Neuropsychologia 38 (2000) 1593 – 1606 www.elsevier.com/locate/neuropsychologia
Hemispheric asymmetries and gender influence Rembrandt’s portrait orientations James A. Schirillo * Department of Psychology, Wake Forest Uni6ersity, Winston-Salem, NC 27109 -7778, USA Received 11 June 1999; received in revised form 7 April 2000; accepted 12 April 2000
Abstract For centuries painters have predominantly painted portraits with the model’s left-cheek facing the viewer. This has been even more prevalent with females ( :68%) than males ( : 56%). Numerous portraits painted by Rembrandt typify this unexplained phenomenon. In a preliminary experiment, subjects judged 24 emotional and social character traits in 20 portraits by Rembrandt. A factor analysis revealed that females with their left cheek exposed were judged to be much less socially appealing than less commonly painted right-cheeked females. Conversely, the more commonly painted right-cheeked males were judged to be more socially appealing than either left-cheeked males or females facing either direction. It is hypothesized that hemispheric asymmetries regulating emotional facial displays of approach and avoidance influenced the side of the face Rembrandt’s models exposed due to prevailing social norms. A second experiment had different subjects judge a different collection of 20 portraits by Rembrandt and their mirror images. Mirror-reversed images produced the same pattern of results as their original orientation counterparts. Consequently, hemispheric asymmetries that specify the emotional expression on each side of the face are posited to account for the obtained results. © 2000 Elsevier Science Ltd. All rights reserved. Keywords: Hemispheric asymmetries; Emotion; Gender differences; Facial asymmetry; Portraits
1. Introduction A comprehensive survey of 1474 portraits painted in Western Europe from the 16th to the 20th century revealed that 891 ( : 60%) were posed with their left cheek facing the viewer. More importantly, within this collection : 68% of the women were painted with their left cheek exposed, while only : 56% of the men exposed their left cheek [41,42]. Previous artists may have chosen to expose their patrons left-cheek more often to display an appealing set of character traits or moods [14]. These ‘left-cheek’ traits may have either been more predominant in women or simply more valued by society when expressed by women. Recent work by Nicholls et al. [44] suggests the latter. They asked males and females to pose for a portrait to be viewed by either their ‘close-knit’ family, where they were to reveal their ‘warm-hearted and
* Fax: + 1-336-7584733. E-mail address:
[email protected] (J.A. Schirillo).
affectionate’ emotions or for a portrait where they were to appear as ‘cool-headed, calm and reasonable, successful scientists’. In the ‘emotional’ condition 64% of males and 58% of females turned their left-cheek toward the camera, while in the ‘impassive’ condition only 43% of males and 43% of females turned their left-cheek toward the camera. Nicholls et al. proposes that their results suggest that both genders intuitively know which side of the face effectively expresses emotion. The current study tests the hypothesis that females have traditionally been posed with their left-cheek exposed, not to make them appear more emotional per se, but actually more reserved than their male counterparts who typically exposed their right-cheek. This implies that left-cheeked models should be judged to be significantly less appealing that their right-cheeked counterparts. Likewise, males have traditionally exposed their right-cheek, not to make them appear more impassive per se, but actually more outgoing and approachable. Such models should be judged to be significantly more appealing than their left-cheeked counterparts.
0028-3932/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S0028-3932(00)00063-4
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
1594
Rembrandt’s extensive lifetime collection of 341 portraits was sampled from because it provides an excellent illustration of the gender-specific laterality phenomenon (Table 1) [9]1. Rembrandt’s biased oeuvre might suggest that his patrons found females left-cheek and males right-cheek more aesthetically pleasing. While this is possible, in modern society the right side of a woman’s face is typically judged to be more attractive by both men and women alike. This has been demonstrated using chimerical faces generated from full-face photographs [57]. Interestingly, the same subjects did not show a specific aesthetic preference for either a right– right or left–left composite of men’s faces. Thus, it is possible that during Rembrandt’s era, patrons also preferred right-cheeked females but viewed their left cheek because it conformed to traditional social norms. For example, the majority of Rembrandt’s self-portraits expose his right-cheek, as do the Table 1 Exposed cheek in Rembrandt’s portraits Females
Males
Left cheek
Right cheek
Left cheek
Right cheek
74% N =77
26% n =27
26% N= 62
74% n= 175
Table 2 Exposed cheek by model’s relationship to Rembrandt Females Left cheek
Males Right cheek
Self portraits Family portraits Other portraits
Left cheek
Right cheek
11%
89%
n= 6
n= 50
63%
37%
6%
94%
n =22
N= 13
n= 1
n= 16
80%
20%
34%
66%
n =55
N= 14
n= 55
n= 109
1 There is now a general consensus regarding Rembrandt’s output, although a great many ‘Rembrandts’ have come to be regarded as studio work since the publication of Gerson’s book in 1969 [9] and even more since the Rembrandt Research Project reported its findings between 1982 and 1989’’.…. ‘‘Nevertheless, Gerson’s book by implication remained the canon, as the majority of paintings he had doubted were not even discussed in Volumes II and III of the Corpus of Rembrandt Paintings, 1635 -1642, by J. Bruyn, the Rembrandt Research Project Editor, 1982’’ [4]. Given the ongoing debate concerning the authenticity of Rembrandt’s works, all the portraits used in the current study were authenticated in Bruyn’s Corpus, however, the statistics quoted are in reference to Gerson’s earlier work.
