Numerous studies have found that approximately 40% of the adult population behave as if they do not know that water remains horizontal, regardless of the orientation of its container. In the present set of experiments, it was shown that subjects who perform inaccurately do so because they have misrepresented the problem. These individuals are not answering the environment-relative question, “What should the water's surface orientation be relative to the environment?” Instead, they attempt to solve an object-relative problem, “How should water incline relative to its tilted container?” It was shown that subjects' problem representations can be influenced by manipulating the manner in which the problem is presented, and that performance is affected accordingly. Experiment 1 was a pretest, designed primarily to identify individuals who perform correctly or incorrectly on the traditional paper-and-pencil task. Experiment 1 also demonstrated that by promoting an object-relative perspective, a majority of subjects can be induced into performing incorrectly on the water level problem. Experiment 2 found that when individuals were led to evaluate the problem from an environment-relative perspective, they performed accurately, regardless of their performance in Experiment 1. Experiment 3 found that subjects who performed incorrectly on the paper-and-pencil problem are not simply more variable than those who are accurate, but rather that the object-relative perspective creates a bias toward drawing liquid surface orientations that incline in the same direction as the tilt of the container. Experiment 4 found that the response bias observed in Experiment 3 is the result of a frame of reference perceptual influence. The orientation of the surrounding container affects the apparent tilt of the liquid surface. In summary, it was shown that people fail on the water level problem, not because they are lacking the relevant knowledge, but rather because they are attempting to solve a different problem—a problem represented in an objectrelative, as opposed to an environment-relative coordinate system. Moreover, object-relative solutions manifest a perceptual bias that is inherent in all people.

The visually guided control of helicopter flight is a human achievement, and, thus, understanding this skill is, in part, a psychological problem. The abilities of skilled pilots are impressive, and yet it is of concern that pilots' performance is less than ideal: they suffer from workload constraints, make occasional errors, and are subject to such debilities as simulator sickness. Remedying such deficiencies is both an engineering and a psychological problem. When studying the psychological aspects of this problem, it is desirable to simplify the problem as much as possible, and thereby, sidestep as many intractable psychological issues as possible. Simply stated, we do not want to have to resolve such polemics as the mind-body problem in order to contribute to the design of more effective helicopter systems. On the other hand, the study of human behavior is a psychological endeavor and certain problems cannot be evaded. Four related issues that are of psychological significance in understanding the visually guided control of helicopter flight are discussed. First, a selected discussion of the nature of descriptive levels in analyzing human perception and performance is presented. It is argued that the appropriate level of description for perception is kinematical, and for performance, it is procedural. Second, it is argued that investigations into pilot performance cannot ignore the nature of pilots' phenomenal experience. The conscious control of actions is not based upon environmental states of affairs, nor upon the optical information that specifies them. Actions are coupled to perceptions. Third, the acquisition of skilled actions in the context of inherent misperceptions is discussed. Such skills may be error prone in some situations, but not in others. Finally, I discuss the contextual relativity of human errors. Each of these four issues relates to a common theme: the control of action is mediated by phenomenal experience, the veracity of which is context specific.

People's ability to extract relevant information while viewing ongoing events is discussed in terms of human capabilities, limitations, and defaults. A taxonomy of event complexity is developed which predicts which dynamical events people can and cannot construe. This taxonomy is related to the distinction drawn in classical mechanics between particle and extended body motions. People's commonsense understandings of simple mechanical systems are impacted little by formal training, but rather reflect heuristical simplifications that focus on a single dimension of perceived dynamical relevance.

The study of visual perception in human infants is confronted by a number of special problems arising from the inaccessibility of verbal reports. In this paper, we discuss the experimental strategy of converging operations in the context of investigating the phenomenal experience of infants. The goal ofthis strategy is to logically and empirically delimit alternative explanations for a given behavior. Knowledge about the mature functioning of a perceptual competence, as well as knowledge about its developmental course, constrains the selection of viable explanations, but cannot produce a unique interpretation. Thisgoal is pursued through theimplementation of an iterative strategy in which competing interpretations are tested until only one plausible alternative remains. A series ofexperiments investigating infants’ sensitivity to biomechanical motions are reviewed as a way of illustrating how this methodology is operationalized.

In five experiments, assessments were made of people's understandings about the dynamics of wheels. It was found that undergraduates make highly erroneous dynamical judgments about the motions of this commonplace event, both in explicit problem-solving contexts and when viewing ongoing events. These problems were also presented to bicycle racers and high-school physics teachers; both groups were found to exhibit misunderstandings similar to those of naive undergraduates. Findings were related to our account of dynamical event complexity. The essence of this account is that people encounter difficulties when evaluating the dynamics of any mechanical system that has more than one dynamically relevant object parameter. A rotating wheel is multidimensional in this respect: in addition to the motion of its center of mass, its mass distribution is also of dynamical relevance. People do not spontaneously form the essential multidimensional quantities required to adequately evaluate wheel dynamics.

Good designs take account of perceptual tendencies and conceptual biases in observers. Report presents overview of evolving knowledge of interactions between video displays and human observers. Discusses relative advantages and disadvantages of static and dynamic displays, with attention to human factors combining with characteristics of video-display medium to affect observer's percepts.

Artificiality of images apparent to subjects and influences experimental results. Report evalutes computer-generated animation in research on perception of motion. Such research programs not pursued without computer animation, report notes. Computer-generated displays afford variability and control almost impossible to achieve otherwise. Medium limited in that computer-generated images present simplified approximations of dynamics of natural events.