Main content

Alert message

Emerson Foulke
(presented at International Symposium on Orientation and Mobility, 
Trondheim, Norway, 1996)

Before I begin my discussion, I would like to explain what I mean by the term "mobility." We customarily talk about the two terms "orientation" and "mobility" as if they were separate and independent activities. If this dichotomy is taken seriously, mobility becomes the ability to move through space safely, without the requirement of knowing where you are, where you want to go, or how to get there. Being able to move through space with the assurance that your next step will not be your last step is undoubtedly comforting, but if this is all we have in mind when we consider mobility, then mobility is not a very interesting activity. We would do better to say that orientation is an essential component of meaningful mobility. I certainly do not mean to suggest that the ability to move safely through space is not an important component of mobility, but when we add orientation as an essential component, mobility becomes a useful ability. Accordingly, when I use the term "mobility," I mean to indicate an activity that is impossible without orientation.

One way to get at the information we must have in order to understand the task in which blind pedestrians engage when they travel independently, safely, and purposefully, is to pose a few questions whose answers, when we have them, will tell us what we need to know. Here are a few questions which, at my current level of understanding, seem to me to be important.

What constitutes an adequate description of the mobility task? What are the capabilities of the perceptual systems that are available for use by the pedestrian who is engaged in the mobility task? How much and what kinds of information about space that is relevant to the mobility task can be acquired by the pedestrian's functioning perceptual systems? What contribution to performance of the mobility task is made by the perceived spatial information that is put to use while the task is in progress? What contribution to the mobility task is made by the perceived spatial information that is stored in memory for later use? When we can answer these questions, we will have come a long way in our effort to understand the mobility of pedestrians, blind or sighted.

In my discussion, I will talk first about the relative capabilities of perceptual systems. Next, I will attempt a description of the mobility task and the space in which it is performed. Then I will discuss differences between blind and sighted pedestrians in regard to the information they acquire by perception, and the uses they make of that information.

The Capabilities of Perceptual Systems

To travel independently, pedestrians must have a continual supply of spatial information. The perceptual systems of humans are obviously not equivalent in regard to their ability to inform their users about space, and the skill with which the mobility task can be performed depends, in large part, on the nature of the spatial information provided by the functioning perceptual systems. Accordingly, in order to know how well and in what way pedestrians can perform the mobility task, it is necessary to compare the human perceptual systems in regard to the acquisition of spatial information. This comparison can be made by posing a number of questions whose answers will define the functioning of the perceptual system to which the questions are addressed.

Reach. What is the reach of a perceptual system? This question can be answered by determining how much of the surrounding space can be observed from one position and on one occasion by that perceptual system. The answer to this question determines the extent of both the immediate space within which safe movement is assured, and the remote space within which landmarks can be observed. It also determines how much perceptual anticipation is possible, and how much integration of the serial perception of spatial facts acquired on different occasions is required for construction of the memorial representation of space.

Focus. How selective is a perceptual system? What ability does it have to exclude some of the surrounding space from observation, and focus on the sector of space where needed information is likely to be found? The answer to this question has a bearing on the vulnerability of a perceptual system to the interference caused by noise, etc. For example, by focusing, the visual system can exclude much of the surrounding space and the interfering stimulation it might contribute from observation. On the other hand, the auditory system is stimulated by acoustic energy from all directions, and has little ability to exclude interfering stimulation.

Analysis. What kind of analysis can be performed by the perceptual system in question? By analyzing the stimulus energy to which it is sensitive, the visual system acquires information about the positions, shapes, and surface characteristics of objects in space. By analyzing the stimulus energy to which it is sensitive, the auditory system acquires information about the temporal organization and extension of events. Consider, for instance, the analysis that discloses the phonemes in a speech sound. Neither the visual system nor the auditory system is, by design, capable of the analysis carimpaired out by the other system. Like the visual system, the haptic system is, by design, suited for the observation of objects in space, but its field of observation is relatively small, and its acuity is relatively poor.

Pattern resolution. What is the ability of a perceptual system to resolve patterns? Where is the threshold below which the details of a pattern are too small to be discriminated? The better a perceptual system is at pattern resolution, the more information it can acquire from a stimulus pattern. The perceptual systems of humans differ widely in this regard. For instance, as already mentioned, although the visual system and the haptic system can determine the characteristics of objects, details that are obvious to the visual system are often below the threshold of the haptic system.

