Category: Advanced Stuff ¤
  Author: Gary Wilkes ¤
   Title: When To Click With No Treat ¤

Dani Weinberg writes:
> I'd like to know more about this unconventional practice - 

Dani,

Look at the Click & Treat Training Kit, again - it's all there, in what has
been called an "entertaining" view of the topic. I've heard the production
values aren't too bad, either. <G>

On a more serious note -- arbitrary rules may help beginners to lay a
foundation, but to get to the next level, it is the animal's behavior that
dictates the rules, not the trainer. It is tempting to ignore the real
foundation of clicker training (behaviorism) which ignores supposed mental
states (Won't it confuse the dog if you don't give a treat after the click?)
, and focuses on WHAT is happening.(The dog attempted the behavior again,
with more force.)  To really master the application of operant conditioning,
 it requires adherence to a larger, softer rule -- "reality dictates
methodology." A valid question to ANY rule is "How come?".  When people
suggest that the click must ALWAYS be followed by a treat, my question is
"How come?" If their reasoning is that you must develop a firm association
between the conditioned reinforcer and the primary, they are correct.
However, it often means that they will do a great job of creating a
conditioned reinforcer -- and a limited job of actually using it. When Karen
and I started doing seminars together, she used to take the time to click and
treat for non-specific behavior, to make a proper link.(The rule of thumb in
the marine world)  My business required that I do it on the fly -- (The rule
of thumb in applied behavior modification) neither one is "correct" as an
absolute -- it is merely in context that one can evaluate them as
"appropriate" and "functional".  

So, the "click without a treat" can act as a particular type of information,
and as a trigger for variability - but there is no way to know if it is
"correct" or not unless you tie it to a context, and evaluate it against how
the animal's behavior changes. I can tell you that I have never seen a dog's
brain fall out, or attempt suicide as a result of this practice, so, as in
many other instances, clicker training is a very forgiving process, and you
can feel free to experiment. <G>

Regards,
Gary Wilkes