For a ‘Cognitive Anatomy’ of Human Emotions and a Mind-Reading Based Affective Interaction

For a ‘Cognitive Anatomy’ of Human Emotions and a Mind-Reading Based Affective Interaction

Cristiano Castelfranchi (CNR, Italy)
DOI: 10.4018/978-1-61692-892-6.ch006
OnDemand PDF Download:
List Price: $37.50


Human emotions are based on typical configurations of beliefs, goals, expectations etc. In order to understand the complexity of affective processing in humans, reactions to stimuli, perception of our bodily reaction to events or just the feeling related to something should be considered but this is not adequate. Besides, our body does not respond just to external stimuli (events); it reacts to our interpretation of the stimulus, to the meaning of the event as well. In order to build affective architectures we also have to model the body, and its perception. In this chapter, with the help of these facts, the author will analyze the cognitive anatomies of simple anticipation-based emotions in addition to some complex social emotions.
Chapter Preview


In this chapter, we present, in a rather synthetic way and without the possibility of extensively discussing the literature:

  • a.

    An explicit and analytical cognitive modeling of human emotions (cognitive ‘anatomies’ in terms of beliefs, goals, etc.);

  • b.

    The limits of this fundamental approach, and the need for its embodiment: modeling and integrating the bodily motions and signals, and what we feel;

  • c.

    Its application to computational models, artificial intelligence (AI), and human-computer interaction (HCI).

The effects of complex emotions processed by humans go beyond reacting to stimuli, perceiving our bodily reaction to events, or feeling something. Especially complex human emotions are based on specific mental states; they are typical configurations of beliefs, goals, motives, expectations etc. In this chapter, we will analyze some typical mental configurations needed for (i) rather simple anticipation-based emotions (‘hope’, ‘fear’, ‘disappointment’, ‘relief’, ‘joy’) and (ii) complex social emotions like ‘shame’, ‘envy’, ‘guilt’, ‘pity’: their ingredients and their coherent structure. In particular, we will analyze shame and guilt in a very synthetic way.

We are in favor of a componential analysis of emotions (and in general, mental states and processes like 'expectation', 'need', 'trust', 'argument', etc.). This allows a systematic explicit model of the relationships within and among the substances to be modeled. However, one should also care about accounting for the unitary character of the mental-behavioral phenomena. On one side, being atomically decomposable, the complex mental states have their own emergent, specific, non-reducible properties and functions on the other side.

Our body does not respond to external stimuli or events based on pattern matching; it also reacts to our interpretation of the stimulus, to the meaning of the event; that is to a mental representation. In addition, the body reacts to merely endogenous representations, to mental events (like a counterfactual imagination). For example, it is always a thought that makes us blush. Of course for a complete real emotion, bodily activation and perception is necessary: at least in terms of the activation of the central memory trace of the bodily reaction (somatic marker), the evocation of some sensation. We feel our bodily response, but we ascribe it to that event or idea; this combination gives an emotional nature to both sides.


Expectations versus Predictions

‘Expectations’ are not just ‘Predictions’; they are not fully synonyms. Therefore, we do not want to use expectations (like in the literature) just to mean predictions, that is, epistemic representations about the future. We consider, in particular, a ‘forecast’ as a mere belief about a future state of the world and we distinguish it from a simple hypothesis. The difference is in terms of degree of certainty: a hypothesis may involve the belief that future p is possible while a forecast has the belief that future p is probable. A forecast implies that the chance threshold has been exceeded.

Putting aside the degree of confidence (a general term for covering weak and strong predictions), for us expectations have a more restricted meaning (and this is why a computer can produce predictions or forecasts but do not have expectations). In ‘expectations’:

  • i.

    the prediction is relevant for the predictor; he is concerned, interested, and that is why

  • ii.

    he is expecting, that is the prediction is aimed at being verified; he is waiting in order to know whether the prediction is true or not.

Expectation is a suspended state after the formulation of a prediction1. If there is an expectation then there is a prediction, but not the other way around.

Complete Chapter List

Search this Book: