- imitating horizontal lines at 2 years 4 months, verti- cessful imitation was defined as the age at which a cal lines at 2 years 6 months, a circle at 2 years child first imitated a figure that he or she had not 11 months, a cross at 3 years 5 months, and a previously imitated in longitudinal observations. square at 4 years. A one-way repeated measures This analysis did not include cases in which chil- ANOVA showed main effects of figure, F(4, 80) = fi < 2 ¼ : – ’ dren succeeded on the very rst trial. The success 36.2, p .001, gp 64. Post hoc Tukey Kramer s of an imitation was evaluated by using the criteria honestly significant difference (HSD) comparisons of the Kyoto Scale of Psychological Development between the factors revealed a significant difference (Ikuzawa et al., 1985; see Appendix S1 in the online between all figures except for horizontal and verti- Supporting Information). cal lines. Table 2 also shows the average age of the Next, we used the failure trials before the first first occurrence of each categorized behavior before successful imitation of the figure for humans and their first success in imitation. They marked the compared the drawing patterns with those of the model figure at 1 year 5 months, drew similar lines chimpanzees. Typical behavior patterns were identi- at 1 year 10 months, drew imperfect imitations at fied by focusing on changes in position and scrib- 2 years 5 months, and traced the model lines at bling touches compared to the first free drawing 2 years 8 months, on average. A one-way repeated trial: (a) scribbled randomly: scribbled without clear measures ANOVA revealed a main effect for change in position and touch; (b) marked the categorized behaviors, F(3, 71) = 22.8, p < .001, The Origin of Representational Drawing 5

Table 2 The Average Ages at the First Occurence of Each Typical Behavior and the First Successful Imitation in Individual Humans

Note. N indicates the number of participants who showed each typical behavior during the longitudinal observation. *p < 0.05 (Tukey–Kramer post hoc multiple comparison test).

2 ¼ : – ’ fi gp 49. Post hoc Tukey Kramer s HSD compari- sentation. These ndings resembled the responses sons between the categorized behaviors revealed of chimpanzees in former studies (Morris, 1962; significant differences between (b) mark the figure Schiller, 1951; Smith, 1973). Model presentation, versus (c) similar lines (p < .05), (c) similar lines however, seemed to enhance reactivity, a finding versus (d) imperfect imitation (p < .05), and (c) that was different from previous studies presenting similar lines versus (e) trace the model lines premarked shapes. Moreover, some adults showed (p < .05). No significant difference was found the ability to control their lines to trace the model. between (d) imperfect imitation versus (e) trace Tracing behavior is not extraordinary for chimpan- the model lines. Examples of the products by cate- zees, as Matsuzawa (1990) has previously reported gorized behaviors and the frequency in each age that a chimpanzee named Chloe spontaneously period for humans and chimpanzees are shown in traced a drawn circle. It is also possible to teach Figure 1. Juvenile chimpanzees scribbled randomly chimpanzees to trace a line with a finger on a (40.7% of trials) or marked the figure (59.3%), touch-screen monitor through the use of reinforce- whereas adult chimpanzees also marked the figure ment (Iversen & Matsuzawa, 1996). Our results (36.7%) and three of four adult chimpanzees drew indicated that tracing was even difficult for human similar lines (15.0%) or traced lines (11.7%). How- children who had just started imitating the easiest ever, none of the chimpanzees attempted to imi- figures. Changing their lines and tracing was only tate the figure. observed in adult chimpanzees but not in juvenile chimpanzees, which demonstrated that the adults had matured motor skills requisite for controlling Discussion their strokes (Figure 2). Through a longitudinal analysis, we clarified the These results indicated that chimpanzees develop human developmental trajectory for scribbling pat- drawing skills and at least three adults had the abil- terns before the emergence of imitative figure draw- ity to finely control their manual movements when ing. First, children moved their scribbles to mark drawing lines to trace model lines. Thus, the the model. Second, they tried to change their scrib- observed lack of representational drawing in chim- bling touches similar to the model’s movement, panzees does not appear to be due to inadequate and third, they intended to imitate figures, which motor ability. ultimately led to success upon maturation of In order to copy the figures, not only are motor motor skills. Although chimpanzees did not draw a skills required, but also the cognitive capacity to distinct figure by imitating the model presentation, compare the shape of one’s drawing to the model they changed the position of their scribbles to mark figure. However, in this experiment, we argue only the model or marked a blank space and, further, for motor ability as we assumed that this task was changed their scribbling behavior after model pre- not imitative model drawing but a task of free 6 Saito, Hayashi, Takeshita, and Matsuzawa

