Project members: Jana Holsanova, Johan Blomberg, Peter Gärdenfors
This project analyses data from scene descriptions in films for blind audiences (Holsanova 1999, Holsanova et al., to appear) and applies theoretical framework of conceptual spaces (Gärdenfors 2000, 2014) and semantic structures of events (Gärdenfors & Warglien 2012).
The starting point is that each utterance (or idea unit) expresses an event.
First, we transcribe and code spoken descriptions of film scenes (Blomberg 2014, Holsanova 2001, 2008).
Second, we segment events in the film according to the event-indexing model (Zwaan, 1996; Zwaan, Magliano, & Graesser, 1995; Zwaan & Radvansky, 1998) and code for shifts in space, time, characters and objects in the film narrative.
Third, we correlate the segmentation of the visual stream with the segmentation of the spoken language description of the scenes.
Which of the event borders and shifts are explicitly verbalised by the interpreter and which are not? Which aspects of a visual event are coded in the rapid real time spoken language descriptions generated by the interpreters? Will the actions or the results of an action be in focus most of the time? The results will reveal a lot about how humans conceptualize and categorize events and motion.