The Evolution
45
after the concept "fruit" is activated in memory, that is responsible
for the difference in reaction times. And of course the same lack of
interlinking that prevents spreading-activation from working effectively
among those concepts containing "redness" can also be expected to hamper
spreading-activation among the class of all words beginning with the
same letter (see Collins & Loftus, p. 416).
The above explanation is clearly a possible way of explaining why
a noun category followed by a restricting letter or adjective is a more
convenient order for memory than the same items in reverse (basically
they are postulating a type of activation asymmetry). It is less
effective in explaining why the order effect should also be present for
simultaneously presented items (Loftus & Loftus, 1974), although admittedly
they could argue that even simultaneously presented items are read
successively, and that therefore it is still somewhat better to have the
noun category first so as to allow its time-consuming spreading-activation
a small head start it would lack if it were placed second. Whether this
very small head start can be regarded as sufficient explanation of the
.13 sec order effect observed for simultaneously presented items is
debatable however.
A more direct and compelling objection to the Collins and Loftus
outlook is that by not using two types of links they cannot explain how
the "Florida eggs" intersection of "orange juice" differs structurally
from the "Florida drink" intersection of "orange juice." To do so they
would have to draw a distinction between the relationship "eggs" has with