Perception of abstract and representative visual art

Transcription

1 Perception of abstract and representative visual art Liisa Uusitalo 1, Jaana Simola 2 and Jarmo Kuisma 3 1 Professor, Helsinki School of Economics HSE, Finland 2 Doctoral student, University of Helsinki 3 Doctoral student, Helsinki School of Economics Liisa Uusitalo is professor of marketing at the Helsinki School of Economics. Her special fields are consumer behavior, marketing communication, and the social and cultural aspects of economic behavior. She has been visiting scholar at UC Berkeley, Stanford University, Science Center Berlin and the Institute for Advanced Study Berlin. Research projects cover issues on consumer rationality, environmental behavior, and cultural marketing. The present research project is Consumer behavior in information economy, in which the two co-authors act as researchers. Jaana Simola is experimental psychologist specialized in research on consumer perceptual and cognitive processing and learning. Jarmo Kuisma is marketing scholar specialized in perception of advertising and the eye measurement techniques. ABSTRACT How do we look at contemporary artwork, and how do we construct our preference toward them? Research on consumer art perception has usually been based on interviews, but the rapidly growing area of neuropsychological behavioral research neuro-economics and corresponding experimental research methodology open new ways to study human art perception. In this paper, we explore perception of contemporary paintings using preference and affective scales but also by measuring viewer s eye movements. We are interested in finding out differences between perception of abstract and representative art. The results show that non-professional art viewers prefer more representative than abstract art, and that representative paintings were evaluated higher on emotional valence, arousal and dominance. Also, preference and affective evaluations are closely associated. However, contrary to our expectations, no remarkable difference was found in the way spectators look at abstract and representative paintings according to the total number and duration of eye fixations. Keywords: art perception, preference, affect, visual attention, eye movement THEORETICAL BACKGROUND In marketing literature we find studies on the perception of art organizations such as image of art museums (Ahola, 2008; Uusitalo & Ahola, 2008) and art gallery venues, festivals and theater performances (Oullet, Savard & Colbert, 2008; Bouder-Bailler & Damark, 2005; Kerttula, 1988). These studies often rely on brand and brand personality theories, or on the influence of emotionally appealing and atmospheric factors on consumer perception of products, services and their sales venues. However, brand theory is less applicable in explaining consumer basic perception processes of visual artworks in situations when the consumer sees them for the first time, as is often the case for new contemporary art. Also in psychological and communication literature art perception has rarely been studied although the perceptual processes as such have received much attention as a first stage of information processing and memory effects. Perception of contemporary art has received very little attention. Lagier and Godey (2007) studied dimensions of consumer evaluation of art and luxury products, and Geissler Zinkman and Watson (2006) the impact of homepage complexity (showing that intermediate level of complexity had 1

2 best response). In design aesthetics, the psychological Gestalt theories of proximity, connectedness etc. have helped to improve modeling of information graphics aimed at consumers (Wattenberg & Fisher, 2004), but there are no studies how these apply to consumer art perception. Takahishi (1995) distinguishes two parallel theoretical paths of research on art perception. The first path, to which also this paper more or less belongs, is experimental psychology of aesthetics, which often focuses on the art object, and points out visual factors as the determinant of aesthetic effects such as pleasure and arousal caused by the painting (Berlyne, 1974; Parker & Derekowski,1990). The other path relies more on the philosophy of the aesthetic experience, and it builds on the interrelation between the person s mind and art objects (Benson, 1993; Goodman, 1976; Wollheim, 1974). We also recognize that the observers have internal intentions in art consumption that possibly affect how art is perceived and evaluated. These intentions may vary, for example, between pure stimulation seeking and cognitive attempts to understand and interpret the content of the artwork. For example, several communication studies show that the various motives of viewing television programs or advertising correspond to the criteria along which the viewers structure and evaluate the content (Eco, 1972; Uusitalo 1977, 1978). In other words, various motivations and functions of art to the consumer, such as intellectual stimulation, sensual aesthetic satisfaction, or providing socially meaningful contents for identification, may affect the way he or she looks at the artwork. Since we aim at studying the differences between representative and abstract art perception, the following aspects need more verification. It has been claimed by empirically-oriented, explorative research that pictures can be interpreted as metaphors of both meaning and mood, in other words, cognitive meanings and affective feelings (e.g. Marks, 1983). For example Takahishi (1995) studied how abstract drawings were able to communicate emotions and moods, and whether there existed any congruence between the painter s and spectator s feelings as related to specific forms and figures. This is an important issue for communication theory in general, not only for understanding the communication between artists and their audiences. One interesting result was that affective coding and interpretation of connotative meanings are interrelated. Also consumer researchers have recently emphasized the importance of somatic, bodily experience in experiencing art (Lakoff & Johnson 1999; Joy & Sherry 2003; Ahola 2007). Moreover, research in neuropsychology shows that affective and cognitive processing of art influence each other (e.g. Cohen & Areni, 1991). Consequently, earlier finding gives us the reason to assume that affective reactions and evaluative, cognitive responses to artworks are interrelated. As to the viewing process, Yarbus (1967) showed that observers tend to look at those parts of the visual scene that contain high contrast and fine details as well as items of biological significance and meaning (e.g. other humans). Importantly, he also showed that the intention of the observer plays a crucial role in determining where the eye fixations concentrate. To isolate the factor of intention, he selected a painting by the Russian artist Ilya Repin ( ) and asked different questions regarding the content of the painting. The results showed that the patterns of observers eye fixations, the scan paths, differed considerably as a function of the externally given task. Some studies show that viewer s gaze tends to draw toward contextually relevant or the high resolution and high contrast parts of the painting. However, the most frequently explored part of visual art is the center. Kaufman and Richard (1969) showed that more time is spent at the center of gravity of a visual scene, which is determined by estimating the perceptual weight of each element in the scene. Although it is controversial as to how perceptual weight should be determined in the context of art, a center weighting seems to be a general rule that applies to all visual perception (Solso, 1994). Many artists are aware of the natural tendency of viewers to focus more on central features than peripheral ones. 2

