A Two-Level Hypothesis
Lieberman (2000, pp 167-170) supports a two-level hypothesis in which there are “two distinct systems in conditioning: a relatively primitive system based on associations, and a cognitive system based on expectations” (bottom of p. 169).
He discusses the report by Bechara et al., (1995) [abstract] which compares the effect of a conditioning procedure used with a subject with damage to the amygdala with the effect of the same procedure used for a subject with damage to the hippocampus.
The participant with damage to the amygdla was able to say that the loud blast of a horn was always preceded by a blue slide, but showed no galvanic skin response to blue slides, as normal participants would.
The participant with hippocampal damage showed a normal conditioned GSR to the blue slides, but was unable to verbally identify which color preceded the noise.