Elsevier

Cognition

Volume 87, Issue 2, March 2003, Pages B59-B67
Cognition

Brief article
Changing perspective within and across environments

https://doi.org/10.1016/s0010-0277(02)00231-7Get rights and content

Abstract

Perspective change within a single environment is a slow and effortful process. However, little research has addressed perspective change across multiple environments. Using a task-set switching paradigm, subjects judged spatial relationships between target locations from differing perspectives. Response times were longer when successive trials probed different perspectives. However, this cost was greater when perspective was changed within a single environment compared to when it was changed across two environments. This result indicates that the processing of perspective change, and perhaps general spatial reasoning, differs in these two cases. Implications for theories of perspective change and environmental knowledge are discussed.

Introduction

Imagine you are in the center of your kitchen facing your stove. Where is your sink in relation to you? Turn to face your refrigerator. Now where is your sink? To answer the first question, information about your kitchen had to be accessed from memory. This information is referred to as an “environmental representation”. Moreover, you had to adopt a certain perspective within that environment, according to a particular heading. To answer the second question, you had to change your perspective by changing your heading within the environment. The adoption of a new perspective within an environment requires one to engage in “spatial reasoning” whereby the relationship of the objects relative to the viewer has to be realigned.

A variety of experiments have demonstrated that spatial reasoning requires time. For example, Rieser (1989) showed that response time (RT) and errors in estimated viewer-to-object direction judgments increase as a viewer changes perspective by imagined turning to face new objects (see also Farrell & Robertson, 1998). Imagining an array of targets rotate around a stationary observer can be even more difficult and time consuming (e.g. Huttenlocher and Presson, 1979, Wraga et al., 1999, Wraga et al., 2000; see also Presson, 1982). Similar costs have also been observed with imagined change in self location (Easton & Scholl, 1995).

Research on perspective change has primarily dealt with target locations within a single, well-defined environment, as in the kitchen example above. However, research on environmental representations has suggested that humans do not maintain a single representation of all known environmental space. Rather, humans maintain a series of multiple, separable representations. The dominant view is that these representations are hierarchically organized, such that we may, for example, maintain separate interconnected representations of a city block, the buildings on that block, and the rooms in a particular building (e.g. Hirtle and Jonides, 1985, McNamara, 1986, McNamara et al., 1989, Stevens and Coupe, 1978, Taylor and Tversky, 1992).

A system of multiple representations provides cognitive economy – we can consider only the relevant local aspects of the world required by a particular situation. A disadvantage exists, however, in that we must often consider the relationship between locations that may be stored in separate representations, thereby requiring the access of several representations. This process is non-trivial. Brockmole and Wang (2002) recently showed that only one representation can be accessed at a time, and that changing the mentally active representation entails a temporal cost.

Because environmental knowledge is distributed across separable representations in memory, it may be fruitful to consider perspective change not only within a single environment, but also across environments. The manner in which a change in perspective influences spatial task performance may vary depending on whether that change occurs within (where one representation is considered) or across (where multiple representations are considered) environmental representations. Any differences in the way in which perspective change influences performance in these two cases could shed light on the structure of the environmental memory system and the cognitive processes underlying spatial reasoning.

The present study compared the perspective change process within a single environment and across two nested environments. On any given trial, subjects judged whether a real-world target was presented correctly relative to an imagined location and perspective. For example, on one trial a subject may have imagined being in his office (i.e. accessed an environmental representation) facing the window (i.e. adopting a particular perspective) and been asked if his desk was to his right. Thus, two factors could be manipulated: the environmental representation and the adopted perspective. In the present study, trials either required a change in perspective only, or a change of both environment and perspective.

