Generalization, abstraction, derivation, inference

[From: Peter Burke (3/2/95 4:07 PM CST)]

Inference should probably be reserved for drawing a logical conclusion.
An inference is elective. The step by which the perception "apple" is
derived from lower-level perceptions seems not to be elective. We can
insist on attending to the lower-level perceptions, but the perception
"apple" doesn't go away unless we turn up some perception that betrays a
I'll say derivation instead. (Generalization, abstraction--there must be
a good word.)

The better word, I think, is induction. Creating higher level
concepts from lower level inputs. This is something that needs
to be understood within the PCT context. How are the new
(higher level concepts) created, especially ones that will
match (in some sense) the higher level reference signal. I
think that it is often the case that when a new concept is
induced from observations, we, at the same time, see what the
concept "should" be (i.e., have a sense of a proper reference
level for it). An excellent book on the subject is by Holland,
Holyoak, Nisbett and Thagard: Induction: Processes of
Inference, Learning, and Discovery. MIT Press 1986.