The Impact of a Human Figure in a Scene on Spatial Descriptions in Speech, Gesture, and Gesture Alone

The presence of a human figure in a scene appears to change how people describe it. About 20% of participants take the human figure's viewpoint (Tversky and Hard in Cognition 110:124-129, 2009. http://doi.org/10.1016/j.cognition.2008.10.008). Five exploratory studies compare descriptions of a scene with no person to descriptions of a scene with a person. About 20% of participants are predicted to use the person's point of view in the "person" conditions. Study 1 replicates the original pattern. Study 2 shows that the pattern holds when object/scene are changed, and that the figure's gaze towards/away from the object does not change the pattern. Studies 3 and 4 show the pattern holds when the object has different positions and when it is moving. Study 5 shows the pattern holds when the describer is talking to an interlocutor, in both speech and co-speech gesture, and when the person is using gesture alone. The presence of a human figure in a scene appears to be a robust variable in shaping spatial descriptions.

Spatial language and converseness

Typical spatial language sentences consist of describing the location of an object (the located object) in relation to another object (the reference object) as in “The book is above the vase”. While it has been suggested that the properties of the located object (the book) are not translated into language because they are irrelevant when exchanging location information, it has been shown that the orientation of the located object affects the production and comprehension of spatial descriptions. In line with the claim that spatial language apprehension involves inferences about relations that hold between objects it has been suggested that during spatial language apprehension people use the orientation of the located object to evaluate whether the logical property of converseness (e.g., if “the book is above the vase” is true, then also “the vase is below the book” must be true) holds across the objects’ spatial relation. In three experiments using sentence acceptability rating tasks we tested this hypothesis and demonstrated that when converseness is violated people's acceptability ratings of a scene's description are reduced indicating that people do take into account geometric properties of the located object and use it to infer logical spatial relations.