portraits of most of his male kin, but far less so for his pictures of male friends and strangers (Table 2) [29]. This suggests that the right cheek might express prowess, dominance and status, making it important for males to display. Females portraying these same traits may have been considered threatening and therefore encouraged to avoid such postures. A related possibility proposed in the chimerical face literature is that the right side of the face is more salient, bearing a greater resemblance to the whole face [25,26]. Portraits of men may have their right cheek exposed to place themselves on social display, while females could remain coy by exposing their left cheek [28,52]. In 1933, Wolff [55] postulated that there is both a ‘private’ and a ‘public’ side of a face. He proposed that the right side of the face expressed the public side of a person, their individual conscious and expressive character. The right side showed vitality and power, whereas the more passive left side of the face contained the private and unconscious aspects of a person. If this is true, an exposed right cheek might express assertiveness and dominance in men, while an exposed left cheek might indicate gentility in women [32]. That is, a woman may have been painted with her left cheek facing the viewer to entice them by appearing more demure. As a case in point, semantic differential ratings of eight of Leonardo da Vinci’s drawings indicated that subjects considered the right side of the face more potent and active [3]. This makes tenable the suggestion of Nicholls et al. that power and self-control, expressed in right-cheeked portraits, account for why artists’ were predisposed to orient males and females portraits differently. Neuropsychological research has also demonstrated greater emotional expression being displayed predominantly on the left side of the face [43,47,54] (for counter arguments see Refs. [20,21]). Yet a number of studies have also advocated a motoric direction hypothesis. This hypothesis states that while negati6e emotions are predominately displayed on the left side of the face, positi6e (i.e. not neutral or controlled) emotions are predominately displayed on the right side of the face [7,15,18,22,33,34,49,50]. Borod et al. [7] and ReuterLorenz and Davidson [45] argue that this is because the right cerebral hemisphere regulates behavioral withdrawal while the left cerebral hemisphere regulates behavioral approach. Since only the lower two thirds of the face are predominantly contralaterally innervated [19,46], this makes the lower eyelid, nose, cheek, lips and neck of the left side of the face appear hostile, while making the lower two thirds of the right side of the face appear approachable. Research from several laboratories using EEG recordings [16,22,53] concur with this hemispheric– morphological distinction. For example, EEG’s measured in 10-month-old infants demonstrate an increase
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
in left-hemispheric activity when expressing joy or anger (i.e. emotions related to approach) with an increase in right-hemispheric activity when expressing distress, disgust or sadness (i.e. emotions related to withdrawal) [24]. In fact, infants as young as 2-days-old show an increase in right-hemispheric EEG activation when given citrus acid, compared to an increase in left-hemispheric EEG activation when given sugar water [23]. These neuropsychological findings complement those found in the field of aesthetics. There, researchers have even suggested that women are typically portrayed with their less attractive side showing [9,39,56]. The question is why? We hypothesize that female left-cheeked portraits are selected because social norms dictate that women should appear more demure and evasive than men. As neuropsychological research would predict, this can be effectively accomplished by portraying withdrawal, evident in the left side of the face. This hypothesis is also compatible with Wolff’s prevalent aesthetic argument that the left side of the face is private. In contrast, males show their right side (i.e. their public side) more often. However, rather than doing so to appear emotionally neutral or controlled [9], we hypothesize that the left-hemisphere regulation of approach makes the right side of the face appear to have positive emotional qualities. Although there have been extensive developments regarding facial expressions in the fields of both aesthetics and neuropsychology, studies linking the two disciplines have been limited. The current study attempts to bridge this gap by developing a psychological scale to analyze and differentiate the social and emotional content of portraits by Rembrandt. Sackheim and Gur [47] found laterality differences when using composite photographs while specifically asking the model to display a certain emotion. The argument in the current study is that Rembrandt chose to paint females left side primarily to capture their reserved qualities, which may have also inadvertently emphasized their negative emotions. Male models were posed to expose their right side to reveal their outgoing qualities, which may have accentuated their positive qualities as well. The advantage to viewing ‘naturalistic’ portraits painted by Rembrandt compared to chimerical faces that typify the literature is that we propose that each side of the face expresses different emotions. Thus, both sides of the face must be viewed simultaneously. The side of the face that is more predominantly displayed will determine the emotions that are judged to be more prevalent. An important concern in conducting this research is the possibility that a hemispheric asymmetry may effect the viewer’s judgments of Rembrandt’s work. That is, the hemispheric differences reported may be more perceptual than expressive. For example, Reuter-Lorenz
1595
and Davidson [45] presented happy or sad versus neutral faces to either the right or left visual field. Reaction times were faster in judging right versus left field presentations when the expression was happy (i.e. judged by the viewer’s left hemisphere) and vice versa (i.e. judged by the viewer’s right hemisphere) when it was sad. This suggests that each hemisphere may be specialized for judging, as well as producing, positive and negative emotional displays (see also Ref. [27]). If this were the case, left-cheeked portraits viewed straight on would be processed more by the right hemisphere, which is more sensitive to negative emotional inflections. Consequently, left-cheeked portraits may appear less pleasant either because they are viewed more directly by the observer’s right hemisphere or because the right hemisphere of the model displays more negative emotions on their left cheek. This latter notion has received support from research related specifically to facial attractiveness, which suggests that it is the physiognomy of right–right and left–left facial composites that determine attractiveness and not the asymmetrical perceptual process that a mirror-reversed image might generate [13]. To control for possible effects of perceptual asymmetries, a second study will compare judgements of original right- or left-cheeked portraits as well as their mirror images. This will demonstrate that viewers react to the lateral physiognomy of the face and are not influenced by the visual field containing the exposed side of the model’s face.
