The Evolution
38
(this would be an effective way of evading activation interference).
This raises the possibility that the doubling of the order effect that
occurs when an interval is inserted may be caused by this trick becoming
no longer available to memory, which in turn raises doubt about the
existence of activation asymmetry. In any case, however, from the
point of view of psycho-evolutionary analysis the issue is somewhat
academic, inasmuch as psycho-evolutionary analysis, though clearly calling
for the existence of activation interference (whose existence the above
experiment appears to confirm), is more or less neutral on the separate
question of whether activation asymmetry must exist as well.
Before beginning the next section it is necessary to dispose of a possible
alternative explanation for the results of Freedman and Loftus (1971). One
could argue that a noun category followed by a restricting letter is the
preferred order for yielding an intersected word because a noun must
be recognized and its concept yielded, whereas a mere letter need only be
recognized (i.e. the letter, unlike the noun, does not yield a concept).
Thus the simple kind of activation asymmetry could be assumed to be the
cause of the difference in observed reaction times.
Fortunately, however, Freedman and Loftus also tested their subject's
ability to carry out intersections using phrases such as food white,
seasoning white, and fuel liquid (i.e. a noun category followed by a
restricting adjective), and obtained essentially the same order effect
they observed earlier for a noun category with a restricting letter,
although admittedly the effect was somewhat smaller.