This paper describes a framework for making sense of the relationship between emotions, cognition, mathematical activity, and mathematical beliefs. Subjects were participants in a special mathematics anxiety program. Beliefs are classified as instrumental and relational approaches to a situation. Emotion is also distinguished as a two-faceted construct which includes the physiological arousal (emotion as state) and the cognitive construction (emotion as acts). Perceptions of participants are described in detail to illustrate the nature of shifts in beliefs about mathematics and to elaborate on a framework for interpreting the relationship between mathematical beliefs and emotional acts. Some cases of individuals who made no change and became more anxious as the program progressed are discussed. Anxiety is not an inherent response to mathematics, but is based on an individual's beliefs about mathematical activity. (YP)