Perceptual anticipation. I will consider this concept in greater detail at a later point in my discussion, but I mention it here because the perceptual systems of humans vary widely with respect to the perceptual anticipation they afford. How much perceptual anticipation is possible? This question is answered in part by knowing the reach of a perceptual system, because the extent of the observable space determines the time that can be made available to get ready for upcoming situations. However, in answering this question, it is also necessary to consider the answers to the questions concerning focus, analysis, and pattern resolution, because the value of perceptual anticipation depends not only on the time it makes available but also on the quantity, relevance, specificity, and accuracy of the information provided in advance of the need for it.

Perceptual error. Answers to the questions just discussed should make clear the extent to which a perceptual system can inform or misinform pedestrians who are depending on that system for the information they need in order to perform the mobility task. For example, the conceptions of space acquired by blind pedestrians are often distorted because the proprioceptive system has less ability than the visual system to detect curvature, and because the auditory system and the proprioceptive system have less reach and less ability than the visual system to estimate the directions and distances of spatial features (Howard & Templeton, 1966).

My discussion of the relative contributions of which our perceptual systems are capable has been brief and superficial. However, I hope that it has been sufficient to suggest that the accuracy with which the mobility task is performed, and the way in which it is performed depends on the spatial information that can be provided by the pedestrian's functioning perceptual systems.

The Contents of Space

Spatial features. The things in space include more than just the trees, buildings, lamp posts, and other articles of spatial furniture we ordinarily think of as objects. I have in mind any part of space that is discriminable enough to be distinguished from other parts of space. Consequently, I am also including streets, sidewalks, curves in streets and sidewalks, slopes, surface textures, and so forth, abstract patterns formed by the arrangements of things in space, and events.

Events. In order to perform the mobility task, pedestrians must know not only about stationary spatial features, but also about spatial features that are in motion. Visual observers are usually able to identify moving objects and to estimate their distance, direction, and rate of movement. If moving objects produce sound as they move, auditory observers can also estimate their distance, direction, and rate of movement, and they are often able to identify a moving object by learning the association between it and the sound it makes. Both blind and sighted pedestrians need information about moving objects in order to predict the future positions of those objects, and to predict the times at which the extrapolated positions will be occupied, because decisions about whether to move, and if to move, in what direction and how fast, must often be based, in part, on predictions.

The information acquired by visual observation concerning distance, direction, and rate of movement is more detailed and accurate than the information acquired by auditory observation, and visual object identification is more precise and dependable than auditory object identification (Foulke & Berlá, 1978). To cite a familiar example, sighted pedestrians routinely cross streets on which cars are moving, at locations where crossing is not controlled by traffic lights. The estimates of direction and rate of movement of traffic approaching on the street about to be crossed, and that are made possible by visual observation, allow them to determine that they will have time to cross before approaching cars can reach them. In similar situations, blind pedestrians usually wait until they do not hear any traffic sounds at all on the street about to be crossed, before crossing (Chew, 1986; Guth, 1986). If they try to base crossing decisions on auditory observation of the distance, direction, and rate of movement of approaching traffic, they often find that by the time cars are close enough to permit reliable observation, there is no longer enough time to cross safely. Furthermore, if there are other sounds in the auditory field of observation that are irrelevant to the task at hand, such as sounds made by other cars in the vicinity, or the sound of the pneumatic drill at the nearby construction site, these interfering sounds may prevent interpretation of the sounds blind pedestrians must interpret in order to get the information they need.

Movement patterns. Pedestrians also acquire some of their knowledge of spatial structure by observing patterns of movement. For example, by observing the patterns expressed by moving cars, sighted pedestrians can determine the boundaries that separate safe zones from dangerous zones. By observing the flow of traffic, they can determine the zone between lanes of traffic where it is relatively safe to stand for a time during the crossing of a busy street. Movement patterns can also be observed auditorily, but they lack much of the definition that makes them useful when they are observed visually. For example, with training, blind pedestrians can, by listening to the sounds made by a line of moving cars, maintain a course that is fairly parallel to the path followed by the cars. When they employ this skill, they are usually walking on sidewalks some distance from the streets along which cars are moving, and they navigate in this way to avoid the numerous obstacles they would encounter if they timpaired to follow the curbs that separate sidewalks from streets. In this case, if they pursue a course that is not parallel to the line of moving traffic and they veer toward the street, they can usually rely on the presence of a curb to rescue them from the consequences of their imperfect knowledge of the boundaries between the safe zone and the dangerous zone. They would be ill-advised to employ the same skill to locate safe zones between lanes of traffic.

Immediate space. Immediate space encompasses all of the spatial features that can be observed from one position and on one occasion, and that are not separated from the observer by intervening unobservable space. The bump in the sidewalk that is just now under foot, the tree immediately to the right of the sidewalk that is discovered by the sound it reflects, the tricycle left in the middle of the sidewalk that is discovered by use of the cane, the step down into the street to the left of the sidewalk that is also discovered by the cane, can be in the immediate space of a blind pedestrian. The blind pedestrian who knows about these spatial features and the relationship among them also knows where it is safe to move. All of these spatial features can be in the immediate space of the sighted pedestrian. However, the sighted pedestrian's immediate space can also include the visually observed house set back some distance from the sidewalk, the visually observed trees in the yard between the house and the sidewalk, the surface of the sidewalk that is visually observable for a considerable distance ahead, and the visually observed activity on the street to the left of the sidewalk. Thus, the immediate space within which the sighted pedestrian can move with the assurance of safety is of much greater extent than the immediate space within which the blind pedestrian can move with the same assurance.

Remote space. Remote space is too extensive to be observed in detail from one position and on one occasion. For both blind and sighted pedestrians, it is a conception that is achieved by integrating information acquired on different occasions. As pedestrians move through space, many features in the space beyond immediate space enter immediate space, become observable, and remain observable until further movement again places them beyond the reach of the available perceptual systems. However, other spatial features in the space beyond immediate space, though separated from the observer by unobservable space, are observable, and remain observable as the pedestrian moves. These features are in the pedestrian's remote space. The office tower in the town center two miles to the north lies within the remote space of sighted pedestrians because, although much of the intervening space can not be observed, the tower can be. Of course, because it can not be observed by blind pedestrians, it is not in their remote space. Because sound is reflected from the surface of the house referred to in the previous paragraph, it can, with learning, become a feature in the blind pedestrian's remote space. However, because the perceptual systems available to the blind pedestrian do not have the reach of the visual system, and because the ability to identify objects is significantly reduced unless they are within hand's reach, the remote space of the blind pedestrian is both much smaller and much less populated with spatial features than the remote space of the sighted pedestrian.

The Mobility Task

I will now turn to a consideration of the task that is performed by pedestrians, and the space in which it is performed. In order to get where they want to go, pedestrians must navigate. They must know, or be able to discover where they are and where they will be when they have reached their spatial goals, and they must be able to determine routes that will get them there. For their movement to be purposeful, they must be able to maintain orientation. To maintain orientation, they must know what things are in space and where those things are (Armstrong, 1977; Foulke, 1982), and they must be able to predict the courses and future positions of things that are moving. Most pedestrians live in constructed environments. To get where they want to go, they move through spaces that humans have furnished with familiar contents, and on which humans have imposed predictable organization. In the constructed spaces through which pedestrians usually move, there are, to borrow a term from J. J. Gibson (1966, p. 285), affordances. That is to say, movement along some courses is easy, whereas movement along other courses is difficult or impossible. These affordances are simply paths. The paths in a space usually intersect. Intersections are connected by path segments. The entire collection of path segments and intersections in a space constitute its path structure, and a space's path structure is a pattern that can be learned and remembered.

When pedestrians move along path segments, they arrive at intersections where they must make decisions. To team what decisions to make in order to select a sequence of path segments that will connect a journey's starting point and its end point is to learn a route. The initial decision in such a sequence is a decision concerning the direction in which to start, and the final decision is a decision about when to stop. It is possible for a route to include only one path segment, in which case only the initial and final decisions must be made. However, most routes include intersections, and when pedestrians follow such routes, they must also make decisions about which way to proceed at intersections. A route affords movement from a starting point to an end point, and more than one route is usually possible.

The space in which pedestrians travel is much more extensive than the space they can observe on any one occasion, but they can remember what they cannot observe. To site a familiar example, pedestrians who have learned the path structure of a space have the information they need to determine more than one route from a starting point to a goal. If they find one route blocked, they have alternatives. It is important to distinguish between observed space and remembered space, because they contribute in different ways to the performance of the mobility task, and their relative contributions vary with visual status.

The Uses of Spatial Information

So far, I have considered the spatial information pedestrians must acquire in order to perform the task, and the perceptual equipment at their disposal for acquiring that information. My remaining task is to consider how, once acquired, spatial information is used to organize and regulate the behavior by means of which the mobility task is accomplished, and I would like to begin by distinguishing between two kinds of information--perceptual information and cognitive information.

Perceptual information. Perceptual information is contemporaneous information. It is information that is acquired directly from the space in which the task is performed, and used while the task is in progress. Pedestrians put such information to several uses. If they can gather enough perceptual information, if it is sufficiently accurate and specific, and if it can be acquired soon enough to allow for the organization and timing of the behavior it dictates, it can control the ongoing mobility task by assuring orientation and safe movement in space, without much help from memory. The perceptual information that can be acquired by visual observation meets these requirements.

