Category: Theory, Terms & Abbreviations ¤
  Author: M. Shirley Chong ¤
   Title: Operant Conditioning Terminology ¤

   I assume you mean the training related definitions, not things 
like ROTFL! ;-)
   Respondent conditioning is also known as classical conditioning--
it's what Pavlov did with dogs, bells and mealtimes.
   Operant conditioning refers to the sort of learning that is under 
the subject's voluntary control (unlike salivation). The basic 
sequence of operant conditioning is that the subject offers a 
behaviour in order to operate on their environment and receive a 
primary reinforcer. (Dog sits in order to get trainer to click and 
treat)
   A conditioned reinforcer is a signal that means "what you want is 
coming up right now!" The signal can be anything (a word, a sound, a 
motion, a light flash, a clicker, etc). It serves as an event marker 
for the subject ("what you are doing NOW is what I will reinforce") 
and makes it unnecessary to have what the subject wants right there 
instantly--there can be a (small) time lag.
   A positive reinforcer is something the trainer gives to the 
subject as a consequence to an offered behaviour which increases the 
likelihood of the behaviour recurring. (dog sits, trainer clicks and 
gives nuked hot dog coin; dog sits again to try to make it happen 
again)
   A negative reinforcer is something that the trainer withdraws from 
the subject as a consequence to an offered behaviour, which increases 
the likelihood of the behaviour recurring. (trainer has shock collar 
on dog set at low intensity, just barely high enough to be irritating;
 dog heads away from trainer in pursuit of interesting smells; 
trainer shocks dog; dog happens to turn back towards trainer and 
trainer instantly stops shock; dog is more likely to turn back 
towards trainer in future--the dog's behaviour has been reinforced by 
the withdrawal of the shock)
   A primary reinforcer is something that the subject wants. It can 
be food, special touches, a toy, a game, the chance to go for a car 
ride, etc. It has to be what the subject wants, NOT what the trainer 
thinks or assumes the subject wants!
   Positive punishment is a consequence that the trainer gives in 
response to an offered behaviour by the subject that reduces the 
likelihood of the behaviour recurring again in the future. (dog jumps 
on trainer; trainer steps on dog's back toes; dog is less likely to 
jump on trainer again in the future)
   Negative punishment is something that the trainer withdraws from 
the subject as a consequence of an offered behaviour which reduces 
the likelihood of that behaviour recurring. (dog in crate at mealtime 
rakes door with her feet as her dinner dish approaches; trainer turns 
away until the raking sound stops; dog is less likely to rake door 
with feet)
   Variable Reinforcement is simply asking for two or more 
repetitions of behaviour before the subject gets reinforced.
   Fixed Reinforcement is a consistent and predictable (to the 
subject) schedule of reinforcement. For example, one sit = one treat. 
Sometimes trainers fall into predictable FRs with poor results (the 
dog knows that every third sit will get a treat--most dogs will start 
doing as little as possible for the two sits in between).
   For a far more in-depth discussion of these principles (and 
practical ways to apply them), a good starting point is DON'T SHOOT 
THE DOG by Karen Pryor and/or LADS BEFORE THE WIND by Karen Pryor.
   For an endlessly fascinating discussion of how to use these 
principles to analyze your training: ClickTrain!

M. Shirley Chong
The Well Mannered Dog
Grinnell IA