Figure 1. Examples of the products of the main categorized behaviors and their frequency in each age period for humans and chimpanzees in Experiment 1. The Origin of Representational Drawing 7

Figure 2. A tester drew a circle as a model presentation (left), and chimpanzee Pan traced the circle (right). drawing after model presentation. Despite receiv- was not caused by a lack of motor control. Experi- ing the same ambiguous verbal instructions, it is ment 2 was designed to investigate the underlying possible that the two species differed in the degree cognitive mechanism, another prerequisite for the of understanding for the task objective to imitate emergence of representational drawing. Human the figure. Therefore, we could not conclude that children in the early stages of representational chimpanzees are unable to imitate the shape of a drawing will often draw faces of humans or model. In fact, some of the adult chimpanzees animals, and it is easy for others to objectively per- spontaneously changed their scribbling pattern to ceive what is represented. To assess the representa- shape similar lines seemingly in an effort to imi- tional ability of children in scribbling stage, some tate the model’s movement. For instance, chimpan- studies have used a design in which an incomplete zee Pan, who ordinarily drew short vertical lines, figure is presented, and the child is asked to com- suddenly drew a long horizontal line during a trial plete it. It has been demonstrated that children who of horizontal lines, and she successfully traced cannot yet draw representational figures by them- long vertical lines. Iversen and Matsuzawa (1997) selves fill in some missing parts inside the contours taught chimpanzees to draw a straight line parallel of illustrated figures (Freeman, 1977; Yamagata, to a presented line using their finger on a monitor. 2001). A study in which children were allowed to Success occurred only when a starting point was scribble on picture books showed that even 1- or 2- provided as a guide on the monitor, and it year-olds who were still in the scribbling stage required a great deal of trial and error. It is note- often marked on human or animal figures, particu- worthy that chimpanzees were not proficient at larly on their faces (Yamagata, 1991). It was also imitating a human’s behavior, particularly when reported that adult chimpanzee Ai marked human the actions were not directed toward another and animal figures that appeared in picture books object or their own body (Myowa-Yamakoshi & (Matsuzawa, 1995). We used incomplete figures to Matsuzawa, 1999). It is likely that imitating a directly compare representation ability of two model figure constitutes an advanced level of imi- species. tation that requires not only simply directing a In order to determine whether the chimpanzees pen toward the paper but also adequately con- would fill in the missing parts, we prepared an trolled manual movements on the paper. In this illustrated figure of a chimpanzee’s face and deleted case, chimpanzees must perceive the relation facial parts to make an incomplete-face stimulus. between their own manual movements and the As the marking behavior on picture books indicated manifest results on the drawing. (Matsuzawa, 1995), chimpanzees did seem to recognize illustrated figures. Chimpanzee Ai even recognized familiar chimpanzees and humans portrayed in line drawings and matched them with the letter Experiment 2 of the alphabet that corresponded to the individ- Results from Experiment 1 showed that adult chim- ual’s name (Itakura, 1994). In the present experi- panzees had sufficient motor skills to control their ment, chimpanzees’ spontaneous drawing behavior lines as required to trace a model line. Thus, the on the incomplete-face stimulus was observed and absence of representational drawing in chimpanzees subsequently compared with that of human chil- 8 Saito, Hayashi, Takeshita, and Matsuzawa dren of different ages. We also analyzed children’s Participants were not instructed on what to draw in comments on stimuli and the drawings that regard to the absence of facial parts. We observed occurred with each categorized behavior in order to as the participants drew on the paper for 30 s or investigate the relation between their recognition more