Page 36 - Zoo Animal Learning and Training

8 1 Learning Theory

VetBooks.ir and a conditioned stimulus, it was Skinner in part but discarding the punishment part
(Catania 1998).
the 1930s who found that if a reward is given
Today we know that punishment decreases
for performing a target behaviour, the rate of
that target behaviour increased the next time the rate of behaviour. Thus, like reinforcement,
it is performed. This was a very important you need to know/look at the behaviour after­
discovery and although it might seem simple, wards to define punishment. For example,
it is not. For example, by giving your dog a if I slap someone for grabbing me, am I punish­
treat, you are not using the concept of rein­ ing them? It’s possible, but it’s also possible that I
forcement. For you to know if you are using could be reinforcing it, just watch the movie
reinforcement, you must know what is hap­ 50 Shades of Grey! Only if that slap decreases
pening to a behaviour after giving the reward. the likelihood of grabbing can one call it punish­
If you don’t know what is happening then you ment. In summary, reinforcement increases the
don’t know if the reward is reinforcing the likelihood of a behaviour occurring in the future
target behaviour or not. So, reinforcement is and punishment decreases the likelihood of a
defined by increasing behaviour. For the sake behaviour occurring in the future.
of my favourite undergraduate professor, Dr Skinner (1953) identified four basic pro­
Hank Pennypacker, I should also point out cedures of operant conditioning. Table 1.1 is
that when we speak about reinforcement, we helpful for understanding the types of oper­
should state that we are reinforcing the target ant conditioning. You will notice the words
behaviour and that we are not reinforcing the positive and negative. These words do not
subject. Dr Pennypacker would always say, mean ‘good’ and ‘bad’. Instead, you should
‘Reinforcing a subject would entail a metal rod think of these words as mathematical terms;
and concrete’. You never forget that image. ‘positive’ means adding a stimulus to the
On the opposite side of the coin is punish­ situation, and ‘negative’ means taking away
ment. In an earlier version of his law of effect, a stimulus.
Thorndike mentioned that behaviours which In positive reinforcement, a stimulus follow­
are punished were eliminated. In experi­ ing a behaviour is added which increases the
ments on human verbal learning, he found likelihood of that behaviour. For example, it is
that if someone said ‘right’, the subject had an positive reinforcement when someone gives
increase in rate of responding. However, if an elephant a piece of watermelon for touch­
the experimenter said ‘wrong’, the effects ing her/his foot to a target (operant behav­
were similar to if the experimenter said iour), and then that elephant touches her/his
nothing. Thorndike took these results as evi­ foot when the target is presented more often
dence against punishment being an effective (increases in response strength).
process to stamp out behaviour (Catania Negative reinforcement is when a response
1998). Thus, he replicated the second half of results in the removal of an event, and the
the law of effect, keeping the reinforcement response rate increases. This is ordinarily

Table 1.1 Overview of the four basic procedures of operant conditioning.

Effect on behaviour
Increases the likelihood of Decreases the likelihood of
Consequence future behaviour future behaviour
Stimulus is added (+) Positive reinforcement Positive punishment
Stimulus is removed (−) Negative reinforcement Negative punishment