2. Experiment 1
2.1. Method 2.1.1. Participants Sixty-five male and 65 female introductory psychology students from a small, liberal arts university participated in this study for course credit. Both genders were included as raters because it has been shown that men and women sometimes rate the social characteristics of identical photographs differently [51]. However, Rembrandt may have painted primarily for male patrons, since they typically commissioned his work. Handedness was also recorded to prevent any particular bias. For example, aesthetic preferences for rightward and leftward balanced pictures have been shown to exhibit a hemispheric bias [40]. Furthermore, left-handed subjects have been reported to perceive happier faces in their left visual field, while right-handed subjects have not shown this bias [27]. 2.1.2. Materials Twenty of Rembrandt’s portraits were randomly chosen from his collection of 341 [9]. Five were leftcheeked males, five were left-cheeked females, five were
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
1596 Table 3 Twenty-four trait-pairs load on four factors Pleasantness Affectionate–reserved Lively–serious Warm–cold Imperturbable–high-strung Relaxed–nervous
Dominance Dominant–submissive Assertive–compliant Astute–naı¨ve Bold–timid Rash–cautious
Agreeableness Agreeable–quarrelsome Emotional–unemotional Sensitive–insensitive Independent–dependent Respectful–disrespectful Regardful–disregardful Polite–rude
Social appeal Approachable–aloof Peaceful–volatile Kind–unkind Sophisticated–unsophisticated Wise–ignorant Trustful–distrustful Profound–superficial
right-cheeked males and five were right-cheeked females. The selected portraits were all busts. Each portrait was scanned into PhotoShop on an IBM computer. The portraits were then projected onto a screen using Microsoft PowerPoint. Fixation was not required nor were the images restricted to either side of the visual field and thereby to a specific cerebral hemisphere. There were five different randomized orders of the 20 portraits to counteract potential order effects. Each subject rated each portrait on 24 personality and emotional trait pairs using a standard form. The traits in each pair were placed on opposite ends of a nine-point continuum, such that a rating of five would indicate a neutral evaluation of the portrait on that specific trait pair dimension. Half of the items were reversed scored, so that positive traits were not necessarily numbers greater than five.
2.1.3. Procedure Subjects were given an evaluation packet upon arrival to the 1-h experimental session. There were 26 subjects per session. Under dimmed illumination the subjects viewed each portrait for 2 min. They were told that they could begin the evaluation whenever they felt ready, but they would have to move on to the next painting when it was displayed. Subjects completed each evaluation well within the 2-min limit. 2.2. Results An ANOVA indicated that there were no statistically significant differences due to handedness, F(1,129)= 0.43, or subject gender, F(1,129)=0.33. The main effects of Experiment 1 were determined after a factor analysis. The factor analysis loaded the 24 trait-pairs on four factors: pleasantness, dominance, agreeableness and social appeal (Table 3). The loading factor for each trait is given in parenthesis in Appendix A. A 2 × 2 (portrait gender× portrait orientation) within-subjects repeated measures ANOVA was then performed on each factor. The statistics reported in the text are for the data sets shown in Figs. 3, 5 and 6. That is, one graph for each factor (pleasantness, dominance, agreeableness and social appeal). These results revealed that overall, left-cheeked females (i.e. those most often portrayed) were perceived as having a number of negative traits compared to males facing either direction and sometimes to right-cheeked females. Specific traits (i.e. factors) will be discussed within each of the following four categories.
Fig. 1. Pleasantness judgments of left and right portraits of males and females.
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
Fig. 2. Self-portrait of Rembrandt facing right (c. 1657) ©1969 Phaidon Press Ltd. Rembrandt, The Complete Edition of Paintings. A. Bredius, revised by H. Gerson.
2.2.1. Pleasantness Overall, female portraits painted by Rembrandt were considered to be less pleasant than male portraits [F(3,129)=67.44; P B0.01] (Fig. 1). Within this category, females and males were judged to possess a complex configuration of emotions. For example, female portraits were judged to have some pleasant traits, that is, they were considered to be both affectionate and lively (mean = 6.88 and 6.69, respectively; see Appendix A for means). However, they were also judged
1597
as having a number of anxious traits, such as being nervous and high-strung (mean= 3.94 and 3.82, respectively). In contrast, although the portraits of men were judged as being somewhat more reserved and serious (mean=4.40 and 4.53, respectively), they were also considered imperturbable and relaxed (mean= 7.98 and 8.92, respectively), thus producing higher scores overall on pleasantness. Interestingly, portraits of right-cheeked males (i.e. those Rembrandt preferred for images of himself and his next of male kin) were judged to be the most imperturbable and relaxed (mean= 8.95 and 9.72, respectively). These images project the emotional antithesis of left-cheeked females (mean= 2.83 and 4.81, respectively). Although viewers were not informed that one of Rembrandt’s self-portraits was included in the set of 20, his right-cheeked face typifies the relaxed appearance of other right-cheeked males (Fig. 2). Note, however, that he may have painted himself from a mirror image, thus this may be his left-cheek. Moreover, familiarity might have also influenced subject’s judgments of this portrait. However, recalculating all the ANOVA’s with this portrait excluded did not alter the statistical significance of any of the reported results.
