ich I carried on for a long while with variations of ordinary habits
of daily life, asking whether a habit associated with a certain
sensory stimulus can function automatically while dispositions for a
different habit, previously acquired, remain in the psychophysical
system. For instance, I was accustomed to carry my watch in my
left-hand vest pocket. For a week I carried it in the right-hand
pocket of my trousers and recorded every case in which I first
automatically made the movement to the vest. After some time the
movement to the right-hand pocket became entirely automatic. When it
was sufficiently fixed, I again put the watch in the left-hand vest
pocket and recorded how often I unconsciously grasped at the right
side when I wanted to see what time it was. As soon as the vest pocket
movement had again become fixed, I went back to the right-hand
trousers pocket. And so I alternated for a long while, always changing
only after reaching complete automatism. But the results in this case
and in other similar experiments which I carried on showed that the
new automatic connection did not extinguish the after effects of the
previous habit. With every new change the number of wrong movements
became smaller and smaller, and finally a point was reached at which
the dispositions for both movements were equally developed so that no
wrong movements occurred when the watch was put into the new
position.[22]

This problem has been followed up very recently in a valuable
investigation at Columbia University,[23] in which various habits of
typewriting and of card-sorting were acquired and studied in their
mutual interference. These very careful experiments also show that
when two opposing associations are alternately practiced, they have an
interference effect on each other, but that the interference grows
less and less as the practice effect becomes greater. The interference
effect is gradually overcome and both opposing associations become
automatic, so that either of them can be called up independently
without the appearance of the other. Many details of the research
suggest that this whole group of interference problems deserves the
most careful attention by those who would practically profit from
increased industrial efficiency.

Finally, in the experimental study of the problem of technical
learning, we cannot ignore the many side influences which may hasten
or delay, improve or disturb, the acquisition of industrial skill