Mood is the longer term (>1 second) emotional state of the virtual human. It shares many of the considerations of emotions discussed in the previous section.
A useful model is that of PAD space - pleasure, arousal, dominance, where:
- The "pleasure" dimension is how "happy" or "unhappy" the person or virtual human is
- The "arousal" dimension is how "emotional"/active/external they are about the emotion - so for instance a bot may be on the "unhappy" pleasure dimension, but if they have low arousal they are quite and this is shown as "sad", whereas is they have high arousal they may show as "angry"
- The "dominance" dimension is how in control the bot or virtual human is, so the unhappy/high arousal person may be angry if in high dominance (they think they can make a difference) or fearful if in low dominance (they think there's not much they can do about it).
So for instance a depressed virtual human (Marvin?) may be "unhappy" on the pleasure dimension, low on arousal but very low on dominance as there's not a lot they feel they can do about things.
Another useful concept to bear in mind is that without further input mood can be taken as heading back towards a nominal position - and this nominal position is a reflection of the user's personality. We often think in terms of half-life, the bot halving its deviation from the norm for each time increment (and those time increments may depend on the emotion/mood etc).
- Up to Mind
- On to Personality