2.2.2. Dominance The dominance category (Fig. 3) again reflects the men’s solemn character. As expected, compared to women, men were judged as dominant, assertive and bold [F(3, 129)= 83.62; PB 0.01; see Appendix A]. However, either side of the men’s faces was likely to portray dominance. Right-cheeked women (those portrayed less often), on the other hand, were judged as timid and cautious (mean= 2.85 and 2.62, respectively)
Fig. 3. Dominance judgments of left and right portraits of males and females.
1598
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
much more disagreeable than men overall [F(3,129)= 57.91; PB 0.01]. Interestingly, portraits of left-cheeked females were judged to be impersonal, that is, quarrelsome, disrespectful, disregardful and rude compared to the other three portrait categories, while right-cheeked males show the opposite trends (i.e. being agreeable, sensitive and polite; see Appendix A). This suggests that these former traits, often considered negative by current social standards, may not have been the overriding aesthetic concern in Rembrandt’s time. Such defensive traits could simply have been the consequence of selecting a reserved (i.e. withdrawn) left-cheeked posture (note the left side of the women’s face in Fig. 4).
Fig. 4. Portrait of a young woman facing left (c. 1665). ©1969 Phaidon Press Ltd. Rembrandt, The Complete Edition of Paintings. A. Bredius, revised by H. Gerson.
compared to left-cheeked women (mean=5.69 and 6.27, respectively).
2.2.3. Agreeableness While pleasantness may be likened to agreeableness, the latter category includes more interpersonal traits, like sensitivity and respectfulness. These traits might conflict with attributes that suggest prowess or dominance. Consequently, this category contains the most revealing findings (Fig. 5). Females were judged to be
2.2.4. Social appeal While related to the pleasantness category, social appeal includes a moral quality. As with pleasantness and agreeableness, females were judged to be significantly less appealing than their male counterparts [F(3,129)= 75.68; PB 0.01] (Fig. 6) and like agreeableness, portraits of left-cheeked females were considered the least positive compared to all other portrait groups. So, while artists may have painted left-cheek females most often to elicit their demure ‘private’ side, modern day viewers judged this reserved characteristic to be socially unappealing. In contrast, subjects rated right-cheeked portraits of males very highly (those most often portrayed). Such images are judged as being the most peaceful and kind (mean= 8.31 and 8.54, respectively). This is consistent with the notion that the right-cheek displays an extroverted ‘public’ side that is judged positively. For example, notice how the terms peaceful and kind can be easily attributed to Rembrandt’s self-portrait (Fig. 2),
Fig. 5. Agreeableness judgments of left and right portraits of males and females.
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
1599
Fig. 6. Appealingness judgments of left and right portraits of males and females.
while the woman in Fig. 4 has a much more constrained and ambiguous disposition.
2.3. Discussion The study of physiognomy claims that one can accurately judge the social and emotional character traits of others from their appearance. For example, it has been demonstrated that humans can consistently stereotype the facial characteristics of ‘socially acceptable’ men (r =0.83) and women (r =0.52) [51]. Likewise, subjects agree in how they attribute character traits to strangers’ photographs, which mediate or reflect their expectations of such strangers’ likely social behavior [28]. Since portraiture was the primary means of illustrating an individual’s expected social behavior, it was critical for artists to capture the complexity of traits that the current social order considered most relevant. Hence, the emotional content present on the left side of the face in Fig. 4 depicts the withdrawn aspects of this woman’s character. In contrast, note the turning of her right eyebrow signaling approach. Part of Rembrandt’s greatness is his ability to capture inconsistent emotions that are present simultaneously, such as those that underlie approach (i.e. right side) and withdrawal (left side). The fact that both approach and withdrawal are present concurrently has been postulated by Kinsbourne [33] (p. 415): ‘‘The left hemisphere seems to marshal reasons for benefits, the right, reasons for costs. Pathological
amplification of these functions appears to be related to left-generated ‘euphoric-indifference’ states and the right-generates ‘dysphoric-depressive’ states. In the former, positive aspects are overvalued and negative aspects ignored. In the latter, the converse occurs. We see, then, hemispheric balance between approach and withdrawal tendencies’’ Thus, turning the head to conform to social convention can make one set of emotions appear predominant, but not exclusive. Unfortunately, historical and cultural differences make it impossible to use today’s standards to ascertain Rembrandt’s intention or the effect his work had on others. However, it is reasonable to consider that his choice of which side of the face to profile was not an arbitrary decision. For example, naı¨ve artists predominantly draw profiles facing left. This is true, even for individuals raised in cultures that read right to left [30,31]. So, why accomplished artists change their initial preferred portrait orientation, especially in painting men, is a mystery. Current research has discovered that left–left composite photographs are perceived as being happier [57]. Yet the above findings suggest that full-face portraits of left-cheeked models are anything but happy. Therefore, narrowly defined traits, such as happiness, may not be as important to depict as the complexity of emotion available in approach/avoidance displays. To ensure that it is the model’s side of the face that determines these characterizations, and not the viewer’s orientation, a second study was conducted.
