In 1913, John Watson identified observable behavior as the proper subject matter for psychology and stated that all behavior is controlled by environmental events. This was where Watson laid out the stimulus-response psychology that started the movement called behaviorism. B.F. Skinner later clarified the distinction between Ivan Pavlov's respondent conditioning (conditioned reflexes) and operant conditioning, in which the consequence of behavior controls the future occurrence of the behavior.
Skinner and others outlined basic principles of behavior, which include reinforcement, prompting, fading, shaping, schedules of reinforcement, etc., etc., etc. These principles comprise the pure science of behavior analysis. It's important to understand that this is the pure science, NOT the applied science. This distinction is made when any science is studied in both pure and applied ways (i.e., physics). When the principles of the pure science of behavior analysis are used to teach (or when used in any applied setting), this practice is called "Applied Behavior Analysis" (or, earlier, behavior modification). This means that the principles used to describe how behavior is lawful, observable and measurable, and has an impact on the environment have been adapted into teaching methods based on those principles.
The reason that people who don't know this get confused about what ABA is stems from the difference between the pure and applied science. The science is one thing, made up of these principles of behavior. The applied science is another, made up of strategies based on those principles. The reason that people say there are different "kinds" of ABA -- which is an incorrect statement and I'll explain why -- is because people hear about "Lovaas," "DTT" "NET," "AVB," etc. What those distinctions actually describe are different APPLICATIONS of the SAME science. Here's how that came to pass.
In the 60s, Ivar Lovaas began working on what would later be described in The ME Book, which laid out his teaching application of the science. He laid out a curriculum of programs, as well as a teaching sequence for them, and discussed how to teach them. The means by which these skills were taught was applied behavior analysis (the application of the science of behavior); the way in which the science was applied was Lovaas's. This is what most people discussing ABA mean: They mean that they teach using Lovaas's application of the science of behavior analysis.
Now, many know that Lovaas unfortunately originally included the use of aversives in his application, which most people rightfully questioned. He later reconstructed things to exclude aversives and at this point any application of behavior analysis that still includes the use of aversives is not only illegal and immoral, but archaic, too. Lovaas' long-in-coming follow-up to The ME Book, Teaching Individuals with Developmental Delays: Basic Intervention Techniques, has just recently become available, which describes how the basic programming has changed over the past few decades. But his original application described a way of teaching a target first in isolation, then with one and two distracters, then teaching another the same way, then putting the two into random rotation. He also put a lot of emphasis on intensity, table-time, eye contact, sitting still, etc. Almost everyone not directly trained by Lovaas or at UCLA who implements the Lovaas application of behavior analysis has changed it in some way, whether wittingly or not, so it's rare to find a true "Lovaas" program these days. Those changes, by the way, can be for better or worse. But Lovaas's program was the first application of the science of behavior analysis widely used to teach individuals with autism, and the importance of that contribution can and should never be forgotten.
In addition to fleshing out operant conditioning, Skinner also analyzed the functions of language and presented his analysis in the 1957 book Verbal Behavior. This described language in terms of its functions (see the section on What is AVB?), mainly mands, echoics, tacts, and intraverbals. People like Jack Michael, Vince Carbone, Mark Sundberg, James Partington, and others, all applied behavior analysts, studied Skinner's pure scientific analysis (Skinner's book does not outline a method for applying verbal behavior; he only describes it). They took the pure science and applied it, which describes what we refer to as applied verbal behavior. Now, this is STILL applied behavior analysis, still an application of the science of behavior, but the focus is on the student's verbal behavior.
There are other differences in applications, such as the type of prompting that's done (no-no-prompts versus errorless learning), the ratio of table-time to teaching done in the natural environment (NET), and the presentation of skills (mass trialing of one target versus several targets within one program at a time versus mixing a number of different skills all at once). There are probably as many different applications of the science of behavior analysis as there are home and school ABA programs. So when someone asks if you are doing "ABA," you can say yes whether you're doing a Lovaas program or an AVB program. You can say this as long as you're teaching based on the principles of behaviorism. The more relevant question is what application of the science of behavior you're using. That will tell much more about the way in which you're teaching.
There are many who say that they "don't like ABA." I think what those people mean is that they actually dislike Lovaas's application (or whichever they've seen) of the science of behavior, not the science of behavior itself. There's very little to argue with the ideas that we do things more often when they have positive consequences (reinforcement), that when we don't know how to do something we need help to learn how (prompting), that such help needs to be eventually removed if we are to do the skill on our own (fading), and that we get better and better at things as we practice them more (shaping). Those, in a nutshell, are the most common principles of behavior. There's more to argue with putting a young child in a chair for 40 hours a week, requiring eye contact and sitting still. That, in a nutshell, is one overgeneralized (mis)interpretation of a strict Lovaas application of the science of behavior.
What I do, along with many others, and have always done is to first form a friendship with the child and to pair myself with reinforcement (pairing is the applied strategy based on reinforcement, the behavioral principle). This is not only the best human approach, but also a solid behavioral technique. I make few, if any demands, keeping the level of reinforcement for learning high and the value of escape (manifested in non-compliant, aggressive, or stimming behavior) low. My first "demand" is to teach the child to mand for (request) what s/he wants. This is completely functional, since it gives the kids an appropriate and effective way to communicate needs and wants. Mands for everything that the child might want are taught, including but not limited to preferred items, help, play, information, negative reinforcement (removing something the kid doesn't want), throughout the course of teaching. Gradually, I increase the "demands," though rarely more than 20% difficult tasks to 80% easy tasks. Those "demands" are to teach the child to label everything in his/her environment, to converse, to read, write, spell, do math, and any of a thousand other functional educational things.
In addition to keeping the work roughly 80% easy, I also make the "difficult" tasks easier by using errorless learning, in which I prompt the child if necessary in order for them to be correct and get reinforced, then immediately transfer the task to an independent level so that s/he learns to do it independently, for which s/he is differentially reinforced (more reinforcement of a better quality for better or independent responses). I also spend usually no more than 30% of our time together teaching at the table; the rest is done everywhere else -- inside, outside, at play, with the TV, etc. This results in several things: 1) The child learns as much as, if not more than, s/he would just at the table and 2) the child instantly learns to use everything I teach everywhere because not only is it all functional, but it's already been learned and used in every environment s/he encounters. I can't send you a description of what exactly I do every day because it changes every day, every minute, with every student; all that's constant is the behavioral techniques I employ to teach and the sheer joy I try to bring (and always experience myself) to every interaction my students and I have.
Is this what folks traditionally mean by "ABA"? Not necessarily, since (as I mentioned earlier) most are referring to Lovaas's application of the science of behavior. Is this "ABA"? Yes, since it is all based in the principles of the science of behavior; it's simply a different application.
The science is solid; it's certain applications of it that need tweaking. The tweaking should be done based on the students and no one else, and THAT, believe it or not, is what the data are for. They, along with our personal, experiential knowledge of the student, tell us what's effective and what's not. Applied behavior analysis is not evil; it's not anti-person, nor is it cold and heartless. Does ABA intend to change behavior? Yes, that's the whole point. However, the behavior that's targeted for change is always chosen such that the life of the student will be improved. Is that a value judgment by teachers? Yes, but no more than those made by anyone who ever attempts to teach anyone anything (as all of us who write to various lists, write books, have conversations, and interact with others in any way do). I think that the important piece is that the teaching is respectful to the student, that the application of the science of behavior that's implemented makes sense and is effective and positive for the student. I think any of us who can say that we believe ourselves to be doing that at least agree on one thing.
A Comparison of Approaches