My reviewer attributes to me a serial view of cognitive processes because I distinguish predictive from retrodictive evaluations of mental actions (Pliushch this collection, pp. 7-8). I do not think that this distinction commits me to serialism however. In my 2013 book, I propose that “a mind should primarily be seen as consisting of a hierarchy of control-and-monitoring loops, and their essentially dynamic interaction with the world, rather than as constituted by the successive states that emerge from this interaction”. Examples of how the dynamics at lower levels of representation can influence higher levels, and the converse, are discussed in chapters 11 and 12, where the case of schizophrenic delusions is analyzed. Hence, I have no problem with the view that low-level appraisal affects higher-level appraisals: these types of influences are part of what it is to have a hierarchy of control. This does not mean, however, that predictive appraisal and retrodictive appraisal should be conflated: they have a different evaluative function, and are based on different dynamic cues. This does not mean, either, that a concept-based judgment can easily influence an affordance-based appraisal. The difficulty of having a prolonged strategic control over one's feelings (based on what one knows, as in the anagram experiment), originates in the different roles of associative cues and inferential relations between concepts in mental activity.[3]
Iuliia Pliushch is right, however, when observing that I stick to the distinction between feelings and their propositional re-description. From the viewpoint of action theory, this distinction corresponds to the contrast between reacting and acting strategically. I subscribe also to her remarks on p. 6, according to which goal representations might change affordance-sensings. The point is: how sustained is this change? A conceptual re-description tends to modify one's representation of the context, and hence of one's goals, which might either favor or reduce further elicitation of feelings (for example, by being ashamed of having felt anger), and even inhibit the influence of feelings on action. This is the case for the participants' epistemic decisions in phase 2 of the Anagram Experiment discussed in the section 5.2 of the article. Their ability to control their feelings, however, cannot resist time pressure and/or divided attention in phase 3.
On the view that I propose, feelings can only be sustainably modulated by having other feelings replace them. There are both automatic and strategic ways of enhancing one's feelings through other feelings (see Proust 2014). Feelings can easily be enhanced by enriching the associative representations constituting an affordance. Deliberately suppressing them, or reorienting them to new targets, however, is very difficult (as rejected lovers know all too well). The Confucian moral practices offer a very good example of a strategic attempt to train new moral feelings in followers (see Reber 2013). As Rolf Reber shows in his fascinating analysis of what he calls critical feelings, strategically redirecting one's feelings to new targets can only be performed by manipulating the fluency of one's own re-descriptions and conceptual rules for acting morally. In other terms, the agents need to be trained until they entertain feelings of ease of processing (i.e., feelings of fluency) when activating target concepts and inferences, rather than merely trying to immediately subsume their own initial feelings under critical concepts.