Perceptual anticipation. Numerous experiments (Barth & Foulke, 1979; Crossman, 1960; Hershman & Hillix, 1965; Levin & Kaplan, 1969; McLean & Hoffman, 1973; Poulton, 1954) have demonstrated the dependence of skillful performance on the ability to anticipate behavioral requirements by observing the features of a situation in advance of the time at which some action will be required. The sighted pedestrian who receives advance notice that a curb is 10 or 15 feet away because it is in plain sight, has time in which to prepare for a sequence of movements which, when executed with proper timing, will ensure that the curb is negotiated skillfully. However, because the space that can be perceived without sight is so small, the blind pedestrian has little time to prepare for perceived spatial features. The curb discovered by the tip of the cane is only a step away. Thus, for the most part, the blind pedestrian lacks the perceptual anticipation that serves the sighted pedestrian. Fortunately, the blind pedestrian can replace the missing perceptual anticipation with what I have called cognitive anticipation. However, before considering cognitive anticipation, it will first be necessary to examine the memory on which cognitive anticipation depends.

Constructing the memorial representation. Informing the behavior of the moment is not the only use to which perceptual information can be put. It can also be remembered. Even when, as is often the case for blind pedestrians, perceptual information is insufficient to specify the behavior required for successful performance of the mobility task, it can be added to information already in memory, and as experience with a space accumulates, a memorial representation of that space can be gradually formed.

Cognitive information. Unlike perceptual information, cognitive information is not obtained directly from the space in which the task is performed. It is, instead, obtained from memory. The information supplied by memory is not limited to representations of perceptual information. It also includes information established by inference (e.g., relationships among remembered spatial features that were not observed on the same occasion), generalizations made possible by the reiterations that characterize constructed spaces, and communicated information; that is, information received by way of spoken or written language.

Piecemeal acquisition. For bad pedestrians, the memorial representation is constructed with piecemeal information acquired on many different occasions. It is probably less accurate than the memorial representation of the sighted pedestrian that is constructed with information gathered in larger chunks on fewer occasions. For instance, because visual observers can observe a relatively large space on one occasion, they can directly observe not only its features, but also the spatial relationship of those features. Blind observers must bring together spatial information gathered on different occasions and establish the spatial relationship of many features by inference. Of course, representations based on piecemeal information are evaluated and corrected by the feedback resulting from the behavior based on them. Nevertheless, it is likely that they are not as accurate and not as reliable as the memorial representations achieved by sighted pedestrians.

Quantitative differences. The memorial representations of blind pedestrians are also different from the memorial representations of sighted pedestrians in respect to the quantity of information they contain. Blind pedestrians can observe very little of the spaces bounded by the paths along which they walk. They learn very little by observation about the buildings, trees, fences, and so forth, commonly found in such spaces, and things not observed and remembered cannot serve as landmarks.

Landmarks

Spatial features may become landmarks. Landmarks allow pedestrians to locate themselves in the space through which they are moving, and by doing so to find out where they are. For spatial features to serve as landmarks, they must be stable, and pedestrians must remember not only those spatial features, but also their relationships to other stable spatial features. Beyond this, pedestrians must maintain the currency of their knowledge of the changing relationships between stable spatial features and themselves as they move through space. Landmarks can be anywhere within the reach of the perceptual system by means of which they are observed. They may be located in the spaces bounded by path segments, or on the path segments, themselves. The large house with the distinctive roof line, the old hollow tree, the unusually sharp curve in the sidewalk, and the bump in the sidewalk caused by the tree root underneath are all spatial features and can serve as landmarks for pedestrians who have learned to identify them and to relate them to other spatial features. Some of these spatial features, such as the bump in the sidewalk, will become landmarks for blind pedestrians. Other spatial features, such as the house with the distinctive roofline, will become landmarks for sighted pedestrians.

Detecting the spatial feature that serves as a landmark is, of course, a perceptual event. However, knowing that it is a landmark is achieved by consulting memory, and thus, the knowing is a cognitive event.

I.andmarks in remote space. The office tower mentioned earlier is in the remote space of visual observers and can become a landmark. Its distance from the observer can be estimated, and if it can be related to other spatial features, it can provide the information needed for orientation. Landmarks of this sort are especially useful because, unlike landmarks in immediate space, they do not quickly become unobservable as the observer moves. Incidentally, features of remote space can become landmarks and provide orientation for pedestrians who never go to the places at which those landmarks are located. The mountain range to the east of the city that looms in the distance can provide orientation for pedestrians who have never been, and never expect to be anywhere near it.