from their first mark. We did not designate and representation. the “completion” as a correct reaction in this task and the tester did not provide feedback by vocal reward. We used fruit to maintain the motivation Method of chimpanzees but did not reward any specific drawing behavior. The drawing behavior of the Participants participants was recorded using two digital video Fifty-seven human participants (23 boys and 34 cameras set at the different view angles. girls) ranging in age from 1 year 6 months to 3 years 2 months participated in Experiment 2. The Data Analysis children were members of “Umikaze” Infant Labo- ratory of the University of Shiga Prefecture and all We analyzed 285 trials from 57 sessions with of them already participated in Experiment 1. The humans and 60 trials from 12 sessions with chimpan- experiments were conducted with other object- zees. We identified five typical patterns of behavior manipulation tasks on the same day (Hayashi, on imperfect Stimuli 2 through 5: (a) marked the 2007; Hayashi & Takeshita, 2009). The children took whole face: scribbling mainly inside the whole face this drawing test 2 or 3 times at intervals of more but not limited to specific parts; (b) scribbled on than 3 months. The chimpanzee participants were blank space: scribbling outside the face area; (c) the same as in Experiment 1 (Table 1). All experi- marked the present parts: scribbling to mark the ments complied with the laws of Japan and were drawn facial parts; (d) completed the missing parts: approved by the Human Ethics Committees of the draw eyes, nose, or mouth to complete the face; and Primate Research Institute of Kyoto University and (e) traced the outlines: tracing on the contours of the the University of Shiga Prefecture. Informed con- face. A main rater (A. S.) determined whether each sent was obtained from the parents of human par- categorized behavior was observed in a trial from ticipants, and the data were treated carefully to the video recordings. To assess interrater reliability, protect their privacy. an additional rater watched video recordings in 190 of 285 trials in humans and 30 of 60 trials in chimpanzees and checked the participants’ behavior. Apparatus The main rater and the second rater agreed with B4-sized (257 mm 9 364 mm) paper and paint the rating of each categorized behavior in 97.1% markers were used for drawing. We prepared an of the trials, indicating strong interrater reliabil- illustration of a chimpanzee face outline from a ity. The frequencies for categorized behaviors were photograph using Adobe Photoshop. The illustra- calculated in each of five human age groups (1 year tions were printed in 2-mm-wide gray lines, which 6 months to 1 year 9 months, 1 year 10 months to were of similar width to the marker lines and dis- 2 years 1 month, 2 years 2 months to 2 years tinguishable from the black lines of participants in 5 months, 2 years 6 months to 2 years 9 months, later analysis. Facial features were deleted to con- and 2 years 10 months to 3 years 2 months) and two struct the following stimuli: (1) normal face, (2) groups (juveniles and adults) in chimpanzees right eye missing, (3) left eye missing, (4) both eyes (Table 3). We compared frequencies using a 1-df missing, and (5) outline only. A session consisted Cochran–Armitage trend test. Additionally, we ana- of five trials, with the stimuli presented in that lyzed the spontaneous speech of human children. order. We checked the presence or absence of comments in regard to missing parts of stimuli before they start drawing such as “no eye” or “eyes missing” and Procedure comments about their drawn parts such as “eye” or The experimental setting was identical to Experi- “mouth” or about their drawn face such as “face” or ment 1, and the procedure was again carried out in “monkey” during or after drawing. To assess the individual face-to-face interactions for both human relation between the verbalization and the drawing and chimpanzee participants. behaviors, we recategorized each trial into one of the A tester passed a black pen to a participant and four independent behavioral categories based on the said, “Let’s draw on this, just as freely as you like.” best performance in each trial: (a) completion; (b) The Origin of Representational Drawing 9

