POINTING AND LABELING DIRECTIONS IN EGOCENTRIC FRAMEWORKS

In three experiments, people learned descriptions of layouts and were then tested on the locations of objects in the layouts. Learning occur red either by using arrow-keys (a pointing-like procedure) or naming d irections (front, back, right, left) to explore the layout, and testin g involved either arrow-keys or verbal responses. In Experiment la, us ing arrow keys during learning and testing, right/left judgments were made faster than front/back judgments (reverse pattern). In Experiment 2a, direction labels were used during learning and testing, and front /back judgments were made faster than right/left, replicating the stan dard pattern obtained by Franklin & Tversky (1990). In Experiment 3 su bjects learned the layout using direction labels and were tested with arrow-keys and the reverse pattern was found again. Response biases (E xperiments Ib and Ic) and phonological biases (Experiment 2b) were rul ed out as explanations for the patterns obtained. The pragmatics and r equirements involved in each modality accounts for the accessibility p atterns obtained with pointing and verbal labeling. (C) 1996 Academic Press,Inc.