So here we are, looking at changing the definition of autism--merging the spectrum, and using severity levels to define specific cases. It makes sense to me, to merge the spectrum; but we need to be extra careful not to fall into the trap of thinking that "severe" autism is a different disorder from "mild" autism.
The traditional concept of what a diagnosis is tends to create the idea that everyone with a specific label must have the same traits, or have traits within a very narrow range. If a doctor diagnoses a cold, we expect a few days of discomfort and are justifiably surprised if we get pneumonia instead. When you look at autism and try to apply that definition of diagnosis to it, things don't add up--Why is this one person over here, who needs only a little help, in the same category as this other person, who needs 24-hour care? Doesn't that mean that those two people have two different things?
Well, no, it doesn't. When you say that LFA/HFA are different, you're not looking at the whole spectrum. Sure, if you picked out someone who's profoundly autistic, and then compared them to a local Asperger's college student, you'd find that they were very different. But if you took everyone, from all over the spectrum, and looked at that, you wouldn't find two distinct groups at all.
Autism isn't two different groups. The transition from LFA to HFA is gradual, and there are so many people in between that it doesn't make sense to think of autism as distinct categories. There are many, many people that are impossible to fit into one category or another--nonverbal people with genius IQs; developmentally delayed people with only mild autistic traits; people who ace the neurodevelopmental tests in one area and get soundly beaten by three-year-olds in another area. GAFs range from the single digits to the high nineties.
The fact is, autism is diverse--very, very diverse. Even knowing what traits a child has doesn't permit you to predict his adult functioning level. Children grow and learn; autistic children are no different. I've talked to parents of autistic children, and to autistic adults, who describe childhood traits that are practically stereotypes of profound autism--nonverbal, very few adaptive skills, unaware of the existence of other human beings. Most of these individuals are now verbal; many are living independently; some are married, employed, or both. They've seemingly moved from one category to another. If they go through an extremely stressful time, they may move right back. For that matter, I've moved into the "low-functioning" category myself, under enough stress, and I'm supposed to be "high-functioning" by any definition. How can you say that their autism now is a totally different thing from their autism as children; and would become a different thing yet again if coping skills were to be lost? You can't, really.
A diagnosis is only as good as what it lets a psychologist do; and what a developmental disorder diagnosis is meant to do is to give a psychologist or other professional a shorthand of what level of support someone is likely to need. For that purpose, dividing autism into profound/severe/moderate/mild makes sense. But only if you are very careful to remember that a diagnosis is a tool, a statement of how things are now; it does not predict what a person will learn, nor does it even begin to fully describe the unique nature of each case of autism.