Black and White Thinking & Lack of Creativity
In ABA, an answer is either correct or incorrect. If it's correct, it will be rewarded. If not, it won't be rewarded - it may even be punished. ABA has no procedures for if the therapist doesn't know the answer, or for correct answers the therapist didn't think of, or answers that could be argued to be both correct and incorrect. If they're asking the question in a discrete trial, they have a correct answer in mind, and any other answer is wrong.
Creativity requires a more open-ended perspective, where an infinite number of potential answers are equally acceptable. If drawing flowers means drawing the particular flowers your teacher has in mind, then how can you learn to paint flowers with all the colors of the rainbow?
Difficulty Generalizing
Overall, a discrete trial tends to be extremely precise, with each presentation involving, as much as possible, exactly the same command in the same wording with the same intonation. When your experiences of learning a skill are so uniform, how are you to know which cues are or aren't important to determine when you should make use of the skill?
In tests of effectiveness of ABA, it's common to do reversal tests, confusingly also known as ABA or ABAB. The point is that when you remove the cue designed to elicit the behaviour change, behaviour reverses to baseline.
Ironically, this exact same response that is often taken as proof of effectiveness of ABA procedures is also frequently framed as a problem - failure to generalize. Even with the exact same contingency, reversal of behaviour change in response to minor environmental changes can be seen as a good thing at one point and a bad thing at another point. It makes little sense to me, reading papers written by and for ABA practitioners. How much less sense must it make to a child being taught by these methods?
Reversal tests not only prove that generalization has failed, they also promote more failure to generalize, by training the child with experiences that prove the contingency they’re learning is situation-specific.
When ABA therapists want a child to generalize a behavior, their usual strategy is to recruit more people in the child’s life to reinforce the desired behavior. This can be a difficult prospect, since even one person who doesn’t consistently enforce the desired contingency will teach the child that there are situations this behavior doesn’t get that reward in. But even if it works for everyone, it simply kicks the problem further down. If the person ever graduates from the ABA program, if they ever achieve independence, they’ll be in a situation where they’re no longer expecting rewards for the trained behavior.
True generalization tends to come from understanding the reasons why a behavior is useful and learning to independently identify when the behavior should be applied. This is not taught by ABA. A child learns to say “blue” not by learning that saying “blue” directs their communication partner’s attention to blue objects and therefore makes it easier for them to understand that a communicated message pertains to a blue object as opposed to the red one beside it - no, they learn that if the therapist shows them a blue object and says “what color?” and they say “blue”, they get an M&M (of any color!). If the person doesn’t ask them what color something is, uses an actual sentence (“what color would you say this ball is?”) rather than a sentence fragment, or isn’t holding a pack of M&Ms, will saying blue be useful? They don’t know.
In fact, verbal behavior theory fails to account for the main goal of communication. If you say “milk please” and actually use that as a communicative phrase, even though you might like to get an M&M, getting an M&M isn’t the goal - getting milk is. Verbal behavior theory treats all rewards for verbal utterances as equal. However, only responses that are actually connected to the message encourage the child to make the connection that the word “milk” and the word “please” actually have meanings and can be used to communicate their thoughts to others.
Sometimes, children receiving ABA will incidentally learn that the skills they’re being taught in ABA actually have a use. A child will sit on a potty and pee to get an M&M, only to incidentally learn that this behavior also keeps them from having to put up with the discomfort of a wet diaper. A child will happen to echo a trained phrase in an untrained situation and get a response that actually reflects the meaning of the phrase. But these things happen by accident, not intention, in ABA. And they don’t always happen for every skill that’s taught in ABA.
Aggression & Self-Harm
Martin Luther famously characterised violent protest as “the language of the unheard”. In children, aggressive behavior is usually attributable to emotional distress, instinctive threat responses, and the communication function of refusal. Self-harm, meanwhile, tends to be more specifically linked to distress.
ABA programs focused on developing communicative behavior tend to overlook refusal. Picture Exchange Communication System, an AAC intervention pioneered by ABA specialists, teaches requesting first, followed by commenting, and never even mentions refusal, not even in the “next steps” once the child has completed training.
This is not an accidental oversight. Many children don’t enjoy prolonged ABA instruction, and if adults were to honor their refusal communications, they would receive far fewer discrete trials. In addition, many goals of ABA involve teaching the child to accept stimuli they’d normally refuse, such as eye contact and physical touch, and that requires deliberately and consistently ignoring refusal communications.
Emotional distress is unacknowledged in most ABA programs. ABA comes from behaviorist theory, a theoretical approach that attempts to explain behavior solely through external factors. Emotional distress is an internal experience, outside the scope of behaviorist theory. The closest that behaviorist theory comes to acknowledging emotional distress is in acknowledging that organisms enjoy certain things more than they enjoy other things - however, this completely ignores the way that overwhelming emotional distress can result in irrational and counterproductive behavior that is beyond the individual’s conscious control. It’s quite common for individuals to engage in behavior that is consistently punished when they’re experiencing emotional distress, because their distress demands action and they don’t know any way to relieve or tolerate distress that won’t get them punished.