3 According to Berlyne (1971) aesthetic evaluation can be based on two types of perceptual explorations: a diversive and a specific exploration style. Both exploration styles have their distinctive ocular scan paths. The diversive exploration is characterized by a hunt for stimulation without regard for content, and consequently, typical scan paths are widely dispersed with sparsely populated clusters of fixations. In the specific exploration mode, incomplete knowledge seeks out specific information, and the eye fixations are concentrated in locations that are relevant for content processing. Previous research also suggests that different styles and periods of art may produce different kinds of perceptual explorations. For example, viewing of Reneissance and Baroque paintings elicited different viewing behaviors (Solso, 1994). The classical art produced large and slow eye movements while Baroque produced small and quick eye movements. The mean duration of eye fixations was about 60 ms shorter for Baroque art. We can assume that it is not the art style itself but rather the complexity of the paintings that has an effect on viewing. A general rule suggests that less complex art which the consumer can more easily understand elicits large and slow eye movements, while viewing of complex art is characterized by small saccades and short fixations (Solso,1994). In line with this we can assume that abstract art, on the average, is possibly perceived by laypersons as more complex than representational art, leading to a diversive mode of exploration and shorter eye fixations. The matter is somewhat difficult to study, because it has also been proposed that the scan paths of the art consumer may differ between initial and subsequent viewings (Locher and Nodine, 1987). Widely dispersed eye fixations, suggesting a diversive exploration type, are said to be typical for the initial viewing. As the viewing continues, the proportion of longer fixations increases and the specific exploration mode becomes more common. GOALS AND HYPOTHESES Previous studies on visual art perception have mainly focused on explaining how the physiology of the human visual system affects art perception (e.g. Livingstone, 2002), while less attention has been paid on the influence of subjective intentions, affect and art preferences. Also, the research has mainly dealt with older art styles not including contemporary art or abstract art. Therefore, the first aim of the present article is to explore the differences in the affect elicited by abstract and representational contemporary artworks. These two art styles differ considerably in composition and content, and hence we expect them to evoke different affective states. Preferences are treated as intervening factors in art perception. General art style preferences can influence perception of art, but in this article we investigate the relationship between artwork preference and affect. Based on earlier research, we can make a hypothesis that the strength of the affect elicited by the artwork is associated with the cognitive processes of evaluation and preference formation. The last main hypothesis is that that abstract and representative art are viewed differently. The diversive, scanning or glancing style with widely dispersed and short fixations will dominate the exploration of abstract art, and the specific style and longer fixations will dominate the viewing of representative art. This is because in the latter type of art, eye fixations most likely concentrate on locations that are relevant for interpreting the content of the painting. Moreover, we do some qualitative analysis based on the eye fixation results that relate to earlier findings that viewing tends to concentrate on central parts of the scene instead of peripheral ones, and, in the case of representative art, whether human or other biologically meaningful items tend to receive most attention. 3

