of wine, water, oil, the nursing-bottle and others,
the sight of which excited him, but he united in one notion the
contents of the different sorts of bottles when what was in them was
white--i. e., he had separated the concept of food from that of the
bottle. Ideas are thus independent of words.

Certain as this proposition is, it is not, however, supported by the
reasons given for it by Kussmaul, viz., that one and the same object
is variously expressed in various languages, and that a new animal or
a new machine is known before it is named; for no one desires to
maintain that certain ideas are _necessarily_ connected with certain
words, without the knowledge of which they could not arise--it is
maintained only that ideas do not exist without words. Now, any object
has some appellation in each language, were it only the appellation
"object," and a new animal, a new machine, is already called "animal,"
"machine," before it receives its special name. Hence from this
quarter the proof can not be derived. On the other hand, the
speechless infant certainly furnishes the proof, which is confirmed by
some observations on microcephalous persons several years old or of
adult age. The lack of the power of abstraction apparent in these
persons and in idiots is not so great that they have not developed the
notion "food" or "taking of food."

Indeed, it is not impossible that the formation of ideas may continue
after the total loss of word-memory, as in the remarkable and
much-talked-of case of Lordat. Yet this case does not by any means
prove that the formation of concepts of the _higher_ order is possible
without previous mastery of verbal language; rather is it certain that
concepts rising above the lowest abstractions can be formed only by
him who has thoroughly learned to speak: for intelligent children
without speech are acquainted, indeed, with more numerous and more
complex ideas than are very sagacious animals, but not with many more
abstractions of a higher sort, and where the vocabulary is small the
power of abstraction is wont to be as weak in adults as in children.
The latter, to be sure, acquire the words for the abstract with more
difficulty and later than those for the concrete, but have them
stamped more firmly on the mind (for, when the word-memory fails,
proper names and nouns denoting concrete objects are, as a rule, first
forgotten). But it would not be admissible, as I showed above, to
conclude from this th