This dissertation addressed two broad questions: a) is it possible to systematically describe dimensions of form on which montage varies in TV advertising, and b) given that dimensions can be identified, how do variations in form affect viewer responses to ads? The first question was addressed through a semiotic analysis of ad form, the second question through an experiment that tested for pace and angle effects on brand learning and brand attitudes. The semiotic analysis built upon the work of C.S. Peirce, Christian Metz, Gerard Genette, and Herbert Zettl. Using our normal experience of time and space as a kind of deep structural baseline, it formally defined such montage variables as pace, angle, horizon, color, the camera cut and camera movement, and paradigmatic/syntagmatic shot syntax. Mathematical or quasi-mathematical notation was developed to describe the range of values possible for these variables. In the experimental portion of the study, the effects of pace and angle were tested within a composite ad response model that combined two competing theories -- distraction theory (which predicts a negative distraction effect for abnormal pace and angle values) and the dual-mediation hypothesis (which predicts a positive attention effect) . The results were consistent with distraction theory, i.e., abnormal executions resulted in significantly lower brand learning and brand attitudes. / Ph. D.