4 The main hypotheses are summarized as follows: 1. The affective responses to abstract and representative art differ in their intensity. 2. The affective responses (emotions) and preference (cognitive evaluation) of an artwork are correlated. 3. The perceptual process of abstract and representative art are different; abstract art tends to elicit a diversive, glancing mode of exploration and dispersed, short fixations, and representative art a specific mode of exploration and long fixations. METHODS Participants A sample of 32 volunteers (56 % female, 44 % male) representing an average middle-aged population (mean age 38.2) in Finland participated in the experiment. Their reported art gallery visits varied from 0 to 24 per year, with an average of 6.3 visits in a year. Artworks The artworks used in the study consisted of 40 recently created, contemporary paintings selected from the artists listed by Galerie Forsblom ( Helsinki, Finland. In order to include both abstract and representative artwork in the sample, four persons representing the same population but not taking part in the experiment rated the paintings as representative or abstract. About half of the selected paintings were rated as abstract and half as representative in the selection stage. The final categorization into abstract vs. representative art was defined on the basis of the participants evaluation (on a 9-point scale definitely representative definitely abstract). Research procedure During the initial viewing phase, each painting was shown on a 22 inch screen for ten seconds, while participants eye movements were recorded with a Tobii 120X remote eye-tracking system. After each painting, a gray screen with a fixation cross in the middle was shown for three seconds in order to control the position of the initial eye fixation on the paintings. Before the first viewing phase, the eye-tracking system was calibrated for each participant by requiring participants to fix their gaze to five points covering the whole screen area. The viewing distance was approximately 70 cm. In the second phase of the study, the same paintings were shown on a 15 inch screen using Psychophysics Toolbox run by Matlab. After each painting, participants were asked to rate on a scale from 1 to 9 whether the artwork was abstract or representative and whether they liked it or not (preference). Also, after each painting the participants rated their affective state using the SAM non-verbal scales (Appendix 1). The participants could proceed in a self-paced manner. The valence scale consisted of nine graphic depictions of human faces in expressions ranging from a severe frown (most negative) to a broad smile (most positive). In arousal scale, there were nine graphical characters varying from a state of low visceral agitation to a state of high visceral agitation. The dominance scale consisted of nine characters varying in size from a very small figure (low feeling of dominance) to a very large figure (high dominance). The ratings were given by typing in a number corresponding to an appropriate figure. The scales were modified versions of the Self-Assessment-Manikin scales developed by Lang (1980). In order to exclude the effects of the order of presentation, the paintings were presented in randomized order during both viewings. 4

5 RESULTS Classification of paintings as representative or abstract by participants Participants first rated the paintings as representative or abstract. The ratings for paintings considered as abstract were more consistent than the ratings for the representative paintings as indicated by the higher variance (around 4) for representative as compared to abstract rating scores (under 2) (Figure 1). Moreover, the rating scores for representative and abstract paintings differed significantly [t(31) = 18.18, p =.000]. This refers to the fact that in contemporary art, the dividing line between abstract and nonabstract, representative art is becoming rather blurred. Through this article, the classification to representative (score lower than 5.07, mean SD) and abstract (score higher than 5.29, mean SD) given by the participants is used when examining other relationships. Variance of classifications painting no. Figure 1. The classification of paintings as representative or abstract Score 1 denotes that a painting is considered as very representative and 9 denotes a very abstract score. The upper line indicates the score values and the lower line is the variance of scores across participants (n = 32). The paintings are sorted according to the scores on the representative-abstract dimension. Preference of artworks The most preferred representative painting was a harmonious landscape and the most preferred abstract painting was a rather symmetric and balanced painting (Figure 2). 5

