Imagine if you couldn’t recognise a person’s face without getting close enough and at the right angle. This would be a very disadvantageous way for humans to perceive the world. We need to be able to communicate with each other, and in the Environment of Evolutionary Adaptation (EEA) nonverbal communication must have been in frequent use, otherwise it would not have carried on into modern humans. For example, a quick raising of the eyebrows toward another person is a sign of greeting and affection (Allyn and Bacon, 2002).
So, being duped by innate perceptions such as the hollow face illusion would be better than being duped into seeing real faces as things that are not, but how does this illusion work? (To have a look at the illusion click here: http://www.youtube.com/watch?v=Rc6LRxjqzkA)
The hollow face illusion still hasn’t been precisely explained, but we know that it has something to do with the human perception of faces. It has been found that when the hollow mask is turned upside down, the effect dwindles, as it does when the face is replaced with hollow, non-objects. Hill and Johnston (2007) used a typical jelly mold, a hollow pineapple and a hollow teddy bear in their experiment, and arguably found that the more human the hollow object, the more affect the illusion had on participants.
Of course there are problems with this, as a pineapple is rarely described as more human than a traditional jelly mold. Also, this is not what Hill and Johnston originally set out to find. They were really just looking at the effect of this illusion on random hollow objects. There is also the problems of only a small amount of participants having been used.
All in all, though, it seems as if the more like a human face a concave object is, the more potent the illusion, and I am sure that this is because of an evolutionary adaptation to nonverbal communication.