Reference: Ventura, P., Liu, T. T., Cruz, F., Pereira, A., Domingues, M., Guerreiro, J. C., & Delgado, J. (2023). From Perugino to Picasso: Holistic processing of faces in paintings. Psychology of Aesthetics, Creativity, and the Arts.
Imagine sitting in an art gallery, looking at the most renowned painting in the exhibition. You absorb all the painting’s details: The masterful stroke, the perfect contrast between colors, and the delicate overlap of paint layers. None of these, on their own, depict anything specific. However, you know what the painting is intended to portray. The strokes, colors, and layers morph into people or landscapes. The painting illustrates something about the real world, but the elements in the painting are not found in the real world – rather, they are representations of the actual objects themselves.
A question underlies this realization: Is the way we process real-world stimuli the same way we process their representations in paintings? In recent research, I have addressed this question by exploring whether the way people process real faces and faces in paintings is the same.
Real faces vs. Faces in paintings: Similarities and differences
Though similar in many ways, actual faces and faces in paintings are not exactly interchangeable: Paintings of faces convey depth but are not three-dimensional like actual faces. Furthermore, paintings differ across artists and art styles, and the resemblance to actual faces varies substantially as a function of these factors. For example, Cubist paintings are heavily distorted and thus resemble real-world objects (such as faces) less than those in Renaissance paintings. These differences could suggest that the processes recruited for actual faces differ from those that faces in paintings engage. Following this, people describe and draw faces and bodies as the sum of their separate parts, which is inconsistent with how people process actual faces.
Face processing mechanisms
How do people process actual faces, then? The face processing literature argues that people represent faces as unified wholes, creating a joint representation of all facial features available – a strategy coined holistic processing. Holistic processing is the opposite of part-based processing, which would be described as attending to the different features of a face one at a time and representing them independently of one another. Extracting a global representation of a face is more efficient than going over each feature separately – and thus addresses the importance of facial stimuli in our environment, which requires that we set apart faces not only accurately, but also quickly.
Experimentally, there are several tasks designed to explore the holistic processing of faces. One of these is the part-whole task. In this task, people first see a target face. After its presentation, people are shown either a) two faces (“whole” trial), or b) two face features (e.g., noses; “part” trial) side by side (see example trials below). The participant’s goal is to select which of the faces they had been presented before (in “whole” trials), or which of the two features belonged to the previously presented faces (in the “part” trials). If faces are processed holistically, performance should be better when features are integrated into a whole face than when presented in isolation. In line with this hypothesis, research consistently finds that people’s performance for real faces is better for whole trials than for part trials.
Ventura and collaborators’ (2023) work
In our research, we extended this paradigm to faces in paintings to explore whether these representations of faces also elicit holistic processing. Adding to real faces (i.e., photographs), we used faces in paintings belonging to four different art styles, namely Renaissance, post-impressionism, expressionism, and Cubism (see below for example). Using different art styles was important because they differ in terms of realism, with the latter applying more distortion to the central elements that allow for the identification of a face as such. It could be that art styles that more closely resemble actual faces are processed holistically, while those that highly distort faces are not. For each of the categories mentioned above, four different faces were selected and used to create a total of 240 trials (48 trials per category; 24 whole and 24 part trials). Thirty Portuguese psychology undergraduates completed the experiment.
Overall, our results replicate previous work in face processing, such that participants had better accuracy performance in whole trials than in part trials – a marker of holistic processing. More importantly, this advantage for features embedded in faces (vs. features represented in isolation) was found across all art styles, with similar magnitude. This finding suggests that people not only process faces in paintings holistically (i.e., as they process actual faces), but that they do so regardless of the art style’s level of distortion.
Interestingly, the extent to which participants processed faces holistically was related to their interest in art. Participants who reported being more interested in art (as assessed by the Vienna Art Interest and Art Knowledge Questionnaire) displayed a marginally smaller advantage for whole (over part) trials. This outcome could represent the tendency for art aficionados to be more interested in details contained within paintings, which would prompt more part-based processing strategies. Conversely, people with better visual recognition abilities (as assessed by the Abstract Art Memory Test) displayed an increased difference in performance between whole and part trials, indicative of increased holistic processing. This outcome is congruent with the idea that holistic processing is important for efficient and accurate stimuli recognition.
Taken together, these results support the idea that faces in paintings are processed like actual faces – through aggregate representations of facial features, or holistic representations. This finding, however, is modulated by individual characteristics, such as general recognition abilities and interest in art.
Just something to think about next time you go to an art gallery!
Image Credits
Featured Image: Photo by Uğurcan Özmen, under Creative Commons License CC0. https://www.pexels.com/photo/people-looking-at-painting-in-art-gallery-16842473/
Part-whole trial example: Original, based on pictures by Barry Langdon-Lassagne, under Creative Commons License 3.0. https://commons.wikimedia.org/wiki/File:Caring_Face.jpg and https://commons.wikimedia.org/wiki/File:Desirable_Face.jpg
Portraits examples (left to right): Images of paintings, under public domain or fair use licensing. https://www.wikiart.org/en/giovanni-antonio-boltraffio/portrait-of-a-lady-as-st-lucy-1500, https://www.wikiart.org/en/paul-cezanne/portrait-of-madame-cezanne, https://www.wikiart.org/en/kees-van-dongen/woman-with-cherries-on-her-hat-1905, and https://www.wikiart.org/en/pablo-picasso/woman-by-the-window-1936
cogbites 
