On the basis of empirical findings about the validation of the affecti
ve Gottschalk-Gleser content analysis scales a prototheory of the emot
ional influence on the content of speech is illustrated with reference
to the briefly described speech-production-model of Mannheim. The eff
ects of state and trait emotions on the single, basically sequential,
processing stages of speech production are explained applying Bower's
emotional associative network model. State emotions show, normally, a
strong effect on the articulation component because the voice seems to
be an integral part of a biologically anchored emotional expression s
ystem. State emotions as well as trait emotions exert a general select
ive influence on the content of speech via the processing components o
f focussing and selection. Emotions appear to control working memory a
ccording to their intensity and in this way to codetermine the content
of speech.