6 Figure 2. The most preferred representative and abstract painting On average, the paintings rated as representative were more preferred than the abstract paintings as shown in Figure 3. This was indicated by a significant difference in preference scores [t(31) = 3.15, p =.004]. Neither age nor gender affected the preference for the two type of paintings. However, the number of art gallery visits per year showed a trend toward increasing the preference of abstract paintings [r =.312, p =.08]. This refers to the fact that acquired cultural competences play a role in art preference; the more familiar the art and artist are to the consumer and the more s/he knows about them, the higher satisfaction the art gives. Affective state According to our first hypothesis the affective response to abstract and representative art differ, and according to the second hypothesis, the immediate affect for an artwork is associated with its preference. We found support for both hypotheses from the SAM scale results. Figure 3 shows that representative paintings scored higher on all affective state dimensions (valence, arousal and dominance) than the abstract paintings. The differences are also indicated by paired samples t-tests: valence [t(31) = 2.01, p =.053], arousal [t(31) = 3.68, p =.001] and dominance [t(31) = 2.26, p =.031]. The association between affect and preference is also clear. According to bivariate correlation scores, preference for representative paintings was correlated significantly with valence [r =. 812, p =.000], arousal [r =.430, p =.014] and dominance ratings [r =.396, p =.025]. Also, preference for abstract paintings correlated significantly with affective state evaluations: valence [r =.828, p =.000], arousal [r =.375, p =.035] and dominance [r =.497, p =.004]. It should be noted that the affective state ratings as such were strongly correlated with each other. Participants also expressed some difficulty to distinguish between the various affect scales, especially the dominance scale remained somewhat obscure. The strong association between preference and affective states can be interpreted in several ways. 6

7 rating score Preferred, familiar art elicits high emotions, or the art is preferred because of the positive emotions it raises. Anyhow, we can conclude that the cognitive and affective processes are interdependent and together important in forming a final art experience representative abstract preference valence arousal dominance Figure 3. Preference and various affective responses of representative and abstract art Preferences measure by 9-point scale and affect by 9-point SAM scale. The way of viewing art: eye movement results Due to calibration inaccuracies eight participants were excluded from the eye movement analyses. Our third hypothesis expected that representative and abstract paintings would differ in perceptual process and that shorter fixations would be more common for abstract than for representative art. However, the total number of eye fixations (representative: mean SD, abstract mean SD) and the total fixation durations in seconds (representative: mean SD, abstract mean SD) did not differ between abstract and representative paintings. Thus, the total fixation counts and fixation durations did not indicate differences in perceptual process between representative and abstract paintings. Figure 4 gives examples of the eye fixation results. Explorative analysis of the eye fixations revealed that the gaze in representative paintings was mostly directed to identifiable figures (heads, texts, points forming the center of gravity) and therefore the fixations were more not less dispersed across the screen than in the case of abstract paintings. In abstract paintings, the eye fixations were mostly focused in the middle, perhaps indicating that without clearly identifiable figures, the viewers are not interested in the peripheral parts of the painting (or the peripheral parts of the paintings that contain less details can be identified with peripheral vision). 7

8 Figure 4. Examples of eye movements in abstract and representative paintings The first column shows the original paintings. The second column shows the total eye fixation frequency of all participants, where red means many fixations within the same area. The third column denotes the eye scan paths of one representative participant. CONCLUSIONS According to our results, the preference of artwork and the intensity of emotions it elicits differ for representative and abstract art, which supports the first hypothesis. Representative art is significantly more preferred than the abstract one, and it raises higher emotions measured by SAM scale. In classifying the artworks, we also found that that it is characteristic of contemporary art that it may be difficult to categorize it either as representative or abstract when it includes identifiable items. Consumers can easier distinguish clearly abstract paintings, whereas there is a higher variation among representative paintings as to whether they are perceived as abstract or representative. To measure emotional affect states we used the well-known SAM-scale. However, it is doubtful how well the different affects (valence, arousal and dominance) can be separated by art consumers. The results show that all three dimensions were positively correlated. Especially the distinction between arousal and dominance scales can be difficult to consumers. The dominance SAM-scale may work better in gallery or museum surroundings when consumers see the paintings in their natural size. Art preference was found to be associated with the affect it elicits, which supports our second hypotheses, as well as earlier findings that affective and cognitive, evaluative visual processes are interrelated, and the art experience is formed in their interaction. The correlation between preference and affect was high for all three dimension of affect, which is not surprising because of their high internal correlations. Contrary to our expectations, viewing abstract and representative art do not notably differ from each other measured by the total number and duration of eye fixations. Explorative analysis of eye movement dispersion over the paintings showed that viewing abstract art seems to be even more concentrated to the 8

9 center than viewing representative art. Therefore the third hypothesis is rejected. One possible reason for the incongruence with earlier research (Yarbus, 1967; Berlyne 1971) is that all our paintings represented contemporary art in which the difference between abstract and representative is not so clear. The eye measurement technique can give a contribution to art perception research by concretely showing the most frequent fixation points in a painting, both for individuals and for the whole sample of viewers. ACKNOWLEDGEMENTS The authors are grateful to Gallerie Forsblom, Helsinki for providing the artworks for the experiment. 9

