Understanding Antecedents

 

Copyright, 2006, Kellie Snider

 

                There are different kinds of antecedents and they work differently.  You may have read or been told that if there is an antecedent and a behavior follows it, that behavior is classified as respondent[1] (classical) behavior.  Actually, respondent antecedents are only one kind.  Operant behavior has antecedent stimuli, too. 

 

Respondent (Classical) Antecedents

            The respondent antecedents are unconditioned stimuli (US) and conditioned stimuli (CS).   US are antecedent events in the environment that elicit behavior from an organism in that environment without the involvement of a consequence.  For example, if a speck of dust flies into your eye, you will blink.  If your doctor hits you in a certain spot below the knee with a rubber mallet your foot will kick.  If you hear a sudden, unexpected loud noise, you will jump.  Blinking, kicking and jumping are unconditioned respondent behaviors.  You did not have to learn or be taught to do them, you just do them when the stimuli occur. 

            Conditioned stimuli start out as neutral stimuli that have no effect on behavior but they morph into stimuli that elicit behavior.  CS are antecedent events in the environment that elicit behavior from an organism without the involvement of a consequence as a result of repeated pairing with a US.  Initially Pavlov’s dogs salivated when meat powder was placed in their mouths (US), but over time they began to salivate when they saw researchers in lab coats.  The lab coated researchers always showed up just before meat powder was inserted into their mouths, so they became conditioned stimuli.  The dog was conditioned to respond to the sight of the guys in lab coats just as if they were the meat powder.  This fascinating discovery led Pavlov to stop studying salivation and start studying reflexive behavior. 

            No reinforcer was required to get the dog to make this connection.  It would be some time before Skinner and his colleagues recognized operant conditioning as a behavioral principle.  In respondent conditioning, the antecedent controls the behavior.  In operant conditioning, the consequence for behavior determines whether or not it will happen.

Operant Antecedent Stimuli

            CS and US are not the only kinds of antecedent stimuli.  There are operant antecedent stimuli as well.  The most common kind in clicker training is the discriminitive stimulus.  It is abbreviated S^d, but you will often see it abbreviated as Sd because of the limited capacity of many computer email applications to create a superscript font.  The cue we teach an animal to respond to in a clicker trained behavior is an example of an S^d. 

            An S^d is an antecedent that lets the organism know that if he performs a certain behavior now, he’s likely to get reinforcement for it.  It helps him discriminate between a good time to perform a behavior and a less opportune time.  For example, if you’re trying to teach a dog to sit when you say, “Sit!” and not sit when you say, “Down,” the cue “Sit” is an S^d that says, “Sitting will probably pay off for you now.”  The cue “Down” is not an S^d for sitting.  If he lies down, he won’t get a treat.  Only if he sits.  There are many examples of S^ds in every day life.  A gas station sign is an S^d for the probable availability of gasoline in exchange for certain behaviors, including removing the nozzle from the hook and depressing the handle.  A restaurant is an S^d for the probable availability of food in exchange for ordering it from a menu. 

            A punishing stimulus (S^P) is an operant antecedent that lets the learner know that if he does the behavior it’s likely to result in an aversive consequence.  If you see fire and put your hand into it, it will hurt.  If you speak to your spouse when she has a certain look on her face, she is likely to yell at you.  The look on her face is the S^P.  If your mother says, “Don’t!” but you do, you might be spanked, depending on your mother’s discipline strategy. 

            It’s also important to know that some antecedents don’t do much of anything relevant to a given behavior.  I drive an elderly Ford Explorer and the “Check Engine” light comes on often.  It doesn’t mean anything.  It just goes on and off a lot.  An irrelevant antecedent is called an S^delta or an S^r. If I see the temperature gage inching over to the hot side, that’s an S^P that says if I continue to drive, it could mean trouble.  If I only see the “Check Engine” light, that’s an S^r.  If I see a blackbird outside my window before I get dressed, that is unrelated to what I will wear that day.  It is an S^r.

Motivating Operations

            Motivating operations (MOs) are antecedent conditions that make a reinforcer more or less reinforcing in a given situation.  There are two types of motivating operations.  Establishing operations (EOs) work to temporarily make reinforcers more reinforcing.  Hunger makes food more reinforcing.  Abolishing operations (AOs) work to temporarily make reinforcers less reinforcing.  Satiation or illness can make food less reinforcing.  MOs work together with S^ds in determining whether a behavior will be performed or not.  If you see a gas station sign it indicates that the opportunity to get gas there is probable, but if your tank is full, you don’t stop.  That opportunity isn’t valuable just then.  If the needle is getting close to the red zone, that is an EO that makes filling up with gas more reinforcing. 

So what? 

            What all this means is that you can’t assume a behavior is respondent just because it has an antecedent.  It’s important to know whether your behavior is respondent or operant before you can change it very effectively.  You have to determine what kind of antecedent you’re dealing with.  One way to do that is to see if the behavior goes away when the relevant consequence goes away.  If a pigeon always pecks a button when a red light comes on, but stops pecking when food stops being delivered as a reinforcer, you know the red light is not a CS or a US.  It is an S^d as long as food is delivered, but loses its power when food stops being delivered.  If a noise occurs and you jump no matter what happens next, you are probably dealing with a CS or a US. 

            If you treat a reinforced behavior as if it were respondent, you’ll have far less luck in changing it than if you recognize the reinforcers involved and adjust them.  A common time this issue comes up is in dealing with emotional behaviors like fear and aggression.  Many people have written that these responses are respondent, and it is commonly believed to be the case.  However, respondent treatments like desensitization take a long time to get results.  In my current canine aggression research with Jesus Rosales-Ruiz, it turns out that these behaviors diminish rapidly when you adjust their reinforcers.  We’ve been treating them like something they’re not.  That rarely works very well! 

            A behavior often is initially emitted when it is elicited by a US or a CS but as soon as it has a chance to get out into the universe, it has a chance to meet up with consequences.  If some of those consequences are reinforcing, the behavior is likely to stick around.  If the behavior sticks around as a result of its reinforcers, it is no longer respondent, it is now operant. 

            The bottom line is that you have to know what is keeping a behavior happening in order to be very effective in managing that behavior.  Understanding that there are different kinds of antecedents is one part of figuring it out. 

                 

   

      Copyright, 2006, Kellie Snider