The vocabulary of behavior analysis is sometimes incorrectly used. Some of the
typical faux pas are listed below. For an excellent and more complete account, see Jack
Michael’s Concepts and Principles of Behavior Analysis (ISBN #0-935868-51-8) available
through the Society for the Advancement of Behavior Analysis (SABA).
*(This document has been regularly updated. For more, contact Jack Michael, Western
Michigan University, Kalamazoo, MI 49008; (616) 344-3941, Email:
Evoke. Evoke means “to call forth”. A behavior can be evoked by an
unconditioned stilumus, conditioned stimulus, or discriminative stimulus. Therefore
operant and respondent behavior can be evoked by stimuli. Note that responses do not
Emit. Emit refers to the occurrence of operant behavior. Note that only organisms
emit responses (a school bell does not emit childrens’ dismissal).
Elicit. Elicit is used in the context of respondent conditioning and refers to the
reflexive relationship between:
Unconditioned stimulus==>Unconditioned response
Condition stimulus==>Conditioned response
We say that an Unconditioned Stimulus (US) elicits an Unconditioned Response (UR) and
a Conditioned Stimulus (CS) elicits a Conditioned Response (CR).
Negative reinforcement discusses one way to strengthen operant behavior:
(a) A stimulus ceases or decreases following a response
(b) This change increases the likelihood that such responding will occur again.
All forms of reinforcement increase the likelihood that responding will occur. (Positive
reinforcement is when the onset or increase in a stimulus, following aresponse, results in
an increased probability of that response in the future (in similar situation)).
Punishment involves either:
(a) A stimulus onset or increase following a response and such responses
become less likely.
(b) A stimulus terminates or decreases following a response and such
responses become less likely to occur.
So all forms of reinforcement make responding more likely to occur whereas all forms of
punishment make them less likely.
Including Emotions in Definitions of Reinforcement and Punishment
Definitions of reinforcement and punishment do not include references to emotions. It is
not the case the positive reinforcers are necessarily things people like or make them feel
good. Likewise, it is not the case that negative reinforcers are necessarily things people
dislike or make them feel badly. Most of the reinforcers and punishers we receive
hundresd of times each day elicit no emotional reactions.
Reinforcement and punishment are defined solely on the basis of how they affect behavior-
-emotional concomitants are not at issue.
“Reward” and “Reinforcer” are Not Synonymous
Reward and reinforcement are not synonyms. Reward is found within hypothetico-
deductive models and is therefore defined within the nexus of a theory’s operational
definitions. Reward also has residual meaning from Thorndike’s early learning theories
where, for example, rewards “stamp in” habits. These or similar meanings may be retained
in some theorizing. Radical behaviorists, on the other hand, prefer definitions that are less
theory laden, therefore they keep a clear distinction between reinforcement and reward.
Operant Behavior-vs-Operant Class
Operant behavior is controlled by its consequences. Specific examples of such behavior
are called “operants” that collectively constitute operant classes. Consequences affect all
members of a response class. The members of an operant class may not resemble each
other topographically and can vary greatly across individuals. For example, reinforcing a
child’s speaking may affect the time spent reading; or, reinforcing asking for a cookie may
affect looking in cupboards or asking for milk.
For practical reasons, all members of an operant class are seldom identified:
1. Behavior can be changed without identifying all operant class members.
Concerns that changing behavior A for the better may adversely affect behavior B have
been researched and shown to be unfounded.
2. Prohibitively long and expensive assessment is needed to identify operant
class boundaries. Such assessment would consume time and resources better spent
providing assistance and doing functional analyses of problematic behaviors.
These practical arguments should in no way be interpreted to mean that operant class
membership is unimportant and will never be needed to effectively carry out clinical and
educational practices. But at this point, identifying such membership typically is not
prerequisite to effective practice.
Discriminative Stimulus (Sd)-vs-Antecedent
An antecedent is any stimulus that occasions an operant response. For example, a child
who is shown a tractor may state “tractor”. The tractor has become an antecedent for
A problem with such training is that it does not ensure precise stimulus control. For
example, at another point, the child may call a pick-up truck a tractor. To precisely
delineate which stimuli will control a response, discriminative stimulus control should be
A discriminative stimulus, sometimes symbolized “Sd” or “SD”, also regularly precedes an
operant response but it was established in a process involving at least two contingencies.
For example, children learning the alphabet often confuse “b” and “d”. To teach “b”
(1) Positive examples: The teacher points to the “b” says “b”, child says “b”,
teacher fades verbal prompt
(2) Negative examples: The teacher places “x” before the student and asks “Is this
a ‘B’?” A “No” is praised; incorrect responses are corrected. This is repeated with other
letters substituted for x.
(3) Repeating steps (1) and (2) will train the student to discriminate “b” from all
other letters. At that point, “b” is a discriminative stimulus for stating the letter b’s name.
The two contingencies here are:
(1) Discriminative Stimulus Response Consequence
“b” ==> State “b” ==> Praise
Is this a “b”?
(2) Discriminative Stimulus Response Consequence
“x” or other letter ==> “Yes” ==> Corrective Feedback
(3) Discriminative Stimulus Response Consequence
“x” or other letter ==> “No” ==> Praise
The question “Is this a “b”?
In summary, developing a discriminative stimulus involves using contingencies with
positive and negative examples. See chapter 8 of Jack Michael’s book for more (e.g., the
evocative role of SDs , suppressing responses with SDs, SDs as motivative variables, etc.)
In behavior analysis, “aversive” is an adjective for “stimulus” or “stimuli”. An
aversive stimulus is a negative reinforcer (its onset following a response makes such
responses less likely to occur; its termination following a response makes such responses
more likely to occur). As such, “aversive” has a technical meaning typically not found in
The word “adverse” has no technical use in behavior analysis. So there are no
“adversive stimuli” or “adversive responses”. “Adversive” and “abverse” are part of the
legal system’s vocabulary, not that of behavior analysis.
Environments–not Individuals–Associate Events
A common error is to state that in classical conditioning, for example, “The subject
associates the CS and US”. This association occurs in the environment, not the individual.
Similarly, in operant conditioning, the subject does not “Associate the response and
reinforcer” or the antecedent and response; these sequences of events occur in the
environment. To suggest that the individual associates anything goes well beyond what
was observed and invokes an explanatory fiction to account for the behavioral changes.