The Evolution
49
The problem with kick, of course, is that it is extravagant to
assume that there are as many definitions of kick as there are par-
ticularized versions of kicking—just as for instance it would be
extravagant to assume that there are as many definitions of red as
there are red objects. In each case the remedy is to assume that the
word, whether kick or red, has as its definition a general meaning,
and that it is this general meaning which yields the various par-
ticularized versions of "kicking" and "redness." This same solution
would, of course, work for container, since one may assume that the
word container yields concept "container," and that it is this general
concept which tends to yield "bottle," "basket," and the other forms
of "container."
The advantage of this simple outlook is that it accords with common
sense and is theoretically economical. It is also flexible enough to
handle words such as ball, which must have at least two general meanings
("round object" and "dance"), inasmuch as the word refers to a set of
objects and a set of events each of which has virtually nothing in common
with the other.
The above position is with some modification that of Anderson and
Ortony (see pp. 177-178). Anderson and Ortony go on to assert that an
intersection search (concerning which they cite Quillian, 1968, 1969)
might be the means by which memory arrives at a particularized image.
This would mean that cola container yields "bottle" by an intersection