How Not to Talk About Reinforcers

In behavior analysis, a reinforcer is defined functionally, in terms of its effect on behavior. Thus, something (or someone or some activity) is said to be a reinforcer if it develops or maintains some response (actually, a cluster or class of responses, called an operant, which all have in common that they are maintained by the reinforcer). A reinforcer also may be demonstrated by showing that its removal decreases the previously maintained behavior.   

The alternative to a functional definition of a reinforcer is a structural one. This sometimes also is called a formal definition.  By a structural definition, some things (or people or activities) are classified by some arbitrary criterion as being reinforcers, such as the fact that the object is a preferred food item, toy, or activity. In these cases, reinforcers are defined in ways other than their behavioral effects, before they are ever tested in a given context to see whether or not they actually change behavior as noted above. 

An historical example of a structural definition of a reinforcer is when it was defined as something that reduces a physiological or psychological need. Here, there is no behavioral criterion or criterion in terms of the effect of the reinforcer on behavior. The reinforcer is defined independently of behavior in terms of other properties - need reduction. 

All of this is building to something that really bugs me about some discussions of reinforcers. In talking about it, I confess at the outset that I am not without sin in this regard, but like all sinners I am trying to be better. It is common among us behavior analysts to speak of “effective reinforcers.” For example, one often hears someone say something like “to shape behavior we begin by finding an effective reinforcer” or “an effective reinforcer is needed to control the response” or just “such and such was an effective reinforcer.”  

Using this expression implies that some reinforcers are effective, but others are not. Such use means that the user subscribes to a structural definition of a reinforcer, something that is anathema to behavior analysts because we believe that reinforcers are, so to speak, made and not born.  Reinforcers are classified as such by their functional effects on behavior, not on the basis of some phantasmagorical (I learned that word from my very capable former graduate student and now colleague from Brazil, Raquel Aló) a priori (before the behavioral test – I learned that word from a high school Latin book) category that allows us to know in some magical, telegraphic way what will and will not change behavior before we have ever made the test.

If you accept the idea that reinforcers should be defined functionally, then a reinforcer is a reinforcer only if it is effective - that is has the requisite behavioral effect - and the expression “effective reinforcer” is redundant. So, if you subscribe to a behavior-analytic world view, you will avoid this term like the plague when speaking of single reinforcers.

There is, of course, a qualification.  “Effective” is a perfectly good word when the discussion is about two or more reinforcers. In this case we can without problem compare the two by observing how each develops or maintains behavior.  We thus can on this basis say that Reinforcer A is, in whatever behavioral sense you are comparing the two, more effective than Reinforcer B. Thus, if Reinforcer A maintains higher response rates than Reinforcer B on the same schedule, or if responding is more resistant to extinction following maintenance with Reinforcer A than with Reinforcer B, or if a child chooses Reinforcer A over Reinforcer B, you can speak of Reinforcer A as being more effective (in that context) than Reinforcer B. But, in an absolute sense Reinforcer B is effective just because it maintains some behavior.

So, all you behavior analysts out there, if you want to be an “effective” behavior analyst, don’t fall into the structuralist trap and talk (dare I say it?) mindlessly about “effective reinforcers” unless you are talking about reinforcers relative to one another, and not in an absolute way. 

Posted by Andy Lattal, Ph.D.

Dr. Andy Lattal is the Centennial Professor of Psychology at West Virginia University (WVU). Lattal has authored over 150 research articles and chapters on conceptual, experimental, and applied topics in behavior analysis and edited seven books and journal special issues, including APA’s memorial tribute to B. F. Skinner.