Steve Nordby1 says that "Using behaviorism to control learning is like using an umbrella to control the weather."
There is a respectable tradition of criticizing the use of bribery in education. It can be ineffective and is often argued to do harm to the child's education. No technique, particularly not one which is ineffective or harmful, should be beyond criticism if this criticism can improve practice.
However, the criticism alluded to is is misdirected.
The picture of behaviorism it comes with is inaccurate at best, describing a single philosophical position including a ridiculous tabula rasa model of behavior and an insulting belief that people are just like dogs or rats or pigeons. This position ostensibly acts as an ideological justification for treatment of children (and everyone, in general) as passive people to be controlled like cattle (or molded like clay.) The coup de grâce is a declaration that reinforcement does not work and does harm to the child's education, and an implication that the sinister and misguided techniques are harmful to people in general.
This behaviorism is a straw man, but it is a ubiquitous straw man which is now coming to be defended by bright and otherwise well informed people who do not know any better.
The first important thing to realize about "Behaviorism" is that there are actually many different behaviorisms, with different content and vigorous disagreements. Speaking of 'behaviorism' generically, then, can be misleading. Let's be clear, then, that when talking about "Behaviorism" people are usually talking about John B. Watson's original behaviorism or B.F. Skinner's radical behaviorism (or, often,a noticeable mishmash of both behaviorisms with a few spectacular misconstruals - like the tabula rasa. See the last chapter of Skinner's About Behaviorism for his not unprecedented clarification that he unreservedly accepts genetic and innate contributions to behavior.)
Another pernicious feature of this picture, bearing directly on the application of behaviorism to education, is that it treats operant and classical conditioning as essentially equivalent, and both as equally implying that children are stupid or machine-like. Some of the confusion, of course, is that many behaviorists have taken up the project of explaining intelligent behavior 'algorithmically,' in terms which could be implemented by a biological organism or computer. Of course, this does not imply that the organism is stupid - only that it is not itself made out of smart stuff.
This issue aside, there are clear and relevant differences between operant conditioning and classical conditioning which are often deliberately glossed over by debaters eager to show that behaviorism is wrong and demeaning. Even the application of the same words, like 'conditioning' and 'reinforcement,' is misleading, because the Skinnerian versions are so much different from any of the prior formulations of classical conditioning or instrumental learning. Here are some of the principal differences.
- In operant conditioning, the probability or rate of response changes; in classical conditioning the strength of response (e.g., the strength of a salivation response as measured in drops of saliva) changes.
- In operant conditioning, functionally and not physically defined classes of stimuli (e.g., 'A' 'a') do not elicit, but come to increase the probability of, functionally and not physically defined classes of responses (e.g., saying 'That's an A' or writing 'It's a' etc.) In classical conditioning, a simple physical stimulus which formerly did not result in a response (say, salivation) comes to do so after repeatedly preceding a similar stimulus which did result in a response.
- In classical conditioning, it does not matter whether the response acts on the environment, and it is usually not intended to (salivation is just preparation for eating.) Operant conditioning operates on the environment to produce an effect which then 'strengthens' the response (increases its future probability, given the right situation).
- In operant conditioning animals try all different things themselves and eventually work out what works best; in other words, animals generate new behavior by taking clues from the environment. In classical conditioning there is no new response, only a new occasion for the old response.
With regard to the ineffectiveness of reinforcement, there is an important and often missed clarification which hinges on the definition of reinforcement. A reinforcer, in simple terms, is an event which causes the reproduction of a behavior by occurring after that behavior. This being the case, a bribe which does not affect or which discourages paying attention, then it is by definition not reinforcing paying attention. Similarly, a promise of a reward or even a known reinforcer is not a reinforcer, at least for the reason that it precedes the response. (If it serves as a signal to the child that reinforcement is available, it is a discriminative stimulus.)
Undoubtedly better techniques will be found in education, but not by confusing distinct positions, rejecting techniques because of their associations with certain philosophies, and not by making excuses for any technique - whether gold stars or treating the student as a passive receptacle for verbal lectures. It is in the common interest of educators regardless of their approach to develop a better understanding of motivation - whether described in operant language as what is reinforcing and what the side-effects of different reinforcers are, or described as discovering how to facilitate intrinsic motivation (or use extrinsic motivation effectively.)
1. Steve Nordby is no one well known, if you were wondering yourself. His quotes page is here: http://members.aol.com/svennord/ed/quotes.htm.