The spatial features in the remote spaces observed by blind pedestrians are, for the most part, objects that emit sounds. However, in order to be observed, audible features must be closer to the observer than visible features. Furthermore, if a sound is to serve as a landmark by supplying the information that allows auditory observers to estimate the position of its source, and perhaps to infer its cause, it must be emitted by a stationary source, and it must be available when auditory observers need to consult it. These conditions are not frequently satisfied. As a result, visually observed features of remote space are both more numerous and more useful as landmarks than auditorily observed features of remote space.

Landmarks in immediate space. Pedestrians do not need landmarks to maintain orientation in immediate space because immediate space is directly observable. However, when spatial features originally perceived in immediate space have been remembered, and when the relations among them have been learned, blind pedestrians can use landmarks in immediate space for orientation in remote space, and the landmarks on which blind pedestrians depend are, for the most part, in immediate space. As they move and the boundaries of immediate space change, landmarks quickly become unobservable and must be replaced by other landmarks in the current immediate space. Sighted pedestrians get from immediate space the information they need for safe movement but, to the extent they can use landmarks in remote space that do not quickly disappear as they move, they can dispense with landmarks in immediate space. They do not need to identify the bump in the sidewalk by the feel of the surface under foot to know where they are. Because the knowledge blind pedestrians have of remote space is meager and undependable, they must depend more heavily for orientation on landmarks in immediate space (Brambring, 1982). Because immediate space is much less extensive than remote space and because their movement quickly places landmarks in immediate space beyond the reach of the perceptual systems available to them, they must learn and remember many more landmarks than sighted pedestrians.

The blind pedestrian's remote space. Because there is less for them to observe in remote space, the conceptions of remote space achieved by blind pedestrians depend, to a considerable extent, on remembering spatial features in immediate space, and learning the relationships among them. Their immediate space consists, for the most part, of the space on and near the paths along which they walk. Because their memorial representations include very little information about the spaces bounded by path segments, when they incorporate in their memorial representations the features of immediate space, and when they have learned the relationships among these features, the conceptions of remote space that emerge resemble mazes, or networks of spatial corridors.

Spatial stereotypes. When pedestrians are operating in spaces that have, with experience, become familiar, their memorial representations provide them with information about those spaces, but sometimes they find themselves in spaces that are not familiar. Of course, if current space is not represented in their memorial representations, they cannot consult their memorial representations for information about it. However, their memorial representations can often provide information about similar spaces that may be applicable. Most people spend most of their time in constructed environments, and in constructed environments, certain patterns are likely to be repeated over and over again. For example, in cities, streets cross other streets, and they often intersect at right angles. Many of them have been given contours that promote drainage. From the center, they slope downward to either side. They are usually bounded by curbs, so that one must step up in order to pass from a street to the adjacent land. On both sides of the street, one frequently finds sidewalks, and in residential areas, these sidewalks are often separated from the streets by grassy verges on which objects such as trees, utility poles, and lampposts are occasionally to be found. Beyond the sidewalks and parallel to them are rows of buildings. The point to be made by this example is reinforced by the ease with which it could be extended.

Because constructed environments exhibit extensive reiteration, generalization is possible, and the memorial representation includes what may be regarded as spatial stereotypes. These stereotypes furnish the generalizations that allow blind pedestrians to make predictions about what they will encounter in spaces not previously experienced. Of course, behavior informed by such generalizations is more likely to result in behavioral errors than behavior informed by direct observation of the space in which the task is performed or by an accurate memorial representation of that space but, in the absence of better information, generalizations are useful.

Blind pedestrians who enter new spaces can have no memory of them to consult, and the observation they can accomplish on first encounter will not be sufficient to inform their behavior. Often, their only recourse is to rely on the generalizations furnished by their spatial stereotypes, and these generalizations, though not entirely trustworthy, will usually suffice until there has been enough direct experience to enable the construction of adequate memorial representations.

