Abstract: This dissertation investigates the use of verbal predicates to determine whether persons tend to exhibit preferred modes of expression that correspond to the concepts of preferred modes of representation asserted by Bandler and Grinder. This dissertation also tests the implied assumption that these preferred modes of expression are stable: over time; over reports of positive and negative experiences; and over reports of past, present and future experiences. Also tested was the Bandler and Grinder assumption that of the five identified modes, the three most often identified will be visual, auditory, and kinesthetic. In addition, five exploratory questions were asked:(1) will predicate usage indicate a least preferred mode of expression? (2) will an easily administered self-report questionnaire produce the same categorization of preferred mode as the categorization by predicates? (3) will preferred and least preferred modes of expression be related to personality characteristics? (4) will the addition of nouns to the predicates change the categorization of modes of expression? (5) will the categorization of preferred modes of expression be changed when based on obviously determined predicates as opposed to being based on predicates checked for use in context? Twenty-seven college students were exposed to a two- track audio tape. On the tape were six stimulus questions. The students' responses were recorded on the same tape. One week later, they were exposed to another tape with six similar but different stimulus questions. On the third week, the subjects were given the California Personality Inventory. On the fourth week, the subjects were given a self-report questionnaire. Results indicated that, in the population tested, people have a deliberate preference for words that reflect one sensory category more than another. However, in the population tested, all subjects identified as having a preferred mode had a kinesthetic preferred mode. These subjects showed stability over time; over reports of positive and negative experiences; and over reports of time reference: past, present, and future experiences (p<.001). The self-report questionnaire produced negative results, indicating there was not a match between categorization by predicates and categorization by self-report. There were nine significant correlations between personality characteristics and proportion of predicates used. Five of the nine positive correlations were between the proportion of auditory predicates used and the CPI measures of Well Being (Wb), Socialization (So), Achievement via Conformance (Sc), and Intellectual Efficiency (Ic). The common thread running through these scales are descriptions of people who tend to be seen in similar ways. The three most often identified modes were visual, auditory, and kinesthetic. These three modes accounted for nearly 100 percent of the predicates used. When determining preferred modes of expression, it was determined that including nouns did not affect categorization. Also, there was no difference in categorization of preferred modes when words were checked for use in context and when they were not.