The Eyes Have It:
Eye movements and the debatable differences between original objects and reproductions.
Helen Saunderson, Alice Cruickshank and Eugene McSorley
In a large void of space, punctuated at eye level by a sliver of shelf, are displayed the Parthenon Marbles in the British Museum.[1]The Acropolis museum, meanwhile, displays dazzling white plaster versions of the ‘missing’ Parthenon sculptures nestled between the other, original,sculptures.[2]The legitimate location of the original marbles is a matter of longstanding ethical, political and intellectual debateand with the planned opening in late 2008 of the new Acropolis museum,the controversy will be re-ignited.The fact that such furious disagreement is still engendered by the Parthenon marbles providesa compelling anecdotal example of the many elements that people find important when considering artefacts and art (including legitimate ownership). Amidst the mix, is the significance of experiencing genuine articles as opposed to reproductions–where “reproductions are copies made for honest purposes”(Savage, 1963:1).
But why is viewing originals so important? A variety of potential reasons can be given, including theidea that the experience gleaned from the original is distinct from that of a reproduction. This possibility has been at the core ofa philosophical discussion about the relationship between original objects and their reproductions;in particular, of art works. An, especially pertinent debate in the British Journal of Aestheticsstarted with Zemach,who proposed that a“painting itself exists wherever there is a sufficiently good replica of that painting”(Zemach, 1986: 245/6), but his arguments were rebuffed in a flurry of papers (Levinson, 1987; Taylor, 1989)to which he replied robustly, (Zemach, 1989)and a temporary full stop was reached. A new statement reopened the argument some time later, when Farrelly-Jackson proposed that “one standard objection for treating reproductions of paintings as token of the very painting themselves is that… such reproductions will differ in certain aesthetically relevant properties.” (Farrelly-Jackson, 1997: 139)And so the debate remained unresolved.
Other literature, from a variety of disciplines, proposes that originals and reproductions or replicas are experienced in different ways. Benjamin’s (1936) classic article, describes original art works as having an ‘aura’ that relates to their unique history and authenticity. However, this relationship is not entirely straight forward as for certain art forms (e.g. printing and film) the distinction between a single ‘original’ and ‘reproduction’ is unclear. In addition, an extrapolation of museum studies discourse about the importance of multiple sensory experiences (Classen, 2005; Drobnick, 2005) also has interesting implications for the relationship between originals and reproductions. Reproductions of an art work that engages explicitly more than the visual sense will automatically be only a limited representation. For example, a photograph of part of the work Simply Boutiful by ChristophBuschel illustrating the freezer in the cargo box only fragmentally represents the experience of encountering the art work, which involved clambering into the freezer, listening to the noises of the person previously descended, and the anticipatory smell of the earth below. So, the diverse literature often identifies interesting differences and sometimes complex relationships between originals and reproductions.
In addition to compelling intellectual arguments, human actionsmayalso provide interesting evidence. One aspect of which is the value placed upon the original objects; for example, the monetary worth implies an importance, especially as a National Gallery print of the Van Goghs ‘Sunflowers’ currently costs £6, yet the actual painting is worth millions. Also, scholars will travel across the world to gain access to original manuscripts. Andrew Motion, the poet laureate, has said that he can’t “look at the original manuscript of a poem… still less actually put my hand on it, without feeling an extraordinary electrical charge is coming through me, and that’s something that can’t be replicated by things that are replicas” (Motion, 2007). Therefore, human behaviour implies that originals (in contrast to reproductions or representations), especially in certain contexts, are extremely valued.
How might an approach informed by the academic discipline of psychology enlighten our understandings of responses toart? In particular, can it be used to probe the differences, if any, between experience of originals and of reproductions? Experimental psychology can give a new perspective to the debate, using scientific methodology and techniques. Examples of such inter-disciplinary work include Leder, Belke, Oeberst & Augustin’s (2004) five step model of perceiving contemporary art work and Livingstone’s (2002) examination of the visual qualities of famous paintings e.g. Monets Impression Sunrise(1872).The following chapter considers a psychological experiment devised and undertaken by Eugene McSorley, Alice Cruickshank and Helen Saunderson, to explore the proposed differences in original and reproduction artwork.
The disciplineof psychology offers many investigative methods, four of which may have particular relevance to examining the experience of art:
1. Self-reports.Individuals provide subjective information by describing their own experience. As this method is shared, in various forms, by a number of other disciplines too, and as it is difficult to generalise from individual experiences to a common model, in itself it is unlikely to contribute to psychology’s provision of an innovative insight into the topic.