Cognitive anticipation. Like perceptual anticipation, cognitive anticipation provides time in which to organize behavior that will shortly be required (Foulke, 1982; Poulton, 1952). As pedestrians gain experience with immediate space, they can, if there is a reason to do so, remember it, and they are then able to obtain the benefits of cognitive anticipation. Sighted pedestrians do not have much reason to remember immediate space, but because the immediate space of blind pedestrians is so restricted, the perceptual anticipation it affords does not gain for them the time they need to prepare for imminent behavioral demands, and they must turn to their memorial representations and cognitive anticipation. In order for cognitive anticipation to replace perceptual anticipation, blind pedestrians must be able, by consulting their memorial representations, to determine their current positions accurately, and they use landmarks in immediate space for this purpose. Landmarks in immediate space give blind pedestrians fairly accurate information about their current positions, because the immediate space they can observe is so small that their current positions are usually the same, or nearly the same as the positions of the landmarks they are using. They are where their landmarks are. When blind pedestrians have had enough experience with a route to team that the curb is three steps beyond the point at which the surface under foot changes from brick to concrete, detecting that change provides the cognitive anticipation they need to prepare and time the sequence of movements that will be required when the curb is reached.

If adequate perceptual anticipation is possible, there is little need to consult memory in order to gain time in which to prepare responses, and the immediate space that can be observed visually is generally of sufficient extent to afford adequate perceptual anticipation. Sighted pedestrians do not need to detect the change in the surface under foot or the distinctive bump in the sidewalk in order to know that a curb is coming up, because they can see the curb soon enough to provide ample time in which to prepare their responses.

A Thought Experiment

While keeping in mind the concepts just discussed, imagine that you are the subject in a thought experiment with three conditions. In the first condition, you are blindfolded and brought to an unfamiliar space with a number of intersecting paths. You can, by the way the surface feels underfoot, distinguish between the path and the surrounding surface, but there are no irregularities in the path surface, no turns at choice points with angles other than 90 degrees, no curves with enough curvature to permit their detection by proprioception, no distinctive sounds that are reliably associated with stationary objects, no objects close enough to the path to be discovered by hearing reflected sounds. Because you are blindfolded, you cannot inform yourself about the contents of the spaces bounded by path segments and, at the outset, you know nothing about the route you are to learn. Your task is to learn the route by traversing path segments and by guessing which way to turn at choice points. You will be told when you have made wrong choices. This task is rather like the task performed by a white rat who must learn the sequence of correct turns in order to progress from the starting box of a maze to its goal box and food. As your experience with the route accumulates, you will gradually learn the sequence of correct turns and will finally be able to traverse the route without error. However, because your performance depends almost entirely on remembered route knowledge, it will be limited in a number of ways.

In order for you to know where you are at any point along the route, you must have in memory, in some form, a representation of the maze, and you must remember what turns you have made. If you have a lapse of memory, you will be lost. If, for some reason, you stray from the path, you will be lost, because your memorial representation cannot provide any information about the spaces bounded by path segments and no landmarks will be represented in it. If you find your path blocked and you have only the knowledge of the route you have learned to rely on, you will not be able to select an alternate route.

On the other hand, if you had enough experience of the right kind to learn not just a route but the entire path structure, or if your experience with the path segments over which you walked suggested to you that the route you learned expressed a predictable pattern, you would have been able to select an alternate route, and you could have done so without relying on landmarks. However, it is likely that you would need many more learning trials to acquire the ability to select alternate routes than a pedestrian whose memorial representation was furnished with landmarks. Of course, blind pedestrians rarely find themselves in situations in which there are no spatial features that can become landmarks and in which they must depend entirely on route knowledge, although a network of corridors in a large hotel can come fairly close to satisfying this condition. However, the distant landmarks used by sighted pedestrians to maintain orientation in remote space are, for the most part, not available to blind pedestrians and in some situations, landmarks in immediate space are scarce. When this is the case, they must depend more heavily for orientation on remembering their own behavior. They must often remember what they have done in order to know where they are.

In the second experimental condition, you are again blindfolded and your task is the same, but the terrain in which your route lies exhibits the variability one would ordinarily expect to encounter. The surface over which you walk has slopes and discontinuities. There are trees and other objects beside the path that can be detected by hearing reflected sounds. The sounds made by moving traffic identify the positions of streets. Such characteristics can be remembered and when you have had a number of learning trials, many of them will become landmarks. Served by these landmarks, you will no longer have to depend so heavily on your memory of past behavior to know where you are. When the route is partially learned, the bump in the walk may tell you that the required turn to the left is a few feet ahead. When learning is more complete, you will, by recognizing the bump as you walk over it, be able to confirm or, if your attention has been wandering and you have become uncertain about your orientation, establish your position on the route. As was the case in the first experimental condition, if you stray from the route you will be lost, because except for objects close to the route, you will not know anything about the contents of spaces bounded by path segments, and if the route is blocked, you will not be able to choose an alternate route. However, you will be able to use spatial information you acquire while you are traveling to maintain orientation, instead of having to depend wholly on your memory of what you have done.

