Teaching Children with Autism, TAGteach, and Shaping
Based on the little I know, I have been trying to think my way through Shaping. If I understand correctly, a single occurrence of a consequence is enough for a learner to form an association between that consequence and the behaviour that it (immediately) follows, but not necessarily (or even usually) a strong enough association that (expectation of receiving) that consequennce becomes a reinforcer for that behaviour. For that consequence to become reinforcing, it usually needs to be paired with the behaviour on several occasions, and teaching/training is usually most effective if that reinforcement is provided on a continuous (reinforcement) schedule until the behaviour is learned, i.e., emitted with some regularity. Once a behaviour is learned, however, the teacher should “thin the reinforcement schedule,” moving from a continuous reinforcement schedule to one that is both intermittent and variable. The answer to the question of just how thin that reinforcement schedule should ultimately be is “as thin as you can get away with and still be able to rely on the behaviour being emitted.”
Continuing with my speculation about how shaping works: It seems to me that thinning the reinforcement schedule is aversive (from the perspective of the learner) but necessary (from the perspective of the teacher) and, if it doesn’t occur too quickly/precipitously, likely to be tolerated by the learner and a “good thing” for all concerned in the long run – in terms of conservation of reinforcers, retention of the learning, etc., etc. However, during shaping, the reinforcement schedule should not be as thin as you can get away with, but should be maintained at a more-or-less constant “not too thin” level – a judgement call on the part of the teacher – until the behavior being shaped reaches it end goal.
In this scenario which I am imagining, the aversiveness of thinning the reinforcement schedule results in an immediate increase in variability of responding, and some of those variations are going to be closer to your end gaol than others; and the teacher will immediately start to continually reinforce those variations while continuing to reinforce enough of the less-desirable responses to maintain the chosen “not too thin” level of reinforcement. How often the less-desirable responses get reinforced will depend, therefore, not only on your chosen level of reinforcement, but also on how often the closer-to-the-end-goal response is being emitted, since those “better” responses have to be occurring frequently enough that you can reinforce the less-desirable responses less frequently and still maintain the level of reinforcement that you have chosen.
This scenario will be repeated over and over as you move closer and closer to your end-goal behavior; and how quickly you will achieve that end-goal behavior will depend on several variables, including what the teacher chooses to reinforce, the strength of the reinforcement, the learner’s innate talent or learning ability, and the difficulty of the behavior to be learned.