Autistic individuals can experience emotional distress for the same reasons as anyone else, but they can also experience distress for autism-specific reasons. For example, the majority of autistic individuals have some degree of sensory hypersensitivity, a neurological condition that causes distress at sensations most people aren’t bothered by. These can include sensations associated with the intended goals of ABA treatment, such as seeing people’s eyes or being touched, as well as incidental sensations such as background noises, fluorescent lighting, etc.
Behaviors that lessen emotional distress are internally rewarding, even if they aren’t externally rewarded. If emotional distress outstrips an individual’s executive functioning, they will compulsively engage in the most powerful emotion distress-reducing behaviors they can think of, regardless of any external consequences. At this point, punishment tends to merely increase the person’s distress, resulting in an increase rather than a decrease in the maladaptive behavior.
Punishment is ineffective at interrupting emotional distress behaviors not only because it increases distress, but also because the overall experience of being in distress is so innately punishing that a single act of external punishment tends not to stand out noticeably in comparison. What does stand out is anything that reduces distress, even momentarily.
Self-harm can lessen emotional distress by causing the release of natural endorphins. Endorphins are the neurotransmitters that opiate drugs mimic, and high levels of endorphins consequently result in similar sensations to taking a low dose of an opiate drug. In addition, physical pain attracts attention, reducing the person’s ability to focus on any other thoughts or feelings, including thoughts and feelings that are contributing to emotional distress.
Many parents of toddlers have noted that a momentary distraction can sometimes shake a child out of a tantrum, and pain is extremely distracting. Pain that is from an external source is both distracting and distressing, and therefore likely to only temporarily disrupt a distress response before amplifying it. However, pain that is under the control of the individual experiencing it is far less distressing, both because it’s predictable and because it can be tailored to be just the right level of pain rather than too much pain. (Although if a self-harming individual accidentally causes themselves more pain than intended, this tends not to function as desired.)
In addition, behaviorists have frequently struggled to predict or control instinctive behaviors - for example, raccoons trained by behaviorist methods to exchange tokens for food rewards eventually begin compulsively washing the tokens, impeding their ability to actually exchange the tokens for the food. Instincts tend to be more strongly controlled by classical conditioning than by operant conditioning, so stimuli associated with a situation that triggers instinctive behavior will trigger the instinct regardless of whether it’s a functional or even counterproductive response.
Humans, too, are animals, and we have instincts as well. One of our most powerful instinctive responses occurs in situations where we perceive ourselves (accurately or not) to be in imminent risk of serious or life-threatening harm - including many situations associated with extreme distress.
In these situations, we have three competing instinctive tendencies - fight, flight or freeze. If the fight instinct is triggered, the result is aggression. This is especially likely in situations where other tendencies are inhibited, especially flight - frequently, an attempt to escape a situation will turn into aggression targeted at whoever or whatever is impeding the individual’s escape. For example, a child in an instinctive threat response might first try to elope, and if restrained, respond with aggression towards whoever is restraining them. If this aggression fails, they may eventually become limp, and it may be difficult to elicit any response from them whatsoever. As long as the instinctive threat response state continues, if they do respond, it’ll be switching back into fight or flight mode. This is not a response subject to operant conditioning.
When ABA treatment elicits emotional distress and/or instinctive threat responses, the therapists are trained to respond in counterproductive ways, since more productive responses - ie doing things that reduce distress and communicate that the situation isn’t a threat - tends to be seen as rewarding the unpleasant behavior the individual is engaging in. For example, offering a comfort toy to a crying child is often an effective way to reduce distress, but it’s also giving the child something they find rewarding in response to a behavior you want to reduce.
People who acknowledge internal experiences must learn to distinguish the difference between crying that is an uncontrollable response to distress and crying that represents an operantly-conditioned voluntary behavior, and provide responses that reduce distress but don’t reinforce operant conditioning responses. However, this balancing act only exists if you acknowledge that behaviorist theory doesn’t address the full complexity of human (or animal) behavior.
Furthermore, since most autistic people have atypical sensorimotor systems, they often have atypical self-soothing behaviors, such as rocking, monotone humming and hand-flapping. Unlike aggression and self-harm, these behaviors are essentially harmless, and therefore among the more adaptive means that many autistic people have to reduce their distress. However, since ABA for autism frequently has a goal of making the child look more superficially normal (often described as “social acceptability”), these atypical motor behaviors are often targeted for reduction regardless of any adaptive function they may serve.
While distress and instinctive threat responses don’t tend to respond well to operant conditioning, they tend to respond very strongly to classical conditioning. Classical conditioning is the kind of conditioning described by Ivan Pavlov, where an instinctive salivary response to food began to be elicited by a stimulus paired with food.
