Date on Master's Thesis/Doctoral Dissertation

5-2006

Document Type

Master's Thesis

Degree Name

M.A.

Department

Psychological and Brain Sciences

Committee Chair

Molfese, Dennis L.

Subject

Language and emotions

Abstract

How do words represent emotional states, and how closely are they clustered to an actual emotion? Opinions vary as to how many emotions there are, and whether (or which) emotions are basic. Generally, many agree that basic emotions include fear, anger, joy, sadness, and disgust. Some researchers include surprise, shame, interest, and others as derived emotions. However, all proponents of basic emotions confirm that each emotion reflects a unique motivational and behavioral tendency. These basic emotions are significant in that they represent distinct modes of action and are physiologically distinguishable. Yet, what of the myriad of other words humans use to describe those same emotions? Curiously, many researchers, including neuroscientists, have approached this topic without understanding how well "representative language" describes emotional states. This question prompts my inquiry of how closely correlated emotionally descriptive words are to each other. This study design is a triadic comparison of selected emotion words in two studies; a Positive/Negative word mix, then all Positive or all Negative word presentations to volunteer subjects. The purposes and outcome parameters of this research is to explore how emotion words actually cluster. Utilizing an online survey questionnaire, subjects were presented ten emotion words, three at a time, from which they were asked to select the two most similar. An important aspect of these studies was also to ascertain the efficacy of a triadic comparison, and the subsequent utility of Hierarchical Cluster Analysis in explaining results. The study population is generally described as university students, along with other volunteers to whom the study is interesting. Subjects are, in aggregate, considered to be somewhat representative of the larger population of like age, education, and background in the United States. The results were consistent in revealing coherence of emotion words and, ultimately, suggested improved methodology protocols for future investigations. While coherence was detected, often the logic of the connection among some emotion words was difficult to explain when subjects viewed a mixture of positive and negative words. Selecting from only positive or negative emotion words provided a more coherent clustering. The utility of triadic comparisons and Hierarchical Cluster analysis in determining what emotion words fit together was shown to be a useful research method for this kind of study. To reach an understanding of emotions and emotion states, research must include knowledge of the psycholinguistic structure and coherence of these emotion states. This study should prompt additional research to determine how emotion words relate and cluster to "core" or "basic" emotions similar to the ones mentioned above.

Download

Share

COinS