This hypothesis had already been made by a japanese team who realised an extensive survey of primate eye morphology, and concluded that human eyes were disproportionately visible, as though they advertised for their owner's mental activity. Besides, we know that human infants are experts at using gaze-direction cues from their conspecifics as early as their first year of life. Tomasello and his team put the "cooperative eyes" theory to the test, with the following experimental design:

A human experimenter "looked" to the ceiling either with his eyes only, both head and eyes, or neither. Great apes followed gaze to the ceiling based mainly on the human's head direction (although eye direction played some role as well). In contrast, human infants relied almost exclusively on eye direction in the same situation.

Two additional control conditions involved the experimenter turning its back on the subject and gazing ahead, or looking up (this was meant to differentiate the effect of "head" from the effect of "face".

Even though technical difficulties did not make a direct comparison of chimps and infants' gazing patterns possible, the data seem to support Tomasello's idea, as well as a related prediction: the converse hypothesis to Tomasello's would be that non-human apes' sclera (the sclera is the fibrous covering around the pupil, which is white in humans), evolved to be black because it prevented the chimp's conspecifics to decipher her intentions. It seems indeed that chimps know how to use gaze directions in order to gain information (Tomasello does an extensive review of the mixed evidence for gaze-following in chimpanzees), but rarely do so, presumably because it's unpractical.

Tomasello acknowledges that one could imagine other functions to the human eye beside vision: for example, Brian Hare recently showed that the sight of a pair of eyes primes higher generosity in the ultimatum game.

I'm enthusiastic about the study, yet I would like to offer an alternative interpretation: chimps, much more than humans, are a tree-dwelling species with an evolutionary history that mainly took place in the forest. In such habitats, looking up over one's head is a frequent and useful habit; humans on the other hand evolved in a vaster and flatter environment where looking ahead and around was, I surmise, the overwhelming tendency (not to mention the flatness of middle-sized german towns where Tomasello's subjects originated - presumably, chimps at the Max Planck institute are raised in cages with places to dwell and play dispatched above the floor, as you can see in zoos and several research institutes). Thus, it might be that observing your conspecifics' vertical head motions made less and less psychological sense in human evolutionary history: one differences between chimps and us is that they have much more interesting things above their heads.

In support of this view is the fact that attention to head motion (which was paramount in chimps and nearly absent in humans), much more than attention to eye movements, provided the crucial difference between humans and chimps in Tomasello's experiment. I wonder what the results would be if the experimenter had gazed aside instead of looking up?

Note also that having white sclera and following our conspecifics' gaze is by no means a human specificity: goats and ravens (as I learned from Tomasello) have also been shown to follow gaze, and (as I learned from Google Images), they have white sclera (see the photographs). As far as I remember the litterature, this also applies to dogs. Is it possible to generalize? Are white sclera and gaze-following correlated? And what would the evolutionary rationale behind that tendency be?

Anyway, since all these species possess white sclera and gaze following without always being hugely gifted at social cognition and cooperation (or maybe it's just my being bitter about goats, since what relations I had with their folk when I was a child was anything but friendly), I find it difficult to support the cooperative eyes hypothesis.

One way to test my hypothesis (that Tomasello's result are due to differences in human and chimp's visual environement, actual or past) would be to submit a group of ravens (whose visual universe is mainly vertical) to goats (whose visual universe is flatter).

The same issue of the Journal of Human Evolution has a paper by Carel van Schaik, illustrating a model of the evolution of Orang Utan's brain size, that might account for some puzzling features of Homo Floresiensis. Van Schaik correlates big brains with a steady and reliably rich diet, without which evolution cuts costs by reducing brain size. Nothing very original, but the data are fine and the application to Floresiensis will delight fans of the Hobbit.