A study of linguistic references to emotions in naturally occurring conversations found that references to unpleasant emotions outnumbered references to pleasant emotions almost two to one.Intellect is to emotion as our clothes are to our bodies; we could not very well have civilized lge without clothes, but we would be in a poor way if we had only clothes without bodies.-Whitehead (20, p. 232) Expressing feelings and responding to their expression by others has been identified as one of five major communication competencies (1).Yet little is known about what constitutes competency in expressing emotions, especially through verbal behavior. For example, do conversants talk more about pleasant emotions or unpleasant ones, and are they inclined to talk about emotions in more personal and immediate terms or in more impersonal and distant ones?If communication competence requires knowledge of rules and the ability to conform to them (I, 13) and if the identification of typical patterns is a precursor to inferring rules (18), then knowing what commonly occurs is essential to understanding what constitutes communication competence. Specifically, if conversants talk more about pleasant or unpleasant emotions and if they use predominantly personal or impersonal forms, these patterns may reflect display rules regarding such talk. Deviations from these patterns might then be used to define competency. For example, people who are described as either "sugarcoated" or "cynical" may deviate in their conversational behavior from typical ratios or frequencies in mentioning pleasanthnpleasant emotions. Similarly, people who deviate from the typical pattern for use of personal and impersonal forms may be perceived as excessive (too "wrapped up" in themselves) or repressive (not "in touch" with their feelings) in their disclosures.To verify the existence of rules for behavior, we need three types of evidence. We must demonstrate that the behavior (a) is controllable, (b) occurs with regularity, and (c) is prescribed or proscribed (18). The fact that people can and do learn to monitor their emotional expressiveness (7,12, 17) demonstrates that such behavior is controllable. Because there are unpleasant consequences for deviating from social rules, people are more likely to conform to rules than to deviate from them (18). Therefore, if a behavior is prescribed by a rule we would expect it to occur more frequently (or, if proscribed, less frequently) than a similar behavior not so governed.' Consequently, frequency counts can be profitably used to pinpoint which behaviors are more likely than others to be governed by rules.This study was designed primarily to address the second type of evidence: that is, to determine whether certain emotions are expressed more frequently than others with statistical regularity. In addition, some qualitative observations are included to provide limited evidence that at least some of the statistical patterns may be prescribed or proscribed. Some rules (e.g., for marriage ceremonies) may have a limited d...