Page 155 - Zoo Animal Learning and Training

7.3 Two‐way Communication 127

VetBooks.ir station until a visual bridge is given. Tactile for the poor performance, labelling the animal
as distracted, aloof, messing with their minds,
markers are good communication tools when
animals are in close proximity to the trainer
ers relieve themselves of responsibility for the
and not looking at the trainer, such as when a etc. By placing the blame on the animal, train-
chimp performs a back inspection behaviour, outcomes, but miss valuable information
or a sea lion has its head underwater but its about how to increase motivation through
body is near the trainer. The most appropri- clear communication of cues and high rates of
ate event marker for the animal and current reinforcement.
conditions should be used. Some trainers believe since the marker is
Many trainers have said things like ‘I tried reinforcement for behaviour they don’t have
clicker training and it didn’t work’, or ‘We to provide a backup reinforcer. They incor-
haven’t started our training programme yet rectly call this a ‘variable schedule of rein-
because we don’t have a clicker’. There is no forcement’. However, if the marker is an
magic in the clicker; the magic is in the very effective conditioned reinforcer (as evi-
act of marking the behaviour and closing the denced by its ability to increase or maintain
gap between criterion performance and the behaviour on its own), the trainer is using a
backup reinforcer. Tightening the contiguity continuous schedule of reinforcement – at
between the behaviour and consequence is least until the reinforcing strength of the
clear communication that seems to help the bridge extinguishes. Each time the bridge is
animal understand the function of its behav- given without a backup reinforcer can thus
iour. In my experience, event markers like the be logically viewed as a respondent extinc-
clicker are great in many conditions. However, tion trial. Just as Pavlov paired the metro-
in some conditions a whistle, verbal, visual, or nome sound with meat powder to elicit the
even a tactile event marker might be a better dog’s salivation at the sound of the metro-
tool for the job. For instance, when training nome alone, trainers pair the marker stimu-
great apes or carnivores behind barriers, some lus to a backup reinforcer, often food. When
trainers find a verbal event marker can be a Pavlov stopped backing up the sound of the
better tool than a clicker or whistle. The verbal metronome with the meat powder, the met-
marker frees up the trainer’s hands to hold tar- ronome extinguished as an elicitor. Similarly,
gets, give hand cues, and deliver reinforcers. when trainers unpair the bridge and the
Plus, without a whistle in the mouth it is easier backup reinforcer, animals eventually stop
for a trainer to give verbal cues and prompts listening to the marker and begin focusing on
along with the event marker. more reliable, salient signals of the backup
In some circles it has become routine to use reinforcer to come. This is often the action of
a marker without a well‐established backup the trainer’s hand moving to the backup rein-
reinforcer. For instance, the animal may per- forcer, e.g. hand to feedbag. This visual bridg-
form three behaviours to criteria and receive ing stimulus can serve the function of the
an event marker for each behaviour, but only intended marker, keeping the animal in the
receive a food reinforcer after the third or training environment a bit longer. However,
fourth behaviour. This is a very common rein- eventually the low rate of backup reinforce-
forcement strategy at zoos, and it is also one of ment reduces motivation and the animal
the most confusing reinforcement strategies either leaves the session, shows aggression,
for animals. I have found that animals trained or the behavioural response deteriorates to
with this type of inconsistent pairing of bridge the point the trainer ends the session.
and backup reinforcer often lose motivation A keeper with a long history of training a
to participate in the training session, exhibit a male gorilla (G. gorilla) with this inconsistent
high level of incorrect responses to cues, and pairing of the marker and backup reinforcer
show frustration‐induced aggression. When agreed to participate in a small experiment.
this happens trainers often blame the animal She first demonstrated her usual training