And now for the final condition of service in this thought experiment. You are still a subject in a wayfinding experiment, but this time you have been prepared for service in the experiment by being allowed to become familiar with the space in which your performance is to be observed. You are then blindfolded, brought, without knowing where, to some location in that space, and told that your task is to reach the goal indicated by the experimenter. At this point, your blindfold is removed. As you look around you, you see a large building which has, because of your previous familiarization with the space in which your task is to be performed, been incorporated in your memorial representation of space and is a landmark. Because it is a landmark, you know where it is in relation to other landmarks, and in the act of seeing it, you have discovered your relationship to it. Consequently, you know where you are. Because of your previous experience in this space, you may have learned a route to the goal, and you may even have learned a path structure. However, it is not necessary for you to learn about routes and path structures in order to know where you are now. If a number of spatial characteristics have become landmarks, you have all of the information you will need to find your way through this space. You will still have to make decisions at intersections, but by supplementing the landmark information supplied by your memory with spatial information acquired while you are performing the task, you will have the information you need to make these decisions.

An Attempt To Validate

I hope this description and analysis of the mobility task and the information on which it depends seems reasonable. However, merely seeming reasonable is by no means sufficient. A way must be found to make an inventory of the facts about space that are established by the pedestrian's functioning perceptual systems, and of the perceived spatial facts that are stored in memory and subsequently used in conjunction with directly perceived spatial facts to inform the pedestrian's behavior.

A few years ago, Dr. John Brabyn, Co-director of the Rehabilitation Engineering and Research Center at the Smith-Kettlewell Institute of Visual Sciences, and I decided to tackle this problem by using a method that is certainly simple and probably naive. We identified a small group of skilled blind pedestrians and asked each pedestrian to mention the facts about space he discovered, to identify the perceptual system responsible for their discovery, and to mention the facts about space that were supplied by memory, as he walked through that space. Each pedestrian carimpaired a wireless microphone, and his comments were recorded on a tape recorder carimpaired by either John or myself. Occasionally, John or I noticed that a pedestrian appeared to be responding to something about the space in which he was operating without mentioning what he had perceived or remembered. When this happened, we felt free to intervene and ask the pedestrian to explain his behavior. After we had accompanied all of the subjects on their walks, I listened to the recording of each subject's comments, and devised a preliminary plan for scoring the performance of subjects. If spatial facts were directly perceived, I wanted to know the perceptual system responsible for the perception. I also wanted to know whether a directly perceived spatial fact was the result of an investigation to confirm a remembered spatial fact. If a spatial fact was remembered, I wanted to know whether it was a specific remembered spatial fact or a spatial fact established by generalization or inference.

This is still a very rudimentary scoring system that will require considerable refinement before it can become useful. However, as it stands, it should give you some idea of the varieties of information about space on which the behavior of a pedestrian is based, and it should suggest that blind pedestrians must use different information than sighted pedestrians, and must therefore perform the task of moving through space safely and purposefully in a different way than sighted pedestrians do.

Each of the protocols we scored is quite lengthy, and I cannot include even a single protocol in this paper without making it much too long. However, I can show you enough of a protocol to give you an idea of what we were trying to do.

The Scoring of Protocols

An assertion is scored P (perception) if it reports a spatial fact that is the immediate result of perception. For instance, if a sighted subject knows that a curb is approximately 10 feet ahead because he sees the curb, the subject's assertion that there is a curb 10 feet ahead receives the score of P.

When P is scored, it is qualified in one or more ways, and its qualifiers are enclosed in parentheses. Qualifiers within parentheses are separated by periods. If the subject reports hearing something, an a (auditory) is added to the score. If the subject reports smelling something, an o (odor) is added. If the subject reports the manual exploration of something, an h (haptic) is added. If information is supplied by the muscle sense, as when a subject notices the particular way a door resists opening when the door handle is pulled, a k (Anesthetic) is added. If the subject reports feeling the sun, wind, rain, etc., or if the skin is touched by some object, a d (dermal) is added.

There are two kinds of proprioception to be distinguished. If the subject reports the feel of the surface under foot, an f (foot) is added. If the subject reports discovering something by touching it with his cane, a c (cane) is added. A perception may also be an observation made to verify an earlier perception, as when a blind pedestrian touches an object with his cane to verify the auditory impression of its location. In this case, c (cane) is followed by v, and the letter that labels the perception being verified. For example, if a subject hears the sound reflected from a pole, and touches the pole with his cane to verify the auditory perception, the score P(a)P(c.va) is assigned.

