Since autistic behavior was first described in the 20th century, many theories have been put forward to try to explain why certain individuals seem to behave and react rather differently from the more typical population. Some of these ideas were popular for a while and then were rejected-- an example is the old belief in the "refrigerator mother" who was so cold and aloof that she severely disturbed her child's psychological development.
Over the last decade or two, an interesting theory has been put forward. This is the idea that autistic individuals for some reason do not develop the skills of "mindreading" that allow the rest of us to understand what others know, want, or believe. Most human beings are quite good at figuring these things out from other people's behavior, and begin to show their skills even when they are toddlers. Because we can "mindread" effectively, we can understand other people's communications even when they are not very clear, and we can figure out what we have to do in order to communicate our meaning to others--- or even to lie to or fool people successfully. If autistic individuals do not have "mindreading" abilities, that would explain why they have trouble with language and with other means of communication, as well as why they seem uninterested in social interactions.
A recent article in "Science" examined "mindreading" ability in individuals with Asperger's syndrome, an autism-related disorder characterized by poor social communication. The study reported in this article (Senju, A., Southgate, V., White, S., & Frith, U., "Mindblind eyes: An absence of spontaneous theory of mind in Asperger syndrome", Vol.325, pp. 883-885, 14 August 2009) looked at the ability of adults with Asperger's syndrome to do "theory of mind" tasks in which they have to figure out what another person may do.
The task done by the adults was one which can also be used with children, because all a person has to do is look in one place or another. On a little stage, a puppet appears and hides a ball in one of two boxes. An actor then stands behind the stage and puts a hand through one of two windows to reach for the hidden ball in a box. After the audience has gotten used to this, the puppet reappears when the actor isn't looking, and transfers the ball to the other box. Typically, either a child or an adult observer will look expectantly at the window near the box where the actor would think the ball is hidden, if he or she does not know the puppet has moved the ball. That behavior is interpreted as meaning that the observer has guessed what the actor knows and will do, just on the basis of whether the actor watched when the ball was moved. The correct guess shows that the typical observer has "theory of mind" and can figure out that mental states like knowledge, belief, or desire help determine what people will do--- and what they do helps us understand what they know or want.
When non-autistic young children watched the actor and the puppet, they looked in the direction of the box where they could assume the actor thought the ball was hidden. When adults with Asperger's syndrome watched the same episode, they did not do this, but instead looked toward the box that the puppet had moved the ball to, which would be unknown to the actor who was not looking. The adults did not seem to understand that the actor might not know something that the adults themselves had observed.
What about the idea that autistic persons, and those with similar disorders, avoid looking at people's eyes? Could it be that the Asperger's adults did not look at the actor's face because they do not like to look at eyes? The researchers were careful to exclude this factor by having the actor where a sun visor that made the eyes invisible to the audience. In addition, they examined how much time the Asperger's individuals looked at the actor's face, compared to how long a non-Asperger's control group looked, and there were no significant differences on this between the groups. The most likely explanation of this research seems to be that disorders like autism involve an inability to make good guesses about others' mental states.
Nevertheless, as these authors pointed out, adults with Asperger's syndrome can say in words where a person is likely to look for something, even though the Asperger's adults themselves do not look where they would expect the other person to search. It may be that human beings operate in (at least) two ways, one by using theory of mind and one by using verbal logic. The two methods may not be related. Happily, this would suggest that we may be able to develop teaching methods that will improve the social communications of individuals with poor skill at understanding other people.