10 REFERENCES Ahola, E-K Art museum image an interplay of consumer and museum characteristics. In L. Uusitalo, ed. Museum and visual art markets. Helsinki: Helsinki School of Economics B-96, Ahola, E-K Producing experience in marketplace encounters: A study of consumption experiences in art exhibitions and trade fairs. Dissertation. Helsinki: Helsinki School of Economics A-299. Benson, C The absorbed self: Pragmatism, psychology and aesthetic experience. London: Harvester Wheatsheaf. Berlyne, D. E. (1971), Aesthetics and psychobiology. New York: Appleton-Century-Crofts. Bouder-Pailler, D. and L. Damark Quelle est l ínfluence du design de l espage de diffusion de l art sur les attitudes et comportementsdes consommateurs. In Proceedingsof the 8th International Conference on Arts and Cultural Management, 3-6 July. F-Colbert, ed. Montreal: Presses HEC. Bradley, M. M. and P. J. Lang "Measuring emotion: The Self-Assessment Manikin and the semantic differential". Journal of Behavior Therapy & Experimental Psychiatry, 25, Cohen, Joel B and Areni, Charles S "Affect and consumer behavior". In Robertson, T. & H. Kassarjian, eds. Handbook of consumer behavior. Englewood Cliffs, New Jersey: Prentice Hall, Eco, Umberto "Toward a semiotic inquiry into the television message". Cultural Studies 3 (August 1972), Birmingham University. Geissler, L.G., Zinhan, G.M. and R.T. Watson The influence of home page complexity on consumer attention, attitudes, and purchase intent. Journal of Advertising 35, 2, Goodman, N Languages of art: An approach to a theory of symbols. Indianapolis, In Hackett. Joy, A. and J.F. Sherry Speaking of art as embodied imagination. a multisensory approach to understanding aesthetic experience. Journal of Consumer Research 30, September 2003, Kaufman, L., & Richards, W. (1969). Spontaneous fixation tendencies for visual forms. Perception & Psychophysics, 5, Kerttula, R Teatteriesitys kommunikaatiotapahtumana (Theatre performance as communication). Helsinki School of Economics D-107. Lagier, J.and B. Godey A scale for measuring aesthetic style in the field of luxury and art products. International Journal of Arts Management 9, 2, Lang, P. J. (1980). Behavioral treatment and bio-behavioral assessment: computer applications. In J. B. Sidowski, J. H. Johnson & T. A. Williams (Eds.), Technology in mental health care delivery systems (pp ). Norwood, NJ: Ablex. Lakoff, G. and M. Johnson Philosophy in the flesh: The embodied mind and its challenge to Western thought. New York: Basic. Livingstone, M. (2002). Vision and Art: the biology of seeing. Harry N. Abrams. 10

11 Locher, P. J. and C. F. Nodine "Symmetry catches the eye". In J. K. O'Regan and A. Levy-Schoen, Eds. Eye movements: From Physiology to Cognition, North-Holland: Elsevier Science Publishers. Nodine, C. F., P. J. Locher, and E. A. Krupinski (1993), "The role of formal art training on perception and aesthetic judgement of art compositions". Leonardo 26, Marks, L.E Categories of perceptual experience. In F.J.M. Buffart and E.L.J. Leeuwenberg, eds.modern issues in perception. Part 2. Amsterda: North-Holland, Oullet, J-F. Savard, M-A. and F. Colbert The personality of performing art venues. International. International Journal of Arts Management 10, 3, Solso, R. L Cognition and the visual arts. Cambridge, Massachusetts: MIT Press / Bradford Books. Takahashi, S Aesthetic properties of pictorial perception. Psychological Review 102, 4, Uusitalo, L. and E-K. Ahola Cultural struggles and theimage of art museum. In: Uusitalo, L. (ed.), Museum and visual art markets. Helsinki: Helsinki School of Economics B-96, Uusitalo, Liisa Consumer perception and preferences of message structure. Management Research News 1: 2. Uusitalo, Liisa Consumer perception and preferences of message structure. Helsinki School of Economics B-21. Wattenberg, M. and D. Fisher Analyzing perceptual organization in information graphics. Information Visualization 3, Wollheim, R On art and the mind. Cambridge, MA: Cambridge University press. Yarbus, A. L Eye movements and vision. (L. A. Riggs, Trans.). New York: Plenum Press. 11

12 Appendix 1. The 9-point Self Assessment Manikin (SAM) scales used in the study. The a) valence, b) arousal and c) dominance scales are modified versions of the original scales developed by Lang a) b) c) 12