‘1
For instance, as illustrated by the names of large numbers,
words and peanings can be generated by rules as the need
arises. This implies that recalling exemplars, for example,
of 'things that I own' may be a generative task as much as a
memory task, and that various generative strategies may be
available for use.
A second point is that synonymy should not be the goal
of a semantic analysis. Such a goal resales in a logical
paradox. If the meaning of a concept is identical to the
meaning of the concepts into which it is analyzed, then the
finding is trivial. If it is not identical, then the
finding is wrong. Rather, the goal of the analysis should
be to determine semantic entailment. This would be a
limited goal, in that there would be no claim to an
exhaustive understanding of the concept. For example,
"birdness" entails "wingedness" because to determine whether
or not a thing is a bird, one must do all of the mental
computations required to determine whether or not it has
wings. Operationally, the use of entailment in the present
empirical study of ownership involved measuring which
criteria of ownership regularly co-occurred with judgements
that something is owned.
Finally, there is a concern with two levels or modes of
semantic processing. One is oriented towards formal, exact