The effect of left and right poses on the expression of facial emotion

https://doi.org/10.1016/S0028-3932(02)00024-6Get rights and content

Abstract

Research, using composite facial photographs has demonstrated that left–left composites are more emotionally expressive than are right–right composites. The present study investigated whether hemifacial asymmetries in expression are apparent in photographs, that feature one side of the face more than the other. Photographs were taken of the models who turned their heads: (a) 15° to the left, (b) 15° to the right or (c) faced directly towards the camera. It was predicted that left hemiface and midline photographs would be judged as more emotionally expressive than right hemiface photographs, where the left hemiface is less prominent. Three hundred and eighty-four participants viewed photographs of the three posing conditions, and rated each photograph along an emotional expressivity scale. Midline and left hemiface portraits were rated as more emotionally expressive than were right hemiface portraits. To investigate whether this effect was caused by observer’s aesthetic/perceptual biases, mirror-reversed versions of the three posing conditions were included. Left hemiface and midline portraits were rated as more emotionally expressive, irrespective of whether they were mirror-reversed. It was concluded that head turns of just 15° can bring about significant changes in the perceived emotionality. The relevance of these findings to painted portraits, which feature the left hemiface more than the right, is discussed.

Introduction

Over the past quarter of a century, a large body of research has been accumulated, which demonstrates that the left hemiface is more emotionally expressive than the right hemiface [1]. For example, Sackheim and Gur [14] investigated facial asymmetries in emotional expression using composite photographs. They divided photographs of faces into left and right halves; mirror reversed half of them, and then rejoined them at their mid-sagittal axes to form left–left and right–right composites. When participants were asked to select the image that appeared more emotional, they selected the left–left composite more often than the right–right composite. Borod and coworkers [2] suggested that the preference for left hemiface composites was more consistent for negative emotions and stronger for male faces. The observer’s preference for left hemiface composites, presumably, reflects the fact that the left inferior portion of the model’s face is innervated by the right cerebral hemisphere, which is dominant for the expression of emotion [13].

The present study sought to determine whether the asymmetries in expression observed for relatively artificial composite photographs could be observed in normal portraits. An impetus for this study comes from investigations of lateral biases in portraiture. In a survey of 1474 painted portraits from Western Europe dating from the 16th to the 20th Centuries, McManus and Humphrey [9] found that 68% of female and 56% of male portraits featured the left side of the face. Nicholls et al. [12] suggested that this leftward bias reflects a desire to emphasise the emotional qualities of the left hemiface. In support of this proposition, they demonstrated that models, when instructed to adopt an ‘emotional’ role, intuitively knew to turn their left cheek towards the camera. Conversely, when models were asked to conceal their emotion, they tended to turn the right cheek. In a follow up study, Nicholls et al. [11] demonstrated that individuals who rated high on an emotional expressivity scale [7] were more likely to turn their left hemiface toward the camera than were individuals with lower emotional expressivity scores.

Bearing in mind the research demonstrating that models turn the left hemiface when expressing emotion, the present study tested whether an observer’s impression of a model’s emotionality is affected by the hemiface which the model turns. Models were photographed as they: (a) turned the left cheek, (b) faced directly toward the camera or (c) turned the right cheek. We predicted that left hemiface and midline portraits (which feature the left, emotional side of the face) would be judged as more emotionally expressive than right hemiface portraits (which partly conceal the left hemiface). It is possible, however, that differences between the posing conditions could be generated by an aesthetic/perceptual asymmetry on behalf of the observer that is independent of asymmetries in the model’s expression. That is, certain posing arrangements may be preferable to others. Observer’s perceptual asymmetries could be generated by a leftward bias (from the viewpoint of the observer) in attention that occurs when viewing faces (e.g. [8], [10]). Alternatively, an asymmetry could be generated by an aesthetic preference for certain poses [6]. To investigate the relative effects of observer’s perceptual/aesthetic asymmetries and facial asymmetries in expression, we included mirror-reversed versions of the three posing conditions (see Fig. 1a). If perceived emotionality is affected by the cheek the model turns, but not by an aesthetic/perceptual asymmetry on behalf of the observer, no effect of orientation (normal/mirror-reversed) should be observed.

Section snippets

Subjects

Three hundred and eighty-four university students (modal age=18 years, m=74, f=274) participated in this study as part of their course requirements. Participants were naı̈ve in relation to aims and expected outcomes of the study.

Stimuli

Six models (m=f) with a mean age of 26 years were photographed. Images were taken professionally using a Kodak DCS460 digital camera, which created 2048×3072 pixel colour images. Models were asked to pose in three conditions. In the midline condition, the model’s body

Results

A score of emotional expressivity was gained by averaging responses across the four questions. Scores could range from one to six, with higher scores indicating higher levels of perceived emotional expressivity.

Table 1 shows the data obtained for each model across the six-posing/orientation conditions. An ANOVA with model, as a within participants factor revealed significant differences between the models in the ratings of emotional expressivity they received [F(5,1735)=161.6, P<0.001, η2

Discussion

The data show a clear effect of posing condition. Left hemiface and midline portraits were judged to be more emotionally expressive than were portraits that featured the right hemiface. The effect presumably reflects the fact that the left hemiface, which is dominant for the expression of emotion [1], was less prominent for right poses than for left or midline poses. One potential problem with the current study is that left, right and midline images were taken separately. It could therefore be

References (16)

There are more references available in the full text version of this article.

Cited by (73)

  • Lateralization of the expression of facial emotion in humans

    2018, Progress in Brain Research
    Citation Excerpt :

    Importantly, investigations using more natural stimuli, such as 3/4 view portrait photographs, confirm that perception of the left cheek's greater expressivity is not limited to the chimeric face paradigm. Nicholls et al. (2002a) presented participants with portrait photographs of models adopting a pleasant neutral expression and posing facing 15 degree to the left (right cheek pose), 15 degree to the right (left cheek pose), or straight-ahead (full face pose). Participants were asked to imagine the person in each photo and rate how emotionally expressive they were using four six-point Likert-scale items from Kring et al.’s (1994) Emotional Expressivity Scale (EES).

View all citing articles on Scopus
View full text