Leo is puzzled about the approximately 4:1 ratio of effect of
reinforcement to punishment. He asks:
>Does this mean that we must generate four times as much energy to guide
>behaviour with this method, than if we use punishment?? If this is true,
>and honostly I do not have reasons to doubt, it explains much.
I will try to be clear about what is a somewhat complex question. Let me
divide it into a couple of parts, starting (because it's easier) with the
last part of Leo's question:
LAST PART: Yes, it does explain much. It explains, for example, why
parents (and governments and bosses) are fond of using punishment as a
behavioral control. It's easier for the "controller" (or appears to be.)
Effective punishment generates an almost instant "stop what you're doing"
response -- so, for the punisher, the effect of using punishment is
immediate, certain, and often personal (remember those principles?) Thus,
the effects of the use of punishment can be very reinforcing for the one who
punishes.
However, the less immediate consequences of punishment nearly always create
more need for behavioral control (which means: need for further punishment,
since that's what the punisher has been reinforced for doing.) Here's a
little story about that:
A young mother, driven to distraction by her two sons' use of swear words,
had tried everything to teach them not to swear. Finally, in desperation,
she went to a counselor who suggested a swift, forceful use of punishment
the next time one of the boys used bad language. "Okay," she thought, "this
is something I can do with pleasure -- I'm so frustrated, I'll just use my
anger against them. Great!"
The next morning, at breakfast, she asked her older son, "What would you
like for breakfast?" He replied, carelessly, "Oh, just give me some of
those damn corn flakes." Rising to her full height, the mother delivered
the swift, forceful punishment she had been itching to give out -- smacking
the boy so hard with her hand that he was sent flying off his chair and slid
across the kitchen floor to end up in a heap in the corner.
Feeling pretty good about herself, she then turned to the younger son and
asked, "Now, what would you like for breakfast?" He looked fearfully at his
brother lying in the corner whimpering, then said, "I'm not sure -- but I
sure as hell don't want any of those damn corn flakes!"
To summarize: punishment may be four times as powerful as reinforcement in
terms of its power to change behavior -- but in terms of effectiveness in
creating the change desired, it is pitifully less effective. That is, it
will create dramatic change -- but it probably won't create the change
desired (and it will definitely create unintended change as well.)
FIRST PART:
With the preceding explanation in mind, you are probably now able to answer
the first part of the question yourself.
My answer is, No, it does not take four times as much energy to change
behavior using reinforcement than it does using punishment. It takes much
less energy. That's because while punishment is powerful, it is very poor
at directing behavior (which means it must be used again and again and
again -- perhaps never "teaching" the desired behavior) and it is very good
at creating other unwanted behavior that then must be dealt with.
In the classroom, when I taught behavioral principles, I used to send one
person out of the room while the rest of the class chose a path we wanted
that person to follow -- something like "She should walk in and immediately
turn right, walk along behind this row of tables, stop here, turn around in
a circle, then continue to that window." The only way of communicating to
the person was through simple reinforcement (by ringing a small bell when
she was doing something right and not ringing it when she was doing
something wrong.) The person would return and one of her classmates would
"teach" her the path by ringing the bell. It usually didn't take more than
a couple of minutes for her to successfully follow the chosen path.
We would try it again with another person using punishment instead (hitting
her on the back with a rolled up piece of newsprint paper when she did
something wrong, not hitting when she did something right.) We were almost
never successful getting her to follow the path and had to give up trying in
every case but one that I ever witnessed over several years of running the
class. Why?
When you asked the person being "taught" what was going on, it became clear
how reinforcement guides behavior while punishment doesn't -- and how being
punished creates emotional responses that interfered with both ability and
willingness to learn. The simplest statement of how reinforcement and
punishment differ I heard was this:
When I did something right and you rang the bell, it was so easy -- I just
kept doing what I was doing. When you stopped, I stopped and tried other
things until the bell rang again. When I got whacked with the paper, I
didn't have a clue what you wanted me to do -- I just knew you didn't want
me to do what I had been doing.
So -- a well-timed ("contingent") reinforcement makes it very clear what the
subsequent behavior should be. An equally well-timed punishment gives
almost no guidance about what the behavior should be -- out of all the
thousands of things I could be doing, all I know is that this one [or the
way I've interpreted what "this one" is -- remember the damn corn flakes]
should be stopped.
SUMMARY
Since punishment often fails completely to guide behavior, one could say
that it would take (nearly) infinite energy to guide behavior using
punishment. By contrast, well-timed reinforcement, even if it must be
applied four times, gets the message across quite efficiently.
--"John Gunkler" <jgunkler@sprintmail.com>
Learning-org -- Hosted by Rick Karash <rkarash@karash.com> Public Dialog on Learning Organizations -- <http://www.learning-org.com>