Enter behaviorism. In an attempt to explain how the body and mind were related--the "mind-body problem"--psychologists like Watson and Skinner (those are the two big names; there were many others) came to the conclusion that if psychology was to be a science, then only observable, material things could ever be contained within its domain of investigation. Behaviorists stated that if you wanted to be a scientist rather than a philosopher, the mind was irrelevant; it was only the observable--the body, the behavior of the organism--that could ever be relevant either to research or to clinical psychology. Mental states had nothing to do with behavior; the only thing that mattered was the environment that the organism was exposed to. According to behaviorists, you don't really "feel hungry"; you are simply in an environment in which you have not eaten in a while, and this environment prompts you to seek out food. Your mental state is irrelevant.
Since then, psychology has grown a great deal. While we are still very much focused on what we can observe, especially when it comes to animal behavior, we have also learned that the mental state of the human being is very much relevant. Cognitive psychology came into the forefront, and we began to understand that how people learn and what information they draw on has a lot more to do with behavior than the immediate environment. Maybe it has something to do with the information revolution: We now know that data is just as real as material behavior; an idea can be as powerful as a physical object. And psychology has come to take this into account.
Well, most of psychology takes it into account. There's one area that's still stuck back in the era of behaviorism, and that's therapy for autistic children. In teaching a child by using ABA, only behavior is taken into account; the mental state is treated as irrelevant. Rather than examine how the child thinks, the therapist observes what the child does. Rather than attempt to give the child relevant information which he can use to change his behavior, the therapist attempts to create an environment which will trigger the child's behavior into taking a particular shape.
Maybe it comes from the fact that it's difficult to communicate with an autistic child, especially a non-verbal autistic child. People do often assume that if they can't see it, it's not there, so it's natural that they would assume that the child with whom they have difficulty communicating has nothing to communicate. This assumption pops up in a lot of the language people use when they describe autism: The child is an "empty shell"; he has been "stolen away" by autism; he is "not there"; he is "in his own world"--assumed to be an empty, desert-like world in which nothing happens.
I think it's about time autism psychology left the 1950s and took advantage of some of the things we've learned since then. The mental state of the autistic child is absolutely relevant. The way we perceive the world, the way we think, and the things we know are crucial to how we learn. Using only environment and stimulus-response to teach a child is to ignore the child's mind and the child's experience of the world. Rather, we should teach communication first of all, by interaction and by participation in the child's life, so that we can transfer that crucial information which will allow the child to connect his mind to ours.