its recurrence, but if the exception does not recur, at least the
response is less sure and less accurate than it otherwise would be. It
tends to destroy self-confidence or confidence in the one who allowed
the exception. Sometimes even one exception leads to disastrous
consequences and undoes the work of weeks and months. This is especially
true in breaking a bad habit or in forming a new one which has some
instinctive response working against it.

There has been a great deal of work done in experimental laboratories
and elsewhere in the study of the formation of particular habits. The
process of habit formation has been shown by learning curves. When these
learning curves are compared, it becomes clear that they have certain
characteristics in common. This is true whether the learning be directed
to such habits as the acquisition of vocabularies in a foreign language
or to skill in the use of a typewriter. Several of the most important
characteristics follow.

In the first place it is true of all learning that there is rapid
improvement at first. During the beginning of the formation of a habit
more rapid advance is made than at any other time. There are two
principal reasons for this fact. The adjustments required at the
beginning are comparatively simple and easily made and the particular
learning is new and therefore is undertaken with zest and interest.
After a time the work becomes more difficult, the novelty wears off,
therefore the progress becomes less marked and the curve shows
fluctuations.

Another characteristic of the learning curve is the presence of the
so-called "plateaus." Plateaus show in the curve as flat, level
stretches during which there has apparently been no progress. The