Tom 44 (2022): Zeszyt 1-2 (August 2022)

The Mona Lisa (1503–6) is probably the most celebrated example of ambiguous expression in art. Soranzo and Newberry (2015) demonstrated that a similar ambiguity can be perceived also in La Bella Principessa (1495–6), another portrait credited to Leonardo da Vinci (1452–1519) by many. The paper aims to show that an ambiguous expression can be perceived in a further painting attributed (although not unanimously) to Leonardo: The Lady with Dishevelled Hair, or La Scapigliata. An experiment was conducted whereby participants rated on a 7-point Likert scale the perceived level of contentment of La Scapigliata and that of a comparable painting created by Andrea di Cione, alias Il Verrocchio. The two artworks were presented in random order to two groups of participants. One group could see the artworks from Close (0.6m) whilst the other group from Far (6m) from a Close (0.6 m) or Far (6 m) condition. Results show that the change of distance affected the perceived level of contentment of Leonardo’s figure but not that of Verrocchio’s. Specifically, whilst both artworks received similar ratings of contentment from the close-up condition, La Scapigliata was perceived to be more content from afar. It is concluded that La Scapigliata exhibits an ambiguous expression, and that this ambiguity is similar to the one observed in the Mona Lisa and La Bella Principessa. This result can be only partially interpreted within the spatial frequency hypothesis advanced by Livingstone (2000) and shows that a phenomenological account of Leonardo’s work might be more suited to capture the full extent of the phenomenon. Specifically, it is suggested that the principles of perceptual belongingness (Wertheimer, 1923) may need to be considered to fully capture the extent of the ambiguity depicted by Leonardo.

Prior to the discovery of linear perspective in the fifteenth century, European artists based their compositions more on imagination than the direct observation of nature. Medieval paintings, therefore, can be thought of as ‘mental projections’ of space rather than optical projections, and were sometimes regarded as ‘primitive’ by historians as they lacked the spatial consistency of later works based on the rules of linear perspective.

There are noticeable differences in the way objects are depicted in paintings of the different periods. For example, human figures in pre-linear perspective works often vary greatly in size in ways that are not consistent with the laws of optics. Some art historians have attributed this to ‘hierarchical scaling’ in which larger figures have greater narrative significance. But there are examples of paintings that contradict this explanation.

In this paper we will consider an alternative to the hierarchical scaling hypothesis: that medieval artists used relative size to elicit empathy and to reflect the perceptual structure of imagination. This hypothesis was first proposed by the art historian Oskar Wulff, but has largely been dismissed since. We argue that artists of this period, far from being naïve, used sophisticated techniques for directing the attention of the viewer to a particular figure in a painting and encouraging them to ‘see’ the depicted space from that figure’s point of view. We offer some experimental evidence in support of this hypothesis and suggest that the way artists have depicted space in paintings has an important bearing on how we imagine and perceive visual space.

In 1993, Agostini and Proffitt showed that perceptual belongingness (the subsumption of some sets of elements into a perceived whole) causes simultaneous lightness contrast to be seen in configurations in which the inducing elements are not adjacent to the target. The aim of the present research was to measure the strength of belongingness in determining the contrast phenomenon when the numbers of the inducing and induced elements and their relative positions are manipulated in Agostini-and-Proffitt-type configurations. In the first experiment, by using a forced choice paradigm, naïve observers indicated which gray disks arranged to form the letter T in two rows (organized with black/white inducers) appeared lighter/darker. In the second experiment, expert observers performed two nulling tasks: 1) the lightness of gray disk(s) was adjusted until it was perceived equal to that of gray target(s) aligned with white/black inducers; 2) the lightness of target(s) organized with white/black inducers was adjusted to match the target(s) organized with black/ white inducers. We found that also when there are few inducers, perceptual belongingness causes the contrast effect to propagate spatially on all the induced elements. Spatial position does not influence the induction effect. Low-level theories cannot account for these phenomena, but higher-level processes must be factored in to explain them.

We ran a pilot experiment to explore, using a new psychophysical method, the hypothesis proposed by Zietz and Werner in the ’30s, that a sound presented simultaneously with an afterimage can change its phenomenal appearance in non-synaesthetes. The method we adopted is able to directly collect and visualise the apparent changes in intensity of the afterimages, by recording observers’ interactions with a physical feedback mechanism (the paths that the observers generated by moving a cursor), without referring to verbal descriptions. These first findings support some of the most meaningful observations reported by Werner (1934) and Zietz (1931), according to which the colours of the afterimages ‘disintegrate’ at the hearing of a low sound and ‘concentrate’ for a high sound. This relationship is particularly evident with the Yellow stimulus, where the perceived colour intensity of its afterimage seems to have a faster negative change with a low-pitched tone sound, and an increase in intensity and duration when perceived simultaneously with a soprano sound. These data are also coherent with the crossmodal correspondences between both pitch and loudness in audition and lightness and brightness in vision reported in the literature.

Even today, film restoration (both photographic and cinematographic) is a challenge, because it involves multidisciplinary competences: from analogue film inspection and conservation to digitisation and image enhancement. In this context, thanks to the high manageability of digital files, the film restoration workflow often follows a digitisation step, which presents many approximations and issues that are often ignored. In this work, we propose an alternative approach to the issues commonly encountered in film restoration (mainly concerning colour and contrast restoration) aiming at restoring the original colour appearance, through models of human colour perception.