Page 39 - Zoo Animal Learning and Training
P. 39

1.4  Operant Conditioning  11

  VetBooks.ir  and behaviour with a fixed ratio 5 (written as   by the military in World War II to train
                                                      pigeons  to  guide  missiles  (Skinner  1960).
             FR 5), then we would give him a piece of
             sweet potato after every fifth response: the
             number of required responses is fixed at 5.  During this time, Skinner took a leave of
                                                      absence  from  the  University  of  Minnesota
                                                      and moved his lab into a flour mill provided
                                                      by General Mills. Taking breaks from train­
             1.4.3  Extinction
                                                      ing the soon to be missile guiding pigeons,
             Extinction was discovered through another   Skinner and his students noticed a wild flock
             serendipitous event. During one of Skinner’s   of pigeons hanging around the mill. One day
             experiments on satiation, a rat was pressing   in 1943, Skinner and his students decided to
             the lever when the pellet dispenser jammed.   teach a pigeon to bowl (Skinner  1958). To
             Skinner was not in his lab at the time and thus   achieve this behaviour the pigeon needed to
             couldn’t simply fill the dispenser. However,   first swipe a ball with its beck. The group got
             when he returned he found an interesting   prepared to  reinforce the first  swipe, but
             curve. Skinner states in his book, ‘The change   nothing happened. In the hopes of speeding
             was more orderly than the extinction of a sali­  up the process they decided to reinforce sim­
             vary reflex in Pavlov’s setting and I was terri­  ply  looking  at  the  ball,  then  a  behaviour
             bly excited. It was a Friday afternoon and   ‘which more closely approximated the final
             there was no one in the laboratory whom I   form’ (Skinner 1958, p. 94). Although Skinner
             could tell. All that weekend I crossed streets   had used shaping before, this was the first
             with particular care and avoided all unneces­  time he had hand‐shaped an animal to per­
             sary risks to protect my discovery from loss   form a behaviour (Peterson 2004). It wouldn’t
             through my death’ (Skinner 1979, p. 95).  be until 1951 when an article by  LOOK
               Extinction is defined as withholding   magazine was published that showed Skinner
             reinforcement from a previously reinforced   training a Dalmatian to jump, that the pro­
             response. As a training process, during   cedure of shaping was disseminated to the
             extinction, there is a zero probability of rein­  general public.
             forcement for the response you were previ­  Today the term shaping is widely known
             ously reinforcing. As a behavioural process,   to anyone that has tried to train their dog or
             extinction is a decline in the rate of respond­  has taken an introductory psychology
             ing caused by withdrawal of reinforcement.  course. Shaping is training closer and closer
               The  interesting curve that Skinner found   approximations to the target behaviour. For
             was probably an extinction burst. When the   example, a rough shaping plan for getting a
             extinction procedure has started, a sharp   dog  into a  crate  on its own  using  shaping
             increase in responding occurs, followed by a   may look like this:
             slow decline in responding as the trials con­  Start  with  the  dog  2 ft  (60  cm)  from  the
             tinue. An example that you have probably wit­  crate. Decrease the distance between the dog
             nessed is when you wait for an elevator. If you   and the crate. When the dog is close enough,
             press the button and the doors don’t open, you   have the dog touch the crate with its paw.
             press it again. You are then frustrated so you   After the dog is touching the crate you might
             press it 10 times in a row (extinction burst). If   ask the dog for one paw in the crate and then
             the doors still don’t open, you might press it   increase the distance into the crate until the
             one or two times more before taking the stairs.  entire dog is in the crate. Finally you would
                                                      work on closing the crate door whilst the dog
                                                      is inside.
             1.4.4  Shaping
                                                       Shaping allows animal trainers to set out a
             The story of Skinner’s discovery of shaping is   plan that will bring the target behaviour to
             quite interesting so I thought I would include   fruition. However, more importantly, shap­
             it in this chapter. Skinner was commissioned   ing is an essential process in teaching because
   34   35   36   37   38   39   40   41   42   43   44