Modeling emotional tone in stories using tension levels and categorical states
- 60 Downloads
The objective of this work is not to replicate subjective impressions and certainly not to supplant them, but to explore means by which the second dimension of literary impact, qualities of emotional expression, can be objectively studied through the collection and display of measures made possible by the computer. With that goal, this paper has illustrated two approaches to the analysis and display of three fundamental emotional tone scores. The first is the production of a combined score, tension, which has been derived from previous studies of literary text and criterion passages. The second approach is the generation of transition graphs which identify the emotional state of passages of text according to the categories proposed in Mehrabian's theoretical system.
Both of these approaches to the modeling of emotional tone scores generate meaningful displays of data which can be used in objective comparisons of different stories and which lead to fresh interpretations of the reasons for their impact on a reader. They can be applied to actual samples of the kind of literature that is spontaneously read for pleasure in addition to being of interest for analytic purposes.
Unable to display preview. Download preview PDF.