A direct perception might also be the result of an attempt to verify a remembered spatial fact, as when a blind pedestrian finds a lamp post with his cane in order to verify his memory of its location. An assertion is scored M (memory) if the spatial fact it reports is the remembered result of a perception, inference, generalization, or a communication. Thus, if a blind subject reports that a mailbox is immediately to the left of the curb cut on which he is standing, because he remembers the earlier haptic perception of the mailbox, his score includes M for the remembered mailbox. When M is scored, it is qualified in one or more ways, and its qualifiers are enclosed in parentheses and separated by periods. If a remembered fact is a feature of the current space that was perceived on an earlier occasion, the qualifier s (specific) is added.

Sometimes, although remembered information acquired on some prior occasion enters into the decision of the moment, it is not possible to relate the evoked memory to the perception of any specific feature of the current space. This would occur when, for instance, a blind pedestrian's knowledge that a certain intersection is not controlled by traffic lights enters into his decision about when to cross a street, but the pedestrian cannot point to any specific feature of the current situation that has evoked the memory. Remembered facts of this sort are probably evoked by associative links with other items of remembered information.

If a memory is evoked by a perception, the M component of the score is immediately preceded by the P component. If the memory is evoked by association the M component of the score is not preceded by a P component. If the evoked memory is of other situations that are like the present situation in some degree, so that prediction of features of the current space is possible, even though the current space has not been experienced before, the qualifier g (generalization) is added.

Occasionally, a subject's assertion indicates that his movement, at least for a short distance, has been guided by remembered headings and distances, learned during earlier practice on the current route, rather than by sensory feedback. When this is the case, the M component of the score assigned to the subject's assertion is qualified by s (specific), and by dr (dead reckoning).

Sometimes, a blind pedestrian touches an object with his cane, listens for an echo, or makes some other observation, not to acquire new information, but rather to verify an expectation concerning the location of some remembered feature of the space in which he is operating. If an observation of any kind has been made for the purpose of verifying an expectation, a ve (verify expectation) is added. If, for instance, a blind pedestrian touches a remembered pole with his cane to verify its presence and position, his assertion is scored M(s)P(c.ve).

An assertion is scored I (inference) if the spatial fact it reports is a conclusion drawn from the available evidence. When I is scored, the perceptual and memorial components of the evidence on which it is based are enclosed in brackets. For instance, if a blind subject concludes that an intersection is controlled by traffic lights because he hears the sounds made by moving traffic on the parallel street and by halted traffic on the cross street, followed by the sounds made by moving traffic on the cross street and by halted traffic on the parallel street, his assertion regarding this situation might be scored I[P(a)P(a)M(g)].

By making the distinction between spatial facts that are remembered or inferred, and spatial facts disclosed by direct perception, it will be possible to evaluate the hypothesis that the space in which the blind pedestrian performs the mobility task is, to a considerable degree, a remembered space and a space inferred from evidence at hand, whereas the space in which the sighted pedestrian performs the mobility task is, to a considerable degree, a space that is known immediately by perception. To say that a spatial fact is an inference is to say that it has been established, not by observation, but by performing logical operations on the available evidence. Therefore, if blind pedestrians must make greater use of inference than sighted pedestrians in order to perform the mobility task, it follows that they must depend more heavily than sighted pedestrians on memory and reasoning in order to perform that task. accordingly, performance of the mobility task by blind pedestrians is more strongly influenced than performance of the same task by sighted pedestrians by pitfalls (such as distorted memorial representations, incomplete memorial representations, memory lapses, faulty logic, and the like), that beset memory.

In the partial protocol that follows, the subject and the experimenters are identified by the initial letters of their last names. The protocol also includes statements labeled as comments, that have been added to clarify statements made by the subject or to indicate the need for clarification, to draw the reader's attention to the significance of statements made by the subject, and to explain scoring decisions.

The route on which subjects were observed went south from the front door of the Smith- Kettlewell Institute in San Francisco, California, to the comer of Webster and Clay, west along the north side of Clay to Filmore, south along the east side of Filmore to Bush, east along the north side of Bush to Buchanan, and north along the west side of Buchanan to Clay. All subjects were familiar with the segment of the route from Smith-Kettlewell to the comer of Filmore and Bush, and all of the subjects had little or no familiarity with the segment of the route from the corner of Fdmore and Bush back to Smith-Kettlewell.

Protocol

Subject: Mike Cole

Experimenters: John Brabyn and Emerson Foulke

C: We go out here and I'm going to look for this echoey garage, and I hit this wall here.
IM(s)P(a.ve) 