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
1600
3. Experiment 2
3.1. Method 3.1.1. Participants A second set of 65 male and 65 female students from the previous population was used. 3.1.2. Materials Seventeen new portraits of Rembrandt were chosen from his collection while retaining three from the previous experiment. Again, five were left-cheeked males, five were left-cheeked females, five were right-cheeked males, and five were right-cheeked females. This time, however, the most rightward facing and leftward facing bust portraits painted by Rembrandt were chosen. The selected portraits were all busts. Each portrait was displayed as in the previous experiment; that is, in five different randomized orders. 3.1.3. Procedure Half of the subjects saw 20 original Rembrandt’s while the remaining half saw their mirror-reverse. They were not informed of this manipulation. Each subject rated each portrait on the 24 personality and emotional trait pairs used in the previous study. 3.2. Results An ANOVA indicated that there were no statistically significant differences due to handedness, F(1,129)= 0.38, or subject gender, F(1,129) = 0.41. The three faces reused from Experiment 1 produced a pattern of results consistent with Experiment 1, suggesting that the original findings are replicable. As with Experiment 1, the
main effects of Experiment 2 were determined after a factor analysis. The factor analysis loaded the 24 traitpairs on four factors: pleasantness, dominance, agreeableness and social appeal. The loading factor for each trait is given in parenthesis in Appendix B. In 19 of 20 portraits, subjects typically judged the social and emotional content of a face in a portrait by the qualities that particular side exhibits and were unaffected by the hemispheric asymmetry that a particular view affords. The counter intuitive findings for left-facing females (i.e. those most often portrayed) as being perceived as less pleasant than other portraits was replicated [F(3,129)= 21.23; P B 0.01] (compare Figs. 1 and 7). As in Experiment 1, females were also perceived as being more submissive (compare Figs. 3 and 8), with rightfaced females being the most submissive appearing group [F(3,129)= 36.32; P B 0.01]. There was a slight, but statistically insignificant, decrease in dominance of mirror-reversed males compared to males judged in their original orientation [F(3,129)= 0.12]. The one finding that was statistically significant occurred between original and mirror-reverse males on the agreeableness category (Fig. 9; Appendix B). Males who originally faced left (Fig. 10 is an example) were perceived as less agreeable than when they were mirrorreversed [F(3,129)= 41.66; P B 0.01] (Fig. 11 is an example). That is, the later group appeared as if they had their right-cheek exposed to naı¨ve subjects. Likewise, males who originally faced right (Fig. 2) were perceived as more agreeable than when they were mirror-reversed [F(3,129)= 39.75; P B 0.01] (Fig. 9). In essence, the exposed cheek difference that existed in the original male portraits was eliminated when the portrait mirror images were judged in agreeableness.
Fig. 7. Pleasantness judgments of original and mirror-reversed left and right portraits of males and females.
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
1601
Fig. 8. Dominance judgments of original and mirror-reversed left and right portraits of males and females.
Fig. 9. Agreeableness judgments of original and mirror-reversed left and right portraits of males and females.
As in Experiment 1, subjects rated females less appealing than males and reversing the portraits did not alter these findings appreciably [F(3,129) = 43.76; PB 0.01] (Fig. 12). 4. General discussion Previous studies have demonstrated that Rembrandt typified the work of many artists by painting females predominantly with their left-cheek exposed [9,41,42]. The question is why? A more traditional neuropsycho-
logical perspective has been that, independent of methodology (i.e. composite photos, slow motion videotapes, naturalistic observations), the left side of the face is more active in emotional expression [10,12,43,47]. Recent work by Nicholls et al. (p. 1520) concurs, suggesting that ‘‘…it is possible that emotional expression does not serve to enhance beauty and that the more reserved right side of the face may be perceived as more attractive.’’
1602
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
However, it is unclear why this would be the case. An alternate hypothesis, more consistent with current neuropsychological literature, is that the right side of the face is not necessarily more reserved, but actually appears positive in comparison to the negative qualities displayed on the left side of the face. The current study demonstrates that model’s left side conveys very unappealing character dispositions, while their right side conveys the opposite.
There is an increasing consensus in the neuropsychological literature [15,17,33,34] that the left hemisphere regulates avoidance behavior while the right hemisphere regulates approach behavior. Some take this to mean that the right side of the face appears active while the left side of the face appears inactive. This agrees with a separate body of research on the aesthetics of portraiture that indicates that the right-cheek reflects the ‘public’ side and the left-cheek the ‘private’ side of a model
Fig. 10. Original portrait of Nicolaes Bruyningh facing left (c. 1652). ©1969 Phaidon Press Ltd. Rembrandt, The Complete Edition of Paintings. A. Bredius, revised by H. Gerson.
Fig. 11. Mirror-reversed portrait of Nicolaes Bruyningh originally facing left (c. 1652). ©1969 Phaidon Press Ltd. Rembrandt, The Complete Edition of Paintings. A. Bredius, revised by H. Gerson.
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
1603
Fig. 12. Appealingness judgments of original and mirror-reversed left and right portraits of males and females.