Table 3 Age Groups in Humans and Chimpanzees, Experiment 2

Number of Number of Number

Age groups subjects sessions of trials Mage (SD)

Humans 1 year 6 month to 1 year 9 month 10 10 55 1 year 8 month (1 month) 1 year 10 month to 2 year 1 month 8 8 45 2 year 0 month (1 month) 2 year 2 month to 2 year 5 month 17 17 90 2 year 4 month (1 month) 2 year 6 month to 2 year 9 month 10 10 50 2 year 8 month (1 month) 2 year 10 month to 3 year 2 month 8 8 45 3 year 0 month (1 month) Chimpanzees Juveniles 2 4 20 6 year (0 year) Adults 4 8 40 27 year (3 year)

imperfect completion, for example, drawing too parts of faces. In most human cases, tracing occurred many eyes or indistinct eyes; (c) marking parts; and not independently (except for three cases by a child), (d) marking face. Then, we compared the mean age but simultaneously with other categorized behav- and the frequency of spontaneous verbalization of iors, namely, completion (30.6% of trials), marking children among the four behavioral categories of tri- parts (58.3%), and marking whole face (55.6%). als. Statistical significance was evaluated by an Therefore, we identified four main phases of ANOVA for age followed by Tukey–Kramer HSD development in human children. First, marking analysis and by Cochran–Mantel–Haenszel test fol- within the facial outline; second, marking on exist- lowed by residual analysis for the frequency of each ing parts; third, filling in the missing parts but type of verbalization. imperfect; and fourth, completing the missing parts. Stated otherwise, the marking of existing parts gradually converges on distinct facial parts from Results and Discussion the whole face before the emergence of missing part The frequency of each categorized behavior on completion. We selected trials that contained these missing parts stimuli is presented by age group and types of behavior and recategorized them into the species in Figure 3. The most frequently observed four phases independently based on the best perfor- behavior in younger human groups was “mark the mance in each trial. That is, if a trial was catego- whole face.” The frequency of “mark the present rized in one phase, the trial could not be placed in parts” increased with age in humans (v2 = 12.9, another phase. The mean age and the frequency of p<.01 by 1-df Cochran–Armitage trend test). “Com- children’s spontaneous speech in four behavioral plete the missing parts” increased with age categories of trials with the data for “tracing (v2 = 48.8, p<.001) and was the most frequently outlines” as a reference are shown in Table 4. A observed behavior in the two groups of humans one-way repeated measures ANOVA revealed a aged 2 years 6 months or older. On the contrary, significant difference in ages among four catego- = < 2 ¼ : none of the chimpanzees completed the missing rized groups, F(3, 166) 21.2, p .0001, gp 08. parts; instead, they marked the whole face (90.0% of Post hoc Tukey–Kramer’s HSD comparisons the trials in juveniles, and 27.5% of the trials in showed the significant age difference between adults), scribbled on blank space (0% in juveniles, “completion” versus “marking parts” or “marking and 42.5% in adults), or marked the existing part face,”“failure completion” versus “marking face.” (25.0% in juveniles, and 30.0% in adults). Moreover, A Cochran–Mantel–Haenszel test revealed that the adult chimpanzees traced the outlines of the face in frequency of verbalization before drawing differed 22.5% of trials, a behavior that increased with the by categorized group of trials, v2(3) = 23.4, age in humans, particularly after 2 years 6 months p<.0001, Cramer’s V = 0.35, and verbalization (v2 = 19.9, p<.001). Chimpanzees’ motor skills were during or after drawing by categorized group, more refined in their marking of existing parts and v2(6) = 6.4, p = .039, Cramer’s V = 0.17. Residual marking the existing outline but not, however, in analysis showed that reference to missing parts was completing the missing parts. Conversely, human more frequent in “completion” and “failure comple- children demonstrated an ability to fill in missing tion” groups and less frequent in “marking parts” 10 Saito, Hayashi, Takeshita, and Matsuzawa

Figure 3. Examples of the products of main categorized behaviors and their percentage by different human and chimpanzee age groups (the number of trials with the behavior/total trials). The Origin of Representational Drawing 11

Table 4 Percentages of Spontaneous Referral to Missing Parts Before Drawing and Comments on Drawn Parts or Drawn face During or After Drawing on Stimuli 2 Through 5

Spontaneous verbalization

Age During or after drawing Before drawing Type of behavior N M (SD) Range Missing parts (%) Drawn parts (%) Drawn face (%)

(a) Complete the missing 69 2 year 8 month 2 year 0 month to 55.1** 29.0** 10.1 parts (4 month) 3 year 2 month (b) Fill in the missing parts 34 2 year 7 month 1 year 7 month to 55.9** 23.5 5.9 but imperfect (4 month) 3 year 0 month (c) Mark the existing parts 26 2 year 4 month 1 year 6 month to 11.5* 15.4 0.0 (5 month) 3 year 1 month (d) Mark the whole face 64 2 year 2 month 1 year 7 month to 3.1** 1.6** 10.9 (4 month) 3 year 2 month Trace the outlinesa 20 2 year 8 month 2 year 0 month to 50.0 30.0 10.0 (4 month) 3 year 1 month

