Imagery in L2 captioned video: Investigating incidental vocabulary learning from extensive viewing as a function of modality, contiguity, and spacing

The aim of this thesis is to study the role of imagery in L2 captioned video by examining modality (Study 1), contiguity (Study 2), and spacing (Study 3) effects in incidental vocabulary learning from extensive TV viewing. An experimental design was employed in which one hundred seventy-three Algerian EFL learners in their third year of the Linguistics Bachelor programme were randomly assigned to either a Control, View, or Non-View group. Treatment participants watched two full-length seasons of documentary series extending to eight viewing hours, over a six-week period of two-week intervals. The View group watched the episodes in the form of L2 captioned video while the Non-View group had the imagery hidden and were therefore exposed to L2 audio and L2 captions only. Four levels of word knowledge were measured: meaning recall and recognition (posttest only) and spoken and written form recognition (pretest-posttest).
Study 1 assessed the effect of obscuring imagery on incidental learning of twenty words using a between-participants design. The results showed successful word learning regardless of the presence of imagery. Study 2 investigated the effect of verbal-visual contiguity (the co-occurrence of a word and its visual referent) on incidental learning of twenty-eight words using a within-participants design (View group only). It introduced contigfrequency, contigduration, and contigratio as three measures of contiguity on two timespans (∓7 seconds and ∓25 seconds) that were longer than those used in previous studies. The results showed that the amount of time visual referents appeared on the screen (contigduration), measured in a ∓25 second timeframe relative to the verbal occurrence, was predictive of learning. These results were more pronounced in the meaning recognition test. Study 3 explored whether words would be learned better when their occurrences were spread across viewing sessions (spaced condition), as compared to appearing within a single session (massed condition) by measuring the incidental learning of eight matched word pairs using a between-items design. It also examined whether learning in these two spacing conditions was influenced by the presence of imagery. The results revealed a positive effect of spaced occurrences in the Non-View group but not the View group, suggesting that a spacing advantage is more likely when fewer cues are available. These results were limited to knowledge of meaning only.