2. Measurement of human brain activity. Information about which areas of the brain are most active over time can be recorded. This can help to illustrate how and where the sensory information is processed. Functional Magnetic Resonance Imaging (fMRI)and Positron Emission Tomography (PET)both measure activity of the brain indirectly. Both provide images allowing comparison of areas which are active during different tasks and providing compelling images brains with coloured ‘active’ areas. Magnetoencephalography (MEG) and Electroencephalography (EEG) provide records of the accurate time-course of brain activation. Together these techniques can provide an accurate picture of brain activity. However, the problem with these methods is that knowing when and where inthe brain a signal is processed does not necessarily add to our understanding of how or why it is processed. (seeCaccioppo et al, 2003).
3. Analysing images of the art work. Work in this area has predominantly focused on the visual qualities of the art work (see Graham & Field, 2008;Redies, Hasenstein & Denzler, 2008; Wade, 2003; Zanker, 2004).Livingstone (2002) analysed Monet’s Impression Sunrise, 1872, examining two qualities of the picture: colour and luminance (the amount of light reflected or emitted from an object). Sheshowed that the sun was of equal luminance to the rest of the sky. Therefore, in a greyscale image the sun is invisible, it is only distinct in colour. Livingstone argues that “by making it [the sun] the exact same luminance as the sky, he [Monet] achieves an eerie effect.” (Livingstone, 2002: 39). However, it is clear that this analysis includes a degree of subjective self-report of Livingstone’s own experience of the picture. Deconstructing an art work in this way focuses solely upon the properties of the object, butthe interpretation may become subjective, and will not necessarily reflect entirely anindividual observer’s full experience when encountering the object.
4. Examiningobservers’ behaviour when viewing an artwork. There are many possible measures of behaviour, some easier to record than others. Measures such as body language or facial gesture require an experimenter to make a subjective judgement on the person’s status while measures such as duration spent in-front of an artwork, or visual behaviour can be measured more objectively. Recording eye movements can provide several measures of observers’ behaviour including: where an observer first looks; how long an observer looks at an area before moving their eyes; which areas are looked at most. The experiment undertaken used eye-movement recording, hence the discussions in this remainder of this chapter will focus on this technique.
Eye movements are essential to enable the gathering of information about the visual properties of an object or scene (Henderson, 2003). Take a moment to glance up from reading this book,try to identify which object in the scene stands out the most. It is likely that it will be the element that differs most from its surroundings in luminance, colour or relative contrast.That is, certain aspects of the scene will attract your attention, because of relational and inherent visual qualities. For example, when a visitor descends into the basement depths of the National Museum of Scotland in Edinburgh, a vision unfolds of the sections about the Early Scots people from 10,000 years ago.[3]In the scene, initial attention is drawn to the abstract statue figures: they are dark in contrast to the light sandstone backdrop, and within their structures are lanterns of light, where clear Perspex boxes containing Celtic jewellery are literallyspotlighted. The contrast, colour and brightness of various aspects of the figures in context are identified as the most salient aspects of the scene.In such examples as this, the visual qualities of a scene (or object) drive where the eyes look and therefore what information the brain receives: this is known as bottom-up processing.
In contrast, top-down processing is where prior knowledge and thoughts drive where and what is viewed. If, when visiting the Early Scots gallery,you were armed with the knowledge (i) that Sir Eduardo Paolozzi, who created the sculptures within which the jewellery lies,designed themas abstract in order to reflect our lack of knowledge of how the people of 10,000 years ago would have looked,[4] and (ii) that one of the sculptures has a huge sphere instead of a hand, you might then view the scene in order, say, to search for the ‘hand’ –hence to a large degree ignore the overall visual qualities of the scene, except for elements that might assist you finding the ‘hand’. By studying the eye movements of a person it is possible to detect the direct-end result of the combination of bottom-up and top-down processing, namely where and when we look.
Measures of eye movement are direct, objective and quantifiable, which is why this method was selected as the most appropriate for the investigation discussed here. It is possible to look at many different aspects of eye movements (e.g., size and direction of the movement) and times when the eye is relatively still (e.g., how long it is stationary, known as fixation duration) to gather different information about how an observer is viewing a scene. When looking at gaze position we record the location of the area of the eye that can process the most detail of the scene (the fovea) to determine what the observer is looking at. The fovea is relatively small, for example, whilst reading this sentence, you are able to bring into high focus only a few letters at a time, the surrounding area is relatively blurred. When you read, you are not particularly aware of this as your eyes are moved approximately four times per second through tiny fast movements,i.e., saccades,to another location. The brain processes the information into an experience that appears seamless.
Recording eye movements when people view both original and reproductionartworks has the potential to provide innovative knowledge. However, no-one (to the knowledge of the authors) has previously used eye-movement recordingto investigate this area.