Note. N indicates the number of trials categorized into the four types of behavior independently. aWe excluded this category of data from the statistical analysis as it contains overlapping data with other categories. *p < .05. **p < 0.01 (residual analysis). and “marking face” groups. More than half of the than it was for humans under 2 years 9 months. In children mentioned missing parts before completing contrast, the frequency of the behavior “mark the the task, even if the end result was imperfect, indi- present parts” was lower for adult chimpanzees cating that they spontaneously tried to complete than for humans older than 2 years 6 months. the missing parts, in spite of their lack of motor skill. Verbal behavior during or after drawing was observed more frequently in “completion” and less General Discussion in “marking face.” Further, 15.4% of the children Cognitive Foundation for Representational Drawing (n = 4) who marked the existing parts mentioned drawn parts, while none of them mentioned the This study approached the emergent representa- face or object during or after drawing, suggesting tional drawing ability of young children and illus- that they recognized the drawn parts and perhaps trated their similarities and differences with intended to draw but did not notice the parts miss- chimpanzees by directly comparing the two species ing. On the other hand, in “marking face,” only in two experiments. The longitudinal study of 1.6% of the children (n = 1) mentioned drawn parts, Experiment 1 investigated and shed light on the while 10.9% (n = 7) mentioned the drawn face later, developmental trajectory of human drawing from indicating that they recognized the face but had not scribbling to successful copying of a shape. Chim- noticed the drawn or missing parts. panzees not only marked the figures, which is a Although chimpanzees were likely to recognize common response to these stimuli, but also traced the illustrated face, as former studies indicated, it is the lines in a similar manner to the human unclear whether they failed to recognize the incom- children in the early stages of imitation ability, plete figure as a face, recognized it but did not showing adequate motor-control skills. Adult chim- notice the absence of the parts, or noticed the panzees exhibited a greater array of drawing pat- absence of the parts but had no motivation to terns on stimuli than did juvenile chimpanzees, complete them. Further investigation is needed indicating more mature motor-control skills in the to address this issue, especially with respect to former. Experiment 2 demonstrated a remarkable chimpanzee’s symbolic capacity when processing difference between the two species. Humans drew the incompletely drawn figures. missing parts despite more limited motor control, However, chimpanzees might direct their atten- whereas adult chimpanzees marked only existing tion toward the outline rather than the target facial figures. These results indicated that the lack of rep- features, since the frequency of the behavior “trace resentational drawing in chimpanzees was not due the outlines” by the adult chimpanzees was higher to a motor deficit but derived from a cognitive 12 Saito, Hayashi, Takeshita, and Matsuzawa process necessary for drawing missing parts to parts to complete their images although the situa- complete an image. Although Premack (1975) tions were obviously not the same with our experi- reported that a chimpanzee named Sarah correctly mental settings. For example, the famous bison of configured facial parts, three other chimpanzees Altamira were drawn on swells of the dome, and failed. The task used here, namely completing a the contours of the bodies were sometimes conflu- face by filling in missing parts without assistance, ent with natural cracks on the rock. There is also a appears more complex than simply organizing “mask” with eyes filled in on the hanging parts of existing parts. rock that resembles a face silhouette. Humans have Figures on a paper may trigger imagination in a strong tendency to imagine novel configurations, humans that is not possible in chimpanzees. Per- even in ambiguous images such as a spot on the haps this is the faculty that assists humans in draw- wall or clouds in the sky (Gombrich, 1972; Guthrie, ing representational figures. In Experiment 2, some 1993). This cognitive trait may be a defining feature mothers whose children completed the missing distinguishing Homo sapiens from other ancestors parts explained that it was the first time their chil- who did not develop representational drawing. dren had drawn a representational figure. In addition, many children over 2.5 years old drew the Why Do We Have the Cognitive Trait of Imagination? representational figures inspired by the models observed in Experiment 1 (Figure 4). They may In our study, motorically capable chimpanzees have conceived an imaginary shape of objects from marked strictly on the small parts or lines, that is, a a composition of lines and completed it by adding localized area. This tendency might be related to the missing parts. During longitudinal observations chimpanzees’ lack of global processing in compari- of drawing, about 48% first-time representational son to humans, who have strong global processing figures (by 11 of 23 children) were observed in abilities (Fagot & Tomonaga, 1999). Humans also model drawing trials following the absence of any exhibit greater temporal integration accuracy than clear representation in earlier free-drawing trials do chimpanzees in the task of dynamic shape during the same session (Saito et al., 2011). Some perception under a slit-viewing condition (Imura & stimuli, including even simple abstract figures, can Tomonaga, 2013). Global processing may be related trigger imaginative representations in young chil- to imaginative recreations of abstract figures, as it dren, especially those in the transition period from leads to Gestalt perception and object recognition. scribbling to representational drawing. Imagination can be described as perceiving a Ancient cave art indicated that Paleolithic percept as “something” and categorizing lower humans also imagined animals and drew missing level visual information into the concept of “something” by associating it with a symbol otherwise represented in the mind. This symbolic cognitive system is further evident in the case of human language, and is indeed the premise behind human language, and humans tend to imagine something even in response to ambiguous figures (Humphrey, 1998). Many studies have indicated that chimpanzees have the ability to learn some symbols, such as

