Sequential actions vs. simultaneous states.

The two thrice-repeated demonstrations, #12-14 where BLACK, BLUE and ORANGE BALLS successively collide and #18-20 where BALL IS ON TUBE AND TUBE IS ON PLATE, have similar event structures, in that there are three entities involved, the ‘middle’ one participating in two events. But the former involves sequential actions in time and the latter a simultaneous set of stative relations. It was predicted that sequential actions would lead to a corresponding ordering of NPs representing the entities involved to a greater extent than for the entities in simultaneous stative relationships; furthermore, this ordering should be more stable over repeated sentences describing the action situation than over sentences describing the stative situations. Table n gives the percentage of sentences in which the entities appear in first, second and third NPs as functions of both the perceptual ordering of the entities (in time for BLACK, BLUE and ORANGE BALLS and in space, from top to bottom, for BALL, TUBE and PLATE) and the ordering of successive sentences describing the same situations.

The predicted difference in strength of ordering of NPs does not hold for the first presentations of each type of situation; BLACK (57% 1st NP), BLUE (68% 2nd NP) and ORANGE (61 % 3rd NP) is matched by BALL (57% 1st NP), TUBE (64% 2nd NP) and PLATE (72% 3rd NP). Apparently there is a strong tendency to describe stative vertical relations from top to bottom. However, the ordering of NPs is much more stable across repeated descriptions of the same situation for sequential as compared with simultaneous relations; on the third presentation, we find BLACK (46 % 1st NP), BLUE (61% 2nd NP) and ORANGE (50% 3rd NP), despite instructions to generate different sentences, whereas we find BALL (23% 1st NP), TUBE (57% 2nd NP) and PLATE (34% 3rd NP) by the third presentation for the simultaneous display. It is noteworthy that the ‘ middle5 entity (participating in both events) is the most stable as the second NP in both sequential and simultaneous situations and that the perceptually polar entities rarely appear as the second NPs - as if the most available transformation involved a simple shift in polarity, from first to last and from top to bottom as initially encoded entities. There is also a marked difference between temporal and spatial patterns in terms of how entity relations are expressed, with relative clauses or with prepositional phrases: simultaneous stative relations (#18-20) produced an average of 95 % sentences with prepositional phrases (. . .on the plate) but only 28 % sentences with relative clauses, and most of these included prepositional phrases (. . . which is on the plate); sequential action relations (# 12-14), on the other hand, produced only 22 % sentences with prepositional phrases (. . . into the orange ball) as compared with 36 % sentences with relative clauses (. . . which hits the orange ball).