[55]. In addition, our findings indicate that the rightcheek is not only more ‘public’, but also reflects more positive underlying character traits. Related work has shown that photographs of right side composites of chimerical faces are rated as more attractive [13]. A number of studies have also postulated that the right hemisphere may simply be somewhat more efficient than the left hemisphere in processing emotions [6,8,11,37,38]. This is significant since a left-cheek portrait allows more of the face to fall on the left half of the painting, which projects to the right-hemisphere. However, the findings of Experiment 2 suggest that each side of the face carries a unique emotional disposition and that the asymmetrical results found in Experiment 1 are not due to a perceptual bias originating in the viewer. These findings also address the concerns of Gilbert and Bakan [25] and Grega et al. [26] that right-side facial saliency is due to an artifact of lateral viewing. The current study used only a single full-face, turned to expose predominantly one side versus the unnatural composite images used by Gilbert and Bakan and Grega et al. Moreover, a perceived laterality bias has only been reported for similarity judgments, while the current study addresses the perception of specific emotional dispositions. Related studies [48] have also been limited, in that while they have shown that subjects rate the intensity of emotions higher when the left visual field is accentuated, they fail to address what emotions a subject recognizes when viewing a particular side of the face. In general, males are more lateralized and show greater right hemispheric involvement (i.e. the left side of their face depicts more affect) than females [2].
Consequently, turning males left side away from the viewer minimizes displaying certain negative feelings, making them appear more agreeable and socially appealing than they might otherwise. Females, however, have typically been portrayed with their left-cheek exposed. Thus, the lack of appeal in these images may be physiologically based [5,36]. For example, Alford [1] has shown that females that are able to hide their feelings have a better ability to regulate the left side of their face. Thus, the left side of their face may not reveal salient emotions per se, but selective negative emotions. Kowner [35], using a more modern technique of composing right and left hemifaces, argues that attributions of emotion and personality are not revealed laterality in resting faces and only have a slight effect in posed faces. The current study, which displayed whole faces, does find that the emotions and moods portrayed are highly asymmetrical. Given the assumption that Rembrandt chose to depict specific aspects of the emotional character and social status of his patrons, his decision to expose the left cheeks of females and the right cheeks of males suggests his awareness of lateralized gender differences.
Acknowledgements Special thanks to Alexandra Kutz, Alison Pinnix and Julia Settle for data collection; William Fleeson and Mark Leary for statistical consultation; and Guy Bass for technical assistance.
1604
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
Appendix A. Experiment 1: Original portrait orientation Left female
Right female
Left male
Right male
Pleasantness Affectionate–reserved (0.746) Lively–serious (0.739) Warm–cold (0.641) Imperturbable–high-strung (0.475) Relaxed–nervous (0.370)
7.08 7.13 5.81 2.83 3.02
6.68 6.26 5.14 4.81 4.87
5.33 5.18 5.75 7.01 8.12
3.48 3.89 7.72 8.95 9.72
Dominant Dominant–submissive (0.823) Assertive–compliant (0.718) Astute–naı¨ve (0.652) Bold–timid (0.432) Rash–cautious (0.422)
7.33 7.21 5.07 5.69 6.27
7.47 7.66 5.68 2.85 2.62
8.57 8.81 4.76 8.46 6.13
8.68 9.05 5.21 8.82 6.51
Agreeableness Agreeable–quarrelsome (0.785) Emotional–unemotional (0.725) Sensitive–insensitive (0.648) Independent–dependent (0.426) Respectful–disrespectful (0.421) Regardful–disregardful (0.289) Polite–rude (0.249)
2.13 6.43 5.33 7.94 2.35 2.12 2.01
5.33 5.04 5.47 4.98 5.39 5.21 5.38
4.97 5.19 4.87 6.56 5.22 5.17 5.03
7.12 5.33 7.11 4.74 5.84 5.69 6.61
Social appeal Approachable–aloof (0.763) Peaceful–volatile (0.686) Kind–unkind (0.584) Sophisticated–unsophisticated (0.342) Wise–ignorant (0.144) Trustful–distrustful (0.103) Profound–superficial (0.088)
5.11 4.88 5.02 4.68 4.83 4.85 4.42
5.31 5.67 5.53 5.86 5.86 5.21 5.18
5.69 5.83 5.57 5.76 5.76 6.14 5.83
5.62 8.31 8.54 5.23 5.23 5.39 5.37
Numbers in parenthesis indicate loading factor.
