Eriksen reinterpreted PD as a RESPONSE EFFECT, not a PERCEPTION EFFECT. This is how he did it in behavioral terms. Imagine the sequence:
STIMULUS->PERCEPTION->THOUGHTS AND ASSOCIATES->RESPONSE
For example, a stimulus flashed in a t-scope task is perceived as being a four letter word that ends in k and has an o in it. For example, _o_k. This perception leads to as series of thoughts and associates as to what the work might be. For example, dock, took, pork, etc. These thoughts and associates have the theoretical status of being CUE PRODUCING RESPONSES, according to Miller & Dollard, prominent behavioral theorists at the time. Of importance here is the assumption that these THOUGHTS AND ASSOCIATES differ in strength depending on a persons past experiences with the THOUGHTS AND ASSOCIATES being reinforced/punished.
So the perception, _o_k might produce the following HIERARCHY of THOUGHTS AND ASSOCIATES (strongest at the top, weakest at the bottom)
dock
took
pork
cook
lock
look
hook
The next flash of the stimulus in the t-scope might change the perception to _ook. The subsequent hierarchy of THOUGHTS AND ASSOCIATES would then be in the same order as above, but with the THOUGHTS AND ASSOCIATES pork and lock removed (and any incorrect guess removed). Those thoughts and associates at the top of the hierarchy are there because the individual has experienced reinforcement in association with them. Thus the person would be vigilant in recognizing them. Those thoughts and associates at the bottom of the hierarchy are there because the person has experienced punishment in association with them. Thus the person would be perceptually defensive in recognizing them. Because these thoughts and associates are considered to be responses, this interpretation of PD removed some of the magic and Mystery from the PD EFFECT. WHAT HAD BEEN VIEWED AS A PERCEPTUAL EFFECT was now interpreted as NOTHING MORE THAN THE EMPIRICALLY ESTABLISHED EFFECTS OF PUNISHMENT ON THE OCCURRENCE OF RESPONSES.