The paper deals with the interrelation between patterns in gaze movement when watching dynamic video clips and what is mentioned at what point when talking about events. This interrelation was investigated with respect to dynamic, live-recorded videoclips depicting everyday situations. Speakers of three languages were asked to view the clips and tell what is happening. Attention distribution to different aspects of the clips (causative actions) were measured in 2 identified areas of interest: the area where the agent is located and the area in which the entity acted upon is located. Contrary to studies on the production of single words or clauses relating to pictures (eg. Meyer & Dobel 2003), gaze movement to the areas of interest and the time at which they are mentioned are not directly linked, given real time presentations. Factors that drive patterns of attention and mention over time cross-linguistically will be presented.