2 y 5 m Girl “Railroad” 2y7m Boy“Train” characters and numerals, and even to understand some sign language (e.g., Matsuzawa, 1985a, 1985b). Therefore, we could not conclude that chimpanzees never have imaginative capability only from the result of the present study. Some chimpanzees especially who experienced symbol training may have some primitive capability of imagination based on the primitive symbolical 2y8m Girl “Anpanman” 3y1m Boy “Bus, Starts!” capability, as one of the chimpanzees who engaged in language training correctly configured facial Figure 4. Human children sometimes drew spontaneous repre- parts in the simple task of organizing existing parts sentations inspired by the models in Experiment 1. The indicated ’ age, gender, and verbal explanation by the child. They used their in Premack s (1975) study. imaginations with the presented lines and completed their On the other hand, symbolic systems in humans images by adding some “missing” parts. are much prevalent and also reflected in representa- The Origin of Representational Drawing 13 tional drawings of children as the former studies early representational drawing evokes a great deal indicated. For example, Arnheim (1954) disputes of vocal communication between children and their the assumption that problems of form in young parents and others given the joint attention drawn children’s drawing can simply be decoded into to their figures. Besides, humans, unlike chimpan- problems of content, by discussing the meaning of zees, are socialized starting in infancy to use paper well-known phenomenons such as “transparency” for drawings, to view pictures, and other two- and “tadpole man.” As children’s representational dimensional imagery. Drawing enables the sharing drawings are very symbolic, as opposed to a copy of an inspired image in one’s mind, and it is likely of the real object (Luquet, 1927), they might directly that this is a strong motivation for making repre- reflect the development of knowledge while the sentational drawing in human children. children are expanding their conceptions of objects Human infants are brought up to socialize in as a symbol. A phenomenon known as orientation upright positions and are encouraged to engage in indifferent representation arises in the early repre- face-to-face communication with other people sentational period, where children draw a figure in (Takeshita, Myowa-Yamakoshi, & Hirata, 2009). We an inverted or horizontal orientation. This phenom- speculate that this abundance of social interaction enon can be induced by presenting stimulus figures since birth would enhance human children’s ability such as illustrations of ears of a cat in different to internalize others’ varied viewpoints, strongly orientations. Some younger children draw facial motivate them to share with others, and play an parts in a rotated orientation on rotated stimuli and important role in the development of imagination. in an upright orientation on upright stimuli. It This development of imagination should manifest seems that the younger children are indifferent to in the completion of missing parts in images. It is the orientation of the face on the plane to draw. believed that humans have evolved a great deal of Since they may know the relative order of the facial their characteristic complex behavior through cul- parts in the whole face, they do not show difficulty tural learning (Tomasello, 1999). Chimpanzees also drawing the rotated face in a given orientation. On have primitive social-learning ability in the acquisi- the other hand, older children always intended to tion of tool-using skill, such as “education by mas- draw facial parts in an upright orientation: They ter-apprenticeship” (Matsuzawa et al., 2001) or reorient the sheet into the upright position before bonding- and identification-based observational they start drawing. These age differences in reaction learning (De Waal, 2001). However, enhanced cul- to inverted stimulus figures indicate a relation tural learning in humans may have played an between the production and development of the important role in more than 30,000 years of art his- facial symbol (Saito et al., 2011). tory among Homo sapiens and, further, seems to be In Experiment 2, human children marked on one of the primary drivers of the emergence of the presented figures, similar to chimpanzees, representational drawing. before they started to complete the missing parts. However, when they noticed the absence of certain features, they embraced the challenge to complete them, despite their lack of motor skills in References comparison to chimpanzees. It is noteworthy that Arnheim, R. (1954). Art and visual perception: A psychology the tester only instructed participants to “draw of the creative eye. Berkeley: University of California freely” and did not acknowledge “completion” as Press. a correct reaction. In addition, some children drew Beltran, A. (2000). The cave of Altamira (Japanese transla- independent representational figures on blank tion, Y. Ohtaka & M. Ogawa, Trans.). Tokyo, Japan: space. In contrast, chimpanzees never completed Iwanami Shoten. the missing features whereas more than half of Boysen, S. T., Berntson, G. G., & Prentice, J. (1987). Sim- ian scribbles: A reappraisal of drawing in the chimpan- the humans aged over 2.5 years did so spontane- zee (Pan troglodytes). Journal of Comparative Psychology, ously. As Freeman (1977), Matthews (1984), and 101,82–89. doi:10.1037/0735-7036.101.1.82 Yamagata (2001) pointed out, even young children Chauvet, J. M., Deschamps, E. B., & Hillaire, C. (1996). in the scribbling stage have the intention of repre- Chauvet cave: The discovery of the world’s oldest paintings. sentation. London, UK: Thames and Hudson. It must be noted that human participants likely Chimpanzee Sequencing and Analysis Consortium. practice drawing outside of the experimental con- (2005). Initial sequence of the chimpanzee genome and text far more often than chimpanzees. Such behav- comparison with the human genome. Nature, 437,69– ior might be encouraged by social motivation, as 87. doi:10.1038/nature04072 14 Saito, Hayashi, Takeshita, and Matsuzawa

