To the best of my understanding, the five factor model of personality comes from a factor analysis on a large list of adjectives that can be used to describe an individual's personality. It is validated, in some sense, by its usefulness in predicting behavior and life outcomes in a wide range of contexts.
It seems like something very similar could be done for emotion. I.e. list words describing a persons present state (angry, remorseful, sad, bitter, frustrated, etcetera) and do a factor analysis on those. Then, see if the resulting model (assuming that the analysis yields something useful) is predictive of shorter term / state dependent behavior.
Has any work like this been done?