Abstract

An eye-tracking and questionnaire study was set up in collaboration with the Van Abbemuseum (Eindhoven, The Netherlands) to investigate the perception and appreciation of three Frank Stella paintings from the 60s (Tuxedo Park and Effingham I from the collection of the museum and a hand-painted replica of Hiraqla Variation II). Effingham and Hiraqla were shown next to a printed copy without fluorescent colors, for a direct comparison between the two versions. The main purpose of the study was to assess whether the works were experienced according to Stella’s prescriptions as defined in his Modernist ‘logic’: all-overness, flatness, instantaneousness and self-referentiality. We found that the perception of Tuxedo resulted in a well-structured, coherent heatmap, while a more or less even distribution of fixations over the surface was found in the case of Effingham and Hiraqla (and their copies), which indicates that Stella’s target of all-overness was achieved better in the last two works. Although Stella claimed to have created “flat and frontal” paintings, depth was experienced, especially in Tuxedo and the Hiraqla replica. In the latter, this was mainly caused by the protruding fluorescent colors. Also, in this work more fixations were found in fluorescent-colored areas when corrected for area size. No such effect was found in the original Effingham painting. Most participants found only Effingham to be instantaneously capturable. In the case of Tuxedo, the specific material qualities, like alkyd and open canvas, were rarely recognized, which undermines Stella’s aim for self-referentiality. Participants noticed the fluorescent effect in the Hiraqla replica, but they did not mention other material qualities. A reverse effect was found for Effingham.

Abstract

Frank Stella’s early works tend to be characterised as displaying the flatness of painted surfaces and an ambition to negate pictorial illusionism. However, beyond their emphasis on flatness, these early series of paintings generate new forms of illusions and, in some cases, initiate another type of pictorial space — one that bodies forth, coming toward the viewer, appearing as if in front of the canvas. We consider the materials of the painting format in Stella’s early work (1959 to 1986) that create or facilitate the emergence of such a protruding or ‘projective’ space: mainly canvas, types and colours of paint. After introducing notions of flatness and illusionism and our respective approaches, we focus on Stella’s use of unprimed, raw, canvas, on the one hand, and his use of reflective and fluorescent paint skins, on the other, and how paint and canvas relate to each other. We focus on the material conditions that Stella sets up to manifest his intentions regarding the perception of space in painting and where he believes painting ‘should’ go. Indeed, in a book published in 1986, Stella describes projective effects from painters who use different tactics than his, but he does not reveal how he achieves his own. We analyse precisely which elements in Stella’s early paintings trick the eye of the viewer into seeing a painting, as it were, in front of itself, and we demonstrate the aesthetic impact of Stella’s chosen materials. Or how colour, paint and canvas, working together in a sort of symbiosis, generate a protruding effect in a new, previously unseen manner, and challenge Stella’s assertions against illusionism.

Although Frank Stella intended to create flat, illusion-less Irregular Polygons paintings, it is not uncommon to experience the illusion of colour depth, based on the interaction between their fluorescent and conventional colours. Some critics praised these artworks’ flatness, while others described odd depth experiences that they categorised as a new kind of illusion. In order to provide a correct reading of these works and to reassess their art historical significance, a scientific case study regarding this colour-depth effect imposes itself. This article discusses an experiment in which we determined whether twenty artists, twenty art historians and twenty laypeople experienced fluorescent colours as protruding, receding or flat in combination with conventional colours. We additionally looked at whether they still perceived colour depth when all fluorescent colours were replaced with their conventional variants. All participants observed fifteen designs, which they had to rate according to the perceived depth of each coloured region with a number between −3 (strongly receding) and +3 (strongly protruding). The results revealed that most participants experience fluorescent regions as strongly protruding, unlike all conventional colours, which were rated as much less protruding. When a fluorescent colour was swapped with a conventional variant, all participants experienced significantly less depth. The differences between the subject groups were statistically negligible when looking at the mean depth ratings for both colour types. However, we discovered that artists experienced more contrast effects, as they gave different ratings to different panels (of identical colour and shape) in the same design, depending on their position.

In his search to create ‘instantaneously capturable’ paintings, Frank Stella started to use Day-Glo alkyd paints as a vehicle to communicate his simple, striped designs. Up till now, art criticism has neglected the visual impact of these fluorescent colours on this concept of ‘instantaneous art’. By presenting participants with Stella’s designs (fluorescent and conventional variants) for short presentation times (8 to 12 ms), we aimed to find out whether fluorescent colour combinations are seen faster (i.e., yield better performance in identifying the specific design) than their conventional counterparts. In general, participants were very good in identifying the correct design among distractors, which means that the pattern and colour combinations based on Stella’s work do seem to be ‘instantaneously capturable’. However, Stella’s formula for ‘instantaneous’ paintings is not identical for the different combinations. When exploring fluorescence in combination with other aspects of the design (colour and pattern), we found two effects that seemed to predict performance. First, performance seemed to depend on specific design patterns. Second, fluorescence seemed to interact with specific colour combinations in predicting performance. The red/yellow designs yielded better performance for the fluorescent variants, while the opposite was found for the green/orange designs. Contrast differences in luminance between the two colours of each colour combination might explain part of the results. On the other hand, the effect of fluorescent colours might have been watered down by the confusion between the hand-printed fluorescent colours and the computer display used for the identification task, which only showed conventional colours.