Representational Momentum and the Landmark Attraction Effect

The effect of a large stationary landmark on memory for the location of a smaller moving target was examined. Forward displacement of the target was larger when the target moved toward the landmark than when it moved away from the landmark. Target size and direction of motion also influenced displacement. When the target passed close by the landmark, forward displacement of the target was larger before the target passed by the cardinal axis of the landmark than after the target passed by the cardinal axis of the landmark. Memory for a target that passed by a landmark was also displaced toward the target along the axis orthogonal to motion. Control experiments ruled out biases toward the centre of the screen as causing the differences in displacement. The data support the hypotheses that representational momentum may combine with landmark attraction effects to influence the displacement of a target along the axis of motion. Spatial memory exhibits a number of consistent biases. One type of bias distorts memory in ways consistent with the operation of environmentally invariant physical principles. For example, memory for the final position of a moving target that vanishes without warning is often distorted forward in the direction of anticipated future motion, and this bias has been referred to as representational momentum (e.g., Freyd & Finke, 1984; Hubbard, 1995b). A second type of bias distorts memory by decreasing the remembered distance between a target and a landmark (e.g., McNamara & Diwadkar, 1997; Sadalla, Burroughs, & Staplin, 1980; Tversky & Schiano, 1989), and this bias has been referred to as a landmark attraction effect (Bryant & Subbiah, 1994). These biases and distortions of memory result in a displacement of the remembered spatial position of a target; in other words, the remembered position of a previously perceived target stimulus is not the same as the actual position previously occupied by that target stimulus — the remembered position is displaced from the actual position. Studies of representational momentum have suggested that memory for spatial location may be influenced by memory averaging (e.g., Freyd & Johnson, 1987). Hubbard (1995b) distinguished between two senses of memory averaging: A temporal sense in which memory for the final position of a target is influenced by the memory of prior positions of the target, and a spatial sense in which memory for the final position of a target is influenced by the memory of nontarget stimuli that were presented concurrently with the target. It may be possible that remembered prior positions of a target could serve as landmarks, and via temporal memory averaging account for displacements of remembered position toward the average of those prior positions. However, it is more obvious and perhaps more commonplace that nontarget stimuli or context presented concurrently with the target may serve as landmarks. The presence of such concurrent nontarget stimuli or context could provide the opportunity for spatial memory averaging. Indeed, the landmark attraction effect as it is currently conceived may be a special case of a more general spatial memory averaging: Both landmark attraction effects and spatial memory averaging predict that memory for the position of a target should be displaced in the direction of a landmark. Spatial memory averaging and landmark attraction effects are both consistent with a number of findings in the representational momentum literature. For example, if a rotating target is surrounded by a larger stationary square frame, forward displacement of the target is increased if the orientation of the frame is rotated slightly forward from the final orientation of the target, and forward displacement of the target is decreased if the orientation of the frame is rotated slightly backward from the final orientation of the target. Similarly, if the surrounding frame is in motion concurrently with the target, forward displacement of the target is increased if the frame is rotating in the same direction as the target, and forward displacement of the target is decreased if the frame is rotating in the direction opposite to the target (Hubbard, 1993). These displacement patterns may result from a combination of representational momentum and landmark attraction effects: When representational momentum and landmark attraction operate in the same direction Canadian Journal of Experimental Psychology, 1999, 53:3, 242-255 Landmark Attraction Effect 243 (i.e., when the frame is in front of the target or moving in the same direction as the target), they combine and the resultant displacement forward along the axis of motion is increased. When representational momentum and landmark attraction operate in opposite directions (i.e., when the frame is behind the target or moving in the direction opposite to the target), they partially cancel and the resultant displacement forward along the axis of motion is decreased. Spatial memory averaging and landmark attraction effects are also found along the axis orthogonal to target motion. In addition to the displacement along the axis of motion, memory for the position of a horizontally moving target is also displaced toward a single larger surface presented above or below the path of target motion. Furthermore, the magnitude of upward displacement when a larger stationary object is above the target is smaller than the magnitude of downward displacement when a larger stationary object is below the target (Hubbard, 1995a). Such an asymmetry may be accounted for if a landmark attraction effect displaces memory toward the larger stationary object, and the landmark attraction effect also combines with representational gravity (i.e., with a displacement in the direction of implied gravitational attraction; see Hubbard, 1995b, 1997). When landmark attraction and representational gravity operate in the same direction (i.e., when the landmark surface is below the target), they combine and the resultant displacement along the orthogonal axis is increased, whereas when landmark attraction and representational gravity operate in opposite directions (i.e., when the landmark surface is above the target), they partially cancel and the resultant displacement along the orthogonal axis is decreased. Also, memory for the position of a vertically moving target is displaced toward a single larger surface presented on the left or right of the path of target motion (Hubbard, 1998), with the magnitudes of leftward or rightward displacement being very similar. Given that landmark attraction effects influence displacement along the axis orthogonal to motion for horizontally or vertically moving targets, it is possible that landmark attraction effects might also influence displacement along the axis of motion for such targets. The frame studies previously discussed demonstrated effects of an enclosing frame on memory for the orientation of the target (and on displacement in the direction of motion), but a landmark is more traditionally considered to be offset in a single specific direction from a target rather than enclosing a target. On the basis of the previous studies, we may hypothesize that a landmark that is offset in a specific direction from a target should influence displacement for that target in the direction of the landmark, and we may adapt the vector notion of Hubbard (1995a, 1995b, 1998) to make explicit predictions. If the target moves directly toward a landmark, then both representational momentum and landmark attraction would bias memory in the direction of the landmark, and so they RM LA M Displ. Motion Toward Landmark Motion Away From Landmark Figure 1. The contributions of representational momentum (RM) and landmark attraction effects (LA) to the displacement of the target along the axis of motion when the target moves directly toward or away from the landmark. would combine to produce a larger displacement forward along the axis of motion (see Figure 1, left). If the target moves away from the landmark, then representational momentum would bias memory away from the landmark, and landmark attraction would bias memory toward the landmark, and so they would partially cancel and produce a smaller displacement forward along the axis of motion (see Figure 1, right). Effects of a landmark on the displacement of a target along the axis of motion need not be limited to cases in which the target moves directly toward or away from the landmark, but might also be found if a target passes close by a landmark. Previous investigators reported that memory for a stationary target was displaced toward a cardinal axis (i.e., displaced toward a horizontal or vertical axis of symmetry; Huttenlocher, Hedges, & Duncan, 1991) or toward a diagonal line (Schiano & Tversky, 1992) of a larger enclosing figure or frame. If a cardinal axis of the landmark may also function as a landmark-like reference, then we may hypothesize an analogous combination of representational momentum and landmark attraction effects: If both landmark attraction effects and representational momentum operate in the same direction (i.e., a target moving toward the cardinal axis of the landmark), they should combine and the resultant displacement of the target forward along the axis of motion should be increased, whereas if landmark attraction and representational momentum operate in opposite directions (i.e., a target moving away from the cardinal axis of the landmark), they should partially cancel and the resultant displacement of the target forward along the axis of motion should be decreased (see Figure 2). Additionally, we should also observe effects of a landmark on the displacement of the target along the axis orthogonal to target motion: When a target passes close by a landmark, the displacement of that target along the axis orthogonal to target motion should be toward the landmark. The following experiments tested these predictions regarding the effects of a landmark on the displacement of a moving target. In Experiment 1, the path of target motion was either directly toward or away from a larger stationary 244 Hubbard and Ruppel RM LA M Displ. RM LA M Displ. Motion Toward Cardinal Axis Motion Away From Cardinal Axis Vertical Cardinal Axis of Landmark Figure 2. The contributions of representational momentum (RM) and landmark attraction effects (LA) to the displacement of the target along the axis of motion when the target passes near a landmark (and toward or away from a cardinal axis of the landmark). landmark. In Experiment 2, the same targets used in Experiment 1 were displayed, but the landmark was not visible. In Experiment 3, the target passed close to a larger stationary landmark, and the path of target motion paralleled the nearest edge of that landmark. In Experiment 4, the same targets used in Experiment 3 were displayed, but the landmark was not visible. Experiments 1 and 3 examined possible effects of landmarks on displacement along the axis of motion (and in Experiment 3, on displacement along the axis orthogonal to target motion), and Experiments 2 and 4 provided control conditions that ruled out a bias toward the centre or cardinal axis of the screen as an explanation for the displacement patterns observed in Experiments 1 and 3. Experiment 1 In this experiment, observers viewed a small moving target and a larger stationary landmark. The target appeared above, below, to the right, or to the left of the landmark; immediately after the target appeared, it moved directly toward or away from the landmark. The size of the target varied across trials, but was constant within a trial, whereas the size of the landmark was constant across trials. The target vanished without warning, and after the target vanished, observers indicated the location at which the target vanished. If representational momentum and landmark attraction combine in influencing displacement along the axis of motion, then forward displacement of targets that moved toward the landmark should be larger than forward displacement of targets that moved away from the landmark. Alternatively, if representational momentum and landmark attraction do not combine in influencing displacement along the axis of motion, then forward displacement should not be influenced by whether targets moved toward or away from the landmark.