KramerBressanPerception2009.doc Clear waters, murky waters: why transparency perception is good for you and underconstrained

Nakayama and Shimojo mention that “transparent surfaces are fairly uncommon in our environment”. One encounters mist and vapors much more often, and transparency perception could have evolved to identify objects through such atmospheric disturbances. However, there is one particular substance that has a transparent surface and is critical to our survival: water. Transparency perception could have evolved to help us determine whether water is clear and drinkable or murky and unhealthy. Nakayama et al (1990) argued that human transparency perception does not take all information into account that is available to it. Their Plate 1B shows a cross with a red inner part and green outer parts. When the red part is given a positive disparity, a red transparent disk is perceived in front of an opaque cross (see also Metzger 1955). If the cross were seen as uniformly green, as claimed by the authors, this would indeed be rather odd, because natural filters seem incapable of attenuating green in such a way that it can appear red. In principle, though, it is also possible that the cross is not perceived as uniform in colour. Underneath the red transparent surface, for example, the cross might be white (or even red), despite that elsewhere it looks green. In this case, the change i