Cox, M. V. (1992). Children’s drawings. London, UK: Pen- Ikuzawa, M., Matsushita, Y., & Nakase, A. (1985). Kyoto guin Books. Scale of Psychological Development [in Japanese]. Kyoto, de Waal, F. (2001). The ape and the sushi master: Cultural Japan: Nakanishiya. reflections of a primatologist. New York, NY: Basic Books. Imura, T., & Tomonaga, M. (2013). Differences between Eng, H. (1954). The psychology of children’s drawings (2nd chimpanzees and humans in visual temporal integra- ed.). London, UK: Routledge and Kegan Paul. tion. Scientific Reports, 3, 3256. doi:10.1038/srep03256 Fagot, J., & Tomonaga, M. (1999). Global-local processing Itakura, S. (1994). Recognition of line-drawing representa- in humans (Homo sapiens) and chimpanzees (Pan troglo- tions by a chimpanzee (Pan troglodytes). Journal of Gen- dytes): Use of a visual search task with compound stim- eral Psychology, 121, 189–197. doi:10.1080/00221309. uli. Journal of Comparative Psychology, 113,3–12. doi:10. 1994.9921195 1037/0735-7036.113.1.3 Iversen, I. H., & Matsuzawa, T. (1996). Visually guided Freeman, N. H. (1972). Process and product in children’s drawing in the chimpanzee (Pan troglodytes). Japanese Psy- drawing. Perception, 1, 123–140. doi:10.1068/p010123 chological Research, 38,126–135. doi:10.1111/j.1468-5884. Freeman, N. H. (1977). How young children try to plan 1996.tb00017.x drawings. In G. E. Butterworth (Ed.), The child’s represen- Iversen, I. H., & Matsuzawa, T. (1997). Model-guided line tation of the world (pp. 3–29). London, UK: Plenum Press. drawing in the chimpanzee (Pan troglodytes). Japanese Gardner, R. A., & Gardner, B. T. (1978). Comparative Psychological Research, 39, 154–181. doi:10.1111/ psychology and language acquisition. In K. Salzinger & 1468-5884.00051 F. Denmark (Eds.), Psychology: The state of the art. Kellog, R. (1969). Analyzing children’s art. Palo Alto, CA: Annals of the New York Academy of Sciences, 309,37–76. National Press Books. doi:10.1111/j.1749-6632.1978.tb29441.x Kellogg, W. N., & Kellogg, L. A. (1933). The ape and the Golomb, C. (1973). Children’s representation of the child: A study of environmental influence upon early behav- human figure: The effects of models, media, and ior. New York, NY: McGraw-Hill. instruction. Genetic Psychology Monographs, 87, 107–251. Ladygina-Kohts, N. N. (2002). Infant chimpanzee and Gombrich, E. H. (1972). Art and illusion: A study in the human child: A classic 1935 comparative study of ape emo- psychology of pictorial representation. London, UK: Phai- tions and intelligence. New York, NY: Oxford University don Press. Press. (Original work published 1935) Goodnow, J. J. (1977). Children’s drawing. Cambridge, Lenain, T. (1995). Ape-painting and the problem of the MA: Harvard University Press. origin of art. Human Evolution, 10, 205–215. doi:10. Guthrie, S. E. (1993). Faces in the clouds: A new theory of 1007/BF02438973 religion. New York, NY: Oxford University Press. Lenain, T. (1997). Monkey painting. London, UK: Reaktion Hayashi, M. (2007). A new notation system of object Books. manipulation in the nesting-cup task for chimpanzees Luquet, G. H. (1927). Le dessin enfantin. Paris, France: and humans. Cortex, 43, 308–318. doi:10.1007/ Librairie Felix Alcan. s10071-009-0273-5 Matsuzawa, T. (1985a). Color naming and classification in Hayashi, M., & Matsuzawa, T. (2003). Cognitive develop- a chimpanzee (Pan troglodytes). Journal of Human Evolu- ment in object manipulation by infant chimpanzees. Ani- tion, 14, 283–291. doi:10.1016/S0047-2484(85)80069-5 mal Cognition, 6, 225–233. doi:10.1007/s10071-003-0185-8 Matsuzawa, T. (1985b). Use of numbers by a chimpanzee. Hayashi, M., Sekine, S., Tanaka, M., & Takeshita, H. Nature, 315,57–59. doi:10.1038/315057a0 (2009). Copying a model stack of colored blocks by Matsuzawa, T. (1990). The world through the eyes of chim- chimpanzees and humans. Interaction Studies, 10, 130– panzees [in Japanese]. Tokyo, Japan: University of 149. doi:10.1075/is.10.2.03hay Tokyo Press. Hayashi, M., & Takeshita, H. (2009). Stacking of irregu- Matsuzawa, T. (1995). Chimpanzees are “Chimpan-jin larly shaped blocks in chimpanzees (Pan troglodytes) (human)” [in Japanese]. Tokyo, Japan: Iwanami Shoten. and young humans (Homo sapiens). Animal Cognition, Matsuzawa, T. (2000). Chimpanzee’s mind [in Japanese]. 12,49–58. doi:10.1007/s10071-009-0273-5 Tokyo, Japan: Iwanami Shoten. Henshilwood, C. S., d’Errico, F., van Niekerk, K. L., Matsuzawa, T. (2009). The chimpanzee mind: In search of Coquinot, Y., Jacobs, Z., Lauritzen, S. E., . . . Garcıa- the evolutionary roots of the human mind. Animal Cog- Moreno, R. (2011). A 100,000-year-old ochre-processing nition, 12,1–9. doi:10.1007/s10071-009-0277-1 workshop at Blombos Cave, South Africa. Science, 334, Matsuzawa, T., Biro, D., Humle, T., Inoue-Nakamura, N., 219–222. doi:10.1126/science.1211535 Tonooka, R., & Yamakoshi, G. (2001). Emergence of Henshilwood, C. S., d’Errico, F., & Watts, I. (2009). culture in wild chimpanzees: Education by master- Engraved ochres from the Middle Stone Age levels at apprenticeship. In T. Matsuzawa (Ed.), Primate origins Blombos Cave, South Africa. Journal of Human Evolu- of human cognition and behavior (pp. 557–574). New tion, 57,27–47. doi:10.1016/j.jhevol.2009.01.005 York, NY: Springer. doi:10.1007/978-4-431-09423-4_28 Humphrey, N. (1998). Cave art, autism, and the evolution Matthews, J. (1984). Children drawing: Are young chil- of the human mind. Cambridge Archaeological Journal, 8, dren really scribbling? Early Child Development and Care, 165–191. 18,1–39. doi:10.1080/0300443840180101 The Origin of Representational Drawing 15

