Human perception appears to automatically segregate any given spatial scene into a figure and a ground. A figure is an entity that, among other things, possesses a dominant shape, due to a definite contour or prominent coloring. The figure stands out against the ground, the part of a scene that is relegated to 'background'.

The phenomenon of figure-ground organisation was pointed out by the Danish psychologist Edgar Rubin in 1915. Rubin proposed a number of perceptual differences between the figure and ground. For instance, a figure appears to be thing-like, a contour appears at the edge of the figure's shape, it appears closer to the viewer and in front of the ground, it appears more dominant and is better remembered. In contrast, the ground appears to be substance-like, is relatively formless, appears further away and extends behind the figure, is less dominant, and is less well remembered. Figure-ground organisation has been influential in cognitive linguistics, and has been generalised to language by Talmy with his notions of figure and ground, also known as reference object, and by Langacker with the theoretical constructs trajector and landmark.

Talmy has proposed that in linguistic terms, smaller and more mobile objects are typically interpreted as figures, while larger, more immovable objects which serve to locate other objects are typically interpreted as the ground. This observation accounts for the asymmetric behaviour of linguistically encoded spatial scenes. For instance, in the following utterances which are straightforward reversals of one another, while the first utterance is acceptable the second is normally judged as being semantically anomalous as indicated by the question mark.

1. The bike is near (the house)
2. ?(The house) is near the bike This suggests that the grammatical organisation of linguistically encoded spatial scenes reflects figure ground organization. While the subject position corresponds to the figure, the object position corresponds to the ground. The unnaturalness of the second sentence is due to the fact that an entity that would be more likely to serve as the ground in a spatial relation holding between a bike and a house is placed in the position associated with the figure