If you consistently respond to distress behaviors and instinctive threat responses in ways that increase the distress the person is experiencing, over time, classical conditioning will result in the internal experiences of distress and/or instinctive threat becoming more likely in situations that contain stimuli shared by the situation that has previously elicited distress and/or instinctive threat responses.
For example, a child who frequently experiences distress at school may come to find the sensation of wearing a backpack or the sight of a school bus also elicit distress. A child who enters an instinctive threat response state in the context of toilet training may eventually enter that state when they simply see a toilet. A child who, during an instinctive threat response state, was physically restrained to stop them eloping, fought back unsuccessfully and then fell into a freeze response may later exhibit any of those three responses to a momentary touch from the same person who restrained them before.
In severe cases, classical conditioning of instinctive threat responses can result in an individual developing mental health conditions that are characterised by overactive stress responses, such as specific phobias or even post-traumatic stress disorder. In children, instinctive threat responses and distress are also subject to neuroplasticity, with a host of permanent biochemical and behavioral consequences. For example, cortisol production appears to be elevated in many traumatised individuals, and yet permanently suppressed in the most severely traumatised children. Both elevated and suppressed cortisol likely have major implications for both behavior and physical health over the long-term.
The duration of ABA training and the frequently sedentary nature of discrete trial training are also relevant. Physical exercise reduces susceptibility to instinctive threat responses and is likely to release endorphins, just like self-harm does. Children generally are more inclined to be physically active than adults, a pattern seen across many vertebrate species as soon as they become capable of independent locomotion. This means not only that they are more likely to spontaneously engage in physical exercise, but also that they need more physical exercise to get the emotional benefits of exercising, and that they’re more susceptible to boredom and restlessness in the absence of physical activity.
ABA practice has, over time, placed less emphasis on doing discrete trial training with a child seated at a table, instead moving towards doing discrete trials on the floor surrounded by toys or in other contexts that feel more natural for a small child, the constraints of discrete trial training still generally require inactivity. If a child is running circles around the room, for example, they’re not looking at the block you’re touching when you say “what color?”
This enforced inactivity isn’t a huge concern if it occurs for brief periods, increasing gradually in duration as a child’s developmental need for physical activity decreases. However, ABA practice generally encourages intensive, early intervention, resulting in children as young as 18 months old being subjected to discrete trials for 30 to 40 hours a week. This is extremely developmentally inappropriate, since even 6 and 7 year olds are likely to struggle with being inactive for that much time during a week.
One way that children can attend to stationary activities and still meet some of their need for physical activity is by engaging in irrelevant small-motor movements - fidgeting. Since autistic children’s sensorimotor systems function differently from most children, these fidgeting movements, like the self-soothing movements described above, often look atypical in autistic children.
Unfortunately, despite fidgeting being essentially harmless and frequently enhancing the capacity to pay attention, it is sometimes distracting - to other learners, if present, and especially to the therapist or instructor who is actively monitoring the child’s responses. Furthermore, if fidgeting looks atypical, “social acceptability” goals can encourage a therapist to attempt to suppress this behavior even more so than normal fidgeting. Regardless of whether it looks typical or atypical, suppressing fidgeting increases the negative psychological consequences of prolonged, enforced inactivity.
In addition, not moving when you want to move is a form of behavioral inhibition, as is paying attention only to what someone else wants you to attend to. Inhibition is an effortful usage of executive control, and executive control is subject to fatigue. Executive control is also used to suppress maladaptive instinctive and/or emotional responses, and therefore, executive fatigue increases the risk that an individual will behave compulsively in response to their emotions and/or instincts.
Children not only need more executive control to inhibit movement since they have a greater drive to move - they also have less executive control overall, since the prefrontal, the brain region primarily responsible for executive control, is continuing to develop up until the mid-twenties. In addition, autism is associated with a generalized impairment in executive functioning, including impaired executive control - and if the individual has ADHD as well, a frequent comorbid condition, this is even more true.
A child with impaired executive control being required to inhibit activity during developmentally inappropriate prolonged sedentary activities requiring intense mental focus is extremely prone to executive fatigue. This is probably one of the big reasons that, despite the common recommendation that autistic children receive 30 to 40 hours a week of ABA treatment, better outcomes have actually been reported for children who receive less-frequent ABA treatment. Lower intensity treatment presents the child with learning opportunities with less risk of triggering executive fatigue.
This is the same reason why 4 and 5 year old typically developing children, if they are in formal preschool or kindergarten educational settings, frequently attend for half-days or every second weekday, resulting in much less time per week in formal educational settings than older children. Unfortunately, many people who recognize the increased need for downtime and physical activity in young typically developing children are nonetheless willing to require many hours of effortful control in same-age or even younger developmentally disabled children, despite the needs for downtime and activity being the same or greater in many of these children. Few people who support early intensive ABA for autistic toddlers would be comfortable with a daycare program that expects 18 month old typically developing children to sit still and focus on academic tasks for 30-40 hours a week. Is it surprising that a child subjected to such developmentally inappropriate instruction might experience escalating distress responses?