The closest previous eye movement investigations in essence and relevanceare two classic psychological studies. Yarbus(1967) recorded eye movements and found that when displaying paintings of scenes, the largest influence on where and when the eyes looked was what task the viewer was given. For example, if instructed to work out the ages of the people in the picture, observers predominantly looked at the faces; if instructed to work out the wealth of the individuals, observers looked more at clothes and artefacts. As the intention of ourinvestigation, however, was to mirror real life experiences as closely as possible, it was important not to give instructions to the person; rather, the viewer looked at the objects with self directed purpose, as they usually do during a museum experience.
The importance of self, of individual differences, was also identified as key in the second classic study, by Buswell (1935). He recorded eye movements of 200 people looking at various images of aesthetic objects (including paintings, sculptures ceramics and architecture), onto photosensitive materialand subsequently analysed the resulting 18,000 feet of film. The results (including fixation durations) identified that the predominant factor influencing eye movements was differences between individuals. Clearly, this means that it is hard to compare one person to another. Therefore, in our investigation, individuals’ results were compared to themselves (a method known as a repeated measures design), so that when an individual viewed the original paintings the eye movements from that would be compared to those recorded when they viewed the photographic and monitor reproductions of the artworks.
Any difference between individual viewers could also be due to individual differences in art expertise. Would there indeed be a discrepancy if art experts and novices were compared? Thus when recruiting the 24 people who took part in the investigation, half were from the Fine Art Department at the University of Reading (designated as ‘art experts’) and the rest from the Psychology Department and other members of staff at the same University (‘art novices’).
All these people were shown sixoriginal paintings (on canvas), and reproductions on photographic paper and computer monitor display. The reason for choosing the two forms of reproductions was that in the field of art it is common for people to consume images of artworks as photographs or high quality prints in books, or sourced from the internet, or viewed as part of computer presentations. Therefore, the types of material selected for the reproductions arguably related to the dominant types of reproductions of art encountered and used on an everyday basis.
The original paintings werecontemporary semi-abstract designs with a limited palette of colours (fig 1). The paintings were professionally digitally photographed and printed onto photographic paper, and transferred to computer format. The original and reproduced images wereclearly different in visual and material qualities: in the painting the textured canvas background and the brush strokes were clearly visible. This isin contrast to the photographic reproductions which,although they were photographed to capture the texture,had a glossy sheen finish. The material differences also had an impact on the luminance of the images (measured in candelas per meter squared, cd/m2). The monitor reproductionshad the highest luminance (27.09 cd/m2), with the luminance of the original paintings being 16.51 cd/m2 andthe photographic reproductions, 15.81 cd/m2. Sinceprevious research suggests(e.g., Livingstone, 2002) that luminance ofobjects viewed has an impact, it seemed very likely that the monitor reproductions in particular may yield distinctly different results.
‘It’s like the machine in Clockwork Orange,’ commented one individual of the contraption used to detect and record observers’ eye movements. It was in fact the harmless helmet-like SR Research Eyelink 2, which works by shining a beam of infra-red into the eye, and detecting differences in how the infra-red light is reflected to tracks the location of the pupil of the eye. The ‘helmet’ also has a centrally mounted camera that picks up signals emitted by infra-red light sources on the display-monitor. This enables the eye tracker to take into account head movements when providing information about where the person was looking.
Each participant viewed the 18 images for 30 seconds. Monitor reproductions were presented on a flat computer monitor, original paintings and photographic reproductions were presented by being fixed to the screen. The order in which the type of images were shown, original painting, photographic or monitor reproduction, was varied, so that each of the six potential order combinations was adoptedfour times, in equal numbers for both art novices and experts. This was done to eliminate the possible effect on results of the order in which the images were presented, and/or the boredom and fatigue that can occur towards the end of an investigation.The images were shown to individuals sitting relatively comfortably. The viewing distance was the same for all observers. As images were changed over participantswere asked to shut their eyes, to prevent accidental previewing of the images.
When the data were analysed comparing the original paintings to the photographic and monitor reproductions, and art experts to art novices, no significant differences were identified. However, in relation to viewing time statistically significant discrepancies were found. The graph (fig 2) shows,that all viewers, viewing all image types, had shorter fixation durations in the first 15 seconds than in the last 15 seconds.For example,when looking at the original paintings the art experts had a mean fixation duration of 376ms during the 0-15 seconds period, and 445ms during the 15-30 seconds phase. This suggests that initially the eyes are spending relatively small amounts of time examining various large areas, and that as the viewing time goes on spend longer looking within one part of the image. For saccade amplitudes, the relationship was inverted (see fig 3): saccade amplitudes were initially larger, and later smaller. For example, whilst looking at the photographic reproduction, art novices initially had averages saccade amplitudes of 2.79°[5]and later 2.72° saccade amplitudes, while art experts had averages of 2.98° during the 0-15 seconds and 2.85° for 15 seconds and over.