The logic of the experimental design follows that of a cued task-set switching paradigm (e.g. Allport et al., 1994, Jersild, 1927, Rogers and Monsell, 1995, Spector and Biederman, 1976). In this paradigm, subjects are presented with a continuous set of trials. On a particular trial subjects are engaged in some task. On the subsequent trial, subjects may execute the identical task (no-switch trial) or engage in a different task (switch trial). If changing task requires additional time, then RT for the switch trials will be longer than that for no-switch trials. This conclusion is strengthened if a pre-cue indicating the identity of the nature of the upcoming trial reduces this switch cost, since it should allow subjects time to become prepared for the next task before the start of the trial. If no additional time is needed to respond on the switch trials, then it is inferred that no additional processing was required during the switch in task.

In the context of the present study, in an unpredictable order, subjects evaluated the correctness of target placement from two different perspectives. Thus, on some trials, perspective was held constant between successive trials (no-switch trials) and on other trials perspective was changed (switch trials). A cue indicated what perspective to adopt for that trial. The elapsed time between the cue and target presentations, the cue-to-target interval (CTI), was 0 or 2000 ms. If changing perspective requires time to complete, then RT to switch trials should be longer than for no-switch trials. Furthermore, the magnitude of this cost should be smaller at the longer CTI as it provides subjects a head-start to change their perspective prior to the presentation of the target.

In separate conditions, perspective change occurred either within a single environment (within condition), or across two nested environments (across condition). In the within condition, subjects changed perspective between facing north and facing east in the middle of the psychology building. In the across condition, they changed perspective between facing west in the middle of the psychology building and facing north in the middle of their office.

Note that compared to the within condition, which only requires a change in perspective, the across condition requires two processes that potentially take time. One is perspective change (e.g. Rieser, 1989). The other is an environmental representation switch (Brockmole & Wang, 2002).1 Thus, there are three potential ways the costs associated with changing perspective could differ when perspective is changed across environments compared to when it is changed within an environment. First, switching environment and changing perspective may occur sequentially; thus, the time required to complete both processes is additive. This hypothesis predicts that the cost in the across condition should be longer than in the within condition. Second, switching environment and changing perspective could be accomplished in parallel. In this case, the time required to change both perspective and environment would simply be the duration of the longer of the two processes. Based on past research, it is expected that the perspective change will require more time than switching environment, and thus it would be predicted that the switch cost in the across condition should equal that in the within condition. Finally, the switch cost in the across condition could be smaller than that in the within condition. This most interesting of possibilities would suggest that the processing that underlies perspective change across environments is different than perspective change within a single environment. The experiment reported here was designed to examine these possibilities. Therefore, the critical comparison concerns the switch cost in the within condition and the across condition.

Section snippets

Subjects

Eight professors at the University of Illinois participated after providing informed consent. Subjects were compensated with gift certificates to a local coffee shop.

Stimuli

Stimuli consisted of a circle with eight equally spaced dots around its perimeter presented on a computer screen. The perimeter of the circle represented the 360 degrees of space surrounding the subject in an imagined location (i.e. the top dot represented the area of space directly in front of the subject in the imagined

Results

Trials were excluded if the response was incorrect or if RT was ±3 standard deviations from the mean RT for correct trials, on a subject-by-subject and cell-by-cell basis. In the within condition, 4% of responses were incorrect; in the across condition, 3% of responses were incorrect. For both conditions, the RT trim discarded 2% of the remaining data. For all analyses, an alpha level of less than 0.05 was adopted as the criterion for statistical reliability.

Mean RT is summarized in Table 1,

Discussion

Changing perspective across environments required less time than changing perspective within a single environment. This is surprising given that past research has shown that both changing perspective within a single environment (Rieser, 1989) and switching between different environmental representations in memory (Brockmole & Wang, 2002) require time. The non-additivity of these processes, however, indicates that switching environment and changing perspective do not occur sequentially.

Acknowledgements

This research was supported by a NSF Graduate Research Fellowship to J.R. Brockmole and a University of Illinois Institutional Research Board Grant to R.F. Wang. We are deeply indebted to all of our subjects, especially the faculty in the Psychology Department at the University of Illinois who so willingly reversed their typical role in research to participate in this study. We also thank David Kaneshiro for his help with data collection.

References (20)

There are more references available in the full text version of this article.

Cited by (0)

View full text