Appendix B. Experiment 2: Original and mirror reversed portrait orientation Original Left female
Right female
Left male
Right male
Reversed Left female
Right female
Left male
Right male
Pleasantness Affectionate–reserved (0.863) Lively–serious (0.844) Warm–cold (0.550) Imperturbable–high-strung (0.511) Relaxed–nervous (0.463)
6.98 7.22 6.12 2.63 2.95
6.72 5.97 5.33 5.21 5.83
5.78 5.69 5.86 7.23 7.83
3.21 3.54 7.52 8.23 8.83
7.86 7.43 5.34 3.21 2.87
6.99 5.74 5.85 5.74 6.26
5.42 5.12 6.23 6.97 6.88
3.55 3.26 6.83 7.84 8.16
Dominant Dominant–submissive (0.795) Assertive–compliant (0.749) Astute–naı¨ve (0.712) Bold–timid (0.432) Rash–cautious (0.422)
7.16 6.68 5.64 5.16 5.63
7.24 7.21 5.87 3.23 2.97
8.84 8.23 5.11 8.12 5.78
8.72 8.25 5.16 8.47 6.72
7.42 7.27 5.67 5.45 5.97
7.16 6.97 5.73 3.17 3.42
8.36 8.13 5.23 7.82 5.23
8.65 8.16 5.72 8.26 5.70
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
Agreeableness Agreeable–quarrelsome (0.898) Emotional–unemotional (0.609) Sensitive–insensitive (0.602) Independent–dependent (0.513) Respectful–disrespectful (0.456) Regardful–disregardful (0.245) Polite–rude (0.217)
1605
2.23 5.88 5.01 7.34 2.12 2.1 2.21
5.62 4.97 5.33 4.43 5.64 5.37 5.96
5.24 4.97 4.68 6.25 4.86 4.79 4.72
7.23 5.16 7.27 4.91 5.93 5.73 6.92
2.56 6.14 5.27 7.54 3.21 2.45 2.43
5.12 4.32 4.92 5.34 5.26 5.27 5.24
5.67 5.36 5.15 6.57 5.21 5.23 6.34
6.93 5.22 6.33 4.86 4.92 5.63 6.23
Social appeal Approachable–aloof (0.680) 4.91 Peaceful–volatile (0.596) 4.83 Kind–unkind (0.553) 5.22 Sophisticated–unsophisticated (0.452) 5.21 Wise–ignorant (0.371) 5.47 Trustful–distrustful (0.233) 4.96 Profound–superficial (0.218) 4.82
5.62 5.24 5.22 5.25 5.17 5.52 5.42
5.69 5.21 6.03 5.18 6.16 5.62 5.78
5.82 7.93 8.36 5.53 6.35 5.46 5.76
6.21 5.22 4.86 5.34 5.56 5.34 5.52
5.61 5.34 5.71 5.27 5.52 5.73 5.38
6.24 5.41 5.27 5.86 6.32 5.76 6.16
5.24 7.85 8.74 5.67 6.29 5.02 5.40
Numbers in parenthesis indicate loading factor. References [1] Alford RD. Sex differences in lateral facial facility: The effects of habitual emotional concealment. Neuropsychologia 1983;21:567 – 70. [2] Alford R, Alford KF. Sex differences in asymmetry in the facial expression of emotion. Neuropsychologia 1981;19:605–8. [3] Benjafield J, Segalowitz SJ. Left and right in Leonardo’s drawing of faces. Empirical Studies of the Arts 1993;11:25–32. [4] Blankert A. Rembrandt, A Genius and His Impact. Zwolle, Netherlands: Waanders, 1997:53–4. [5] Borod JC, Caron HS. Facedness and emotion related to lateral dominance, sex and expression type. Neuropsychologia 1980;18:237 – 41. [6] Borod JC, Koff E, Lorch MP, Nicholas M. The expression and perception of facial emotion in brain-damaged patients. Neuropsychologica 1986;24:169–80. [7] Borod JC, Haywood CS, Koff E. Neuropsychological aspects of facial asymmetry during emotional expression: A review of the normal adult literature. Neuropsychology Review 1997;7:41 – 60. [8] Bowers D, Bauer RM, Coslett HB, Helilman KM. Processing of faces by patients with unilateral lesions. I. Dissociations between judgments of facial affect and facial identity. Brain and Cognition 1985;4:258 – 72. [9] Bredius A. The Paintings of Rembrandt. The Complete Edition of the Paintings. London, UK: Phaidon, 1969 Revised by H. Gerson. [10] Campbell R. Asymmetries in interpreting and expressing a posed facial expression. Cortex 1978;14:327–42. [11] Campbell R. The lateralization of emotion: A critical review. International Journal of Psychology 1982;17:211–29. [12] Chaurasia BD, Goswami HK. Functional asymmetry in the face. Acta Anatomica 1975;91:154–60. [13] Chen AC, German C, Zaidel DW. Brain asymmetry and facial attractiveness: Facial beauty is not simply in the eye of the beholder. Neuropsychologia 1997;35:471–6. [14] Conesa J, Brunold-Conesa C, Miron M. Incidence of the halfleft profile pose in single-subject portraits. Perceptual and Motor Skills 1995;81:920 – 2. [15] Davidson R. Affect, cognition and hemispheric specialization. In: Izard CE, Kaan J, Zajonc R, editors. Emotions, Cognition and Behavior. Cambridge, UK: Cambridge Press, 1984:320 – 65.
[16] Davidson RJ, Schwartz GE, Pugash E, Bromfield E. Sex differences in patterns of EEG asymmetry. Biological Psychology 1976;4:119 – 38. [17] Davidson RJ, Ekman P, Saron CD, Senulis JA, Friesen WV. Approach-withdrawal and cerebral asymmetry: Emotional expression and brain physiology I. Journal of Personality and Social Psychology 1990;58:330 – 41. [18] Davidson R, Saron CD, Senulis JA, Friesen WV. Approach– withdrawal and cerebral asymmetry: Emotional expression and brain physiology I. Journal of Personality and Social Psychology 1990;58:330 – 41. [19] DeJong RN. The Neurological Examination. New York: Harper and Row, 1979. [20] Ekman P, Nelson CA, Horowitz FD, Spinrad SI, Sackeim HA, Gur RC. Asymmetry in facial expression. Science 1980;209:833– 6. [21] Ekman P, Hager JC, Friesen WV. The symmetry of emotion and deliberate facial actions. Psychophysiology 1981;18:101 –6. [22] Fox NA. If it’s not left, it’s right. American Psychologist 1991;46:863 – 72. [23] Fox NA, Davidson RJ. Taste-elicited changes in facial signs of emotion and the asymmetry of brain electrical activity in human newborns. Neuropsychologia 1986;24:417 – 22. [24] Fox NA, Davidson RJ. Patterns of brain electrical activity during facial signs in 10-month-old infants. Developmental Psychology 1988;24:230 – 6. [25] Gilbert C, Bakan P. Visual asymmetry in perception of faces. Neuropsychologia 1973;11:355 – 62. [26] Grega DM, Sackheim HA, Sanchez E, Cohen BH, Hough S. Perceiver bias in the processing of human faces: Neuropsychological mechanisms. Cortex 1988;24:91 – 117. [27] Heller W, Levy J. Perception and expression of emotion in right-handers and left-handers. Neuropsychologia 1981;19:263– 72. [28] Hochberg J, Galper RE. Attribution of intention as a function of physiognomy. Memory and Cognition 1974;2:39 – 42. [29] Humphrey N, McManus C. Status and the left cheek. New Scientist 1973;59:437 – 9 p. 581. [30] Jensen BT. Reading habits and left – right orientation in profile drawings by Japanese children. American Journal of Psychology 1952a;65:306 – 7. [31] Jensen BT. Left-right orientation in profile drawing. American Journal of Psychology 1952b;65:80 – 3.