Morris, D. (1962). The biology of art. London, UK: Methuen Takeshita, H. (2001). Development of combinatory manip- Young Books. ulation in chimpanzee infants (Pan troglodytes). Animal Myowa-Yamakoshi, M., & Matsuzawa, T. (1999). Fac- Cognition, 4, 335–345. doi:10.1007/s100710100089 tors inﬂuencing imitation of manipulatory actions in Takeshita, H., Myowa-Yamakoshi, M., & Hirata, S. (2009). chimpanzees (Pan troglodytes). Journal of Comparative The supine position of postnatal human infants: Impli- Psychology, 113, 128–136. doi:10.1037/0735-7036.113. cations for the development of cognitive intelligence. 2.128 Interaction Studies, 10, 252–268. doi:10.1075/is.10.2.08tak Patterson, F. (1986). The mind of the gorilla: Conversation Tanaka, M., Tomonaga, M., & Matsuzawa, T. (2003). Finger and conservation. In K. Benirschke (Ed.), Primates: The drawing by infant chimpanzees (Pan troglodytes). Animal road to self-sustaining populations. New York, NY: Cognition, 6, 245–251. doi:10.1007/s10071-003-0198-3 Springer-Verlag. doi:10.1007/978-1-4612-4918-4_63 Tomasello, M. (1999). The cultural origins of human cogni- Pike, A. W. G., Hoffmann, D. L., Garcia-Diez, M., tion. Cambridge, MA: Harvard University Press. Pettitt, P. B., Alcolea, J., De Balbin, R., . . . Zilhao,€ J. Yamagata, K. (1991). A study of scribbles on picture (2012). U-Series dating of Paleolithic art in 11 caves books by 1- and 2-year-old children [in Japanese]. Japa- in Spain. Science, 336, 1409–1413. doi:10.1126/science. nese Journal of Educational Psychology, 39, 102–110. 1219957 Yamagata, K. (2001). Emergence of representational activ- Premack, D. (1975). Putting a face together. Science, 188, ity during the early drawing stage: Process analysis. 228–236. doi:10.1126/science.1118724 Japanese Psychological Research, 43, 130–140. doi:10.1111/ Saito, A., Hayashi, M., Ueno, A., & Takeshita, H. (2011). 1468-5884.00169 Orientation-indifferent representation in children’s drawings. Japanese Psychological Research, 53, 379–390. doi:10.1111/j.1468-5884.2011.00495.x Supporting Information Schiller, P. H. (1951). Figural preferences in the drawings Additional supporting information may be found in of a chimpanzee. Journal of Comparative and Physiological ’ Psychology, 44, 101–111. doi:10.1037/h0053604 the online version of this article at the publisher s Smith, D. A. (1973). Systematic study of chimpanzee website: drawing. Journal of Comparative and Physiological Psychol- Appendix S1. The Criteria for Imitation in ogy, 82, 406–414. doi:10.1037/h0034135 Humans, Modiﬁcation of Ikuzawa et al. (1985)