and sounds,
whose paths to and from are indicated by _k_ and _l_, as well as by _g_
and _i_. These paths of connection must be of twofold sort. The
excitement can not pass off to the diction-center D on the same
anatomical path as the return impulse from D, because not a single case
is known of a nerve-fiber that in natural relations conducts both
centrifugally _and_ centripetally, although this possibility of double
conduction does occur under artificial circumstances. Apart, then, from
pathological experience, which seems to be in favor of it, the
separation of the two directions of the excitement seems to be justified
anatomically also. On the contrary, it is questionable whether the
impulse proceeding from D does not arrive directly at the motor
speech-center, instead of passing through W, S, or L. The diagram then
represents it as follows (Fig. 4). Here the paths of direct connection
_i_, _l_, and _n_ from D to M represent that which was just now
represented by _i_ L _d_ and _l_ S _e_ and _n_ W _f_, respectively; in
Fig. 4, _i_ conducts only sound-excitations coming from L, _l_ only
excitations coming from S, and _n_ only those coming from W, as impulses
for M. For the present, I see no way of deciding between the two
possibilities. They may even exist both together. All the following
statements concerning the localization of the disturbances of speech and
the parallel imperfections of child-speech apply indifferently to either
figure; it should be borne in mind that the nerve-excitement always goes
_only_ in the direction of the arrows, never in the opposite direction,
through the nervous path corresponding to them. Such a parallel is not
only presented, as I have found, and as I will show in what follows, by
the most superficial exhibition of the manifold deviations of
child-speech from the later perfect speech, but is, above all, necessary
for the answering of the question: what is the condition of things in
learning to speak?

[Illustration: FIG. 4.]

3. Parallel between the Disturbances of Speech in Adults and the
Imperfections of Speech in the Child.

In undertaking to draw such a parallel, I must first of all state that
in regard to the pathology of the subject, I have not much experience of
my own, and therefore I rely here upon Kussmaul's comprehensive work on
speech-disturbances, from which are taken most of the data that serve to
characterize the individual deviations from the rule. In that work also