1606
J.A. Schirillo / Neuropsychologia 38 (2000) 1593–1606
[32] Keating CF. Gender and the physiognomy of dominance and attractiveness. Social Psychology Quarterly 1985;48:61–70. [33] Kinsbourne M. Hemispheric specialization and the growth of human understanding. American Psychologist 1982;37:411 – 20. [34] Kinsbourne M, Bemporad B. Lateralization of emotion: A model and evidence. In: Fox N, Davidson R, editors. The Psychobiology of Affective Development. Hillsdale, NJ: Erlbaum, 1984:259 – 91. [35] Kowner R. Laterality in facial expressions and its effect on attributions of emotions and personality: A reconsideration. Neuropsychologia 1995;33:539–59. [36] Ladavas E, Umilta C, Ricci-Bitti AE. Evidence for sex differences in right-hemisphere dominance for emotions. Neuropsychologia 1980;18:361–6. [37] Ley RG, Bryden MP. Hemispheric differences in processing emotions and faces. Brain and Language 1979;7:127–38. [38] Mandal MK, Tandon SC, Asthana HS. Right brain damage impairs recognition of negative emotions. Cortex 1991;27:247 – 53. [39] McLaughlin JP, Murphy KE. Preferences for profile orientation in portraits. Empirical Studies of the Arts 1994;12:1–7. [40] McLaughlin JP, Dean P, Stanley P. Aesthetic preference in dextrals and sinistrals. Neuropsychologia 1983;21:147–53. [41] McManus IC. Turning the left cheek. In: Humphrey N, editor. Consciousness Regained: Chapters in the Development of Mind. Oxford, UK: Oxford University Press, 1983:138–42. [42] McManus IC, Humphrey NK. Turning the left cheek. Nature 1973;243:271 – 2. [43] Moscovitch M, Olds J. Asymmetries in spontaneous facial expressions and their possible relations to hemispheric specialization. Neuropsychologia 1982;20:71–81. [44] Nicholls MER, Clode D, Wood SJ, Wood AG. Laterality of expression in portraiture: Putting your best cheek forward. Proceedings of the Royal Society of London 1999;226:1517– 22. [45] Reuter-Lorenz P, Davidson RJ. Differential contributions of the two cerebral hemispheres to the perception of happy and sad faces. Neuropsychologia 1981;19:609–13.
[46] Rinn WB. The neuropsychology of facial expression: A review of the neurological and psychological mechanisms for producing facial expressions. Psychological Bulletin 1984;95:52–77. [47] Sackeim HA, Gur RC. Lateral asymmetry in intensity of emotional expression. Neuropsychologia 1978;16:473 – 82. [48] Sackheim HA, Grega D. Perceiver bias in the processing of facial emotional information. Brain and Cognition 1987;6:464– 73. [49] Schiff BB, MacDonald B. Facial asymmetries in the spontaneous response to positive and negative emotional arousal. Neuropsychologia 1990;28:777 – 85. [50] Schwartz GE, Ahern GL, Brown S. Lateralized facial muscle response to positive and negative emotional stimuli. Psychophysiology 1979;16:561 – 71. [51] Secord PF, Muthard JE. Personalities in faces: II. Individual differences in the perception of women’s faces. Journal of Abnormal and Social Psychology 1954;34:238 – 42. [52] Secord PF, Muthard JE. Personality in faces: IV. A descriptive analysis of the perception of women’s faces and the identification of some physiognomic determinants. The Journal of Psychology 1955;39:269 – 78. [53] Silberman EK, Weingarten H. Hemispheric localization of functions related to emotions. Brain and Cognition 1986;5:322 – 53. [54] Skinner M, Mullen B. Facial asymmetry in emotional expression: A meta-analysis of research. British Journal of Social Psychology 1991;30:113 – 24. [55] Wolff W. The experimental study of forms of expression. Character and Person 1933;2:168 – 73. [56] Zaidel DW, Fitzgerald P. Sex of the face in western art: Left and right in portraits. Empirical Studies of the Arts 1994;12:9 – 18. [57] Zaidel DW, Chen AC, German C. She is not a beauty even when she smiles: Possible evolutionary basis for a relationship between facial attractiveness and hemispheric specialization. Neuropsychologia 1995;33:649 – 55.