I originally crossposted this to the Obed-Comp-L (Obedience Competition Mailing list) and Click-L (a list for those interested in Karen Pryor's point of view) in the summer of '95.
I have found that many people confuse negative reinforcement and punishment. These people include dog trainers, college students, and surprisingly, even many college professors and Introduction to Psychology textbooks. In fact, I recently read the book Don't Shoot the Dog! by Karen Pryor and was a bit dismayed to find this confusion there as well. Note that this is not an attempt to flame the book or criticize it unduly. I like the book and learned a lot from it. I recommend it highly to anyone interested in training animals (and sometimes even those interested in training their children). Nonetheless, I do believe that the book suffers from the problem of confusing negative reinforcement and punishment. As such it propagates this confusion because many dog trainers look to this book as a definitive source on the science of behavior.
I should note, however, that in the preface to the book Karen Pryor says (p.16), "Before going any further, I must apologize to any professional behaviorists who are annoyed at my cavalier uses of the vocabulary of reinforcement theory. . . . hair splitting definitions have proliferated. So I have sacrificed technical precision for a vocabulary I think people can understand." While I don't consider myself a professional behaviorist and I wouldn't say I was annoyed, I do believe that clearing up this confusion will lead to a more wise and productive use of the principles of operant conditioning.
Therefore, I'd like make an attempt at clearing up the confusion over negative reinforcement and punishment. Both are considered consequences of behavior and the notion of the consequences of behavior is central to operant conditioning. The basic paradigm of operant conditioning is that:
|R -----> S*|
|Which should be read as:|
|a Response (leads to) a Stimulus consequence|
For example, when the dog sits, (s)he gets a cookie. Note that there exists a contingency between the R and the S*. This means that whether the S* is administered depends upon the occurrence of the R. In other words, the dog will only get the cookie if, and when, it sits.
Keep in mind that things are a bit more complicated. An S* can also be contingent on No R. Thus, there may be a consequence of not responding. For example, if you don't smoke I'll give you a cookie. Things are also complicated by the fact that a Discriminative Stimulus (or S-D) usually signals that a particular contingency is in effect. For example, I tell the dog to "Sit". What this really tells the dog, is that if it puts it's butt on the floor right now, it will get a smile (and maybe even a cookie).
I believe the confusion over negative reinforcement and punishment is due to several factors. The first concerns the term "negative". It suggests something bad or unpleasant. From a scientific perspective, however, the terms positive and negative are used in a mathematical sense. Positive refers to addition and negative to subtraction. In the case of the consequences of behavior, positive refers to the adding or giving of something and negative refers to subtraction or taking something away. Let me elaborate on this reason for the confusion before going on to discuss others.
Basically, there are two kinds of events: pleasant and aversive. An example of a pleasant event is food and an example of an aversive event is electric shock. These events can be viewed as primary (related to a biological need) or secondary (previously paired with a primary). Furthermore, an event can be given (positive) or taken away (negative). Thus, there are four possible stimulus consequences of behavior or S*'s.
|What you do with the event|
|Type of event||Give it||Take it away|
|Pleasant||+ reinforcement||- punishment|
|Aversive||+ punishment||- reinforcement|
It is important to note that the goal of reinforcement is to increase the likelihood of the behavior in the future, while the goal of punishment is to decrease the likelihood of the behavior in the future. Note also that from this point of view, the terms 'negative' and 'aversive' have very different meanings.
In an effort to make this material as clear as possible, I would like to give a formal definition and an example for each of these four consequences.
Example: Dog sits (or does whatever else you might ask) and you give him/her food, play, and/or praise.
Example: Dog sits up on the long down in the AKC obedience exercise and the owner becomes verbally and/or physically firm.
Example: A technique for teaching the retrieve involves releasing an ear pinch or terminating a shock at the moment the dog clasps the dumbbell in its mouth.
Example: Puppie learn bite inhibition during play when we recoil and cease playing for a moment or two when the dog is too rough. Also, consider the earliest stages of teaching a dog to heel. We may give constant praise and encouragement when the desired behavior occurs but we will withhold the praise when the behavior observed does not meet our criteria of "good heeling".
I believe that a second reason for the confusion of negative reinforcement and punishment (which is very related to the first reason) is the use of a popular slogan by the behaviorists -- "Accentuate the Positive". In light of the information I have presented above, this slogan would be more accurate if it read -- "Accentuate the Use of Positive Reinforcement". The problem, however, is that this latter phrase is not as simple or as catchy as the former.
A final reason for the confusion of negative reinforcement and punishment is the fact that they are interrelated. In order to use negative reinforcement, one must typically administer the aversive stimulus in order to be able to terminate it. For example, let us again consider the retrieve. An ear pinch or shock is terminated at the moment the dog clasps the dumbbell in its mouth. In order to terminate the aversive stimulus, however, it had to have been initiated (i.e., administered) at some point. Thus, in order to use negative reinforcement, one must typically employ positive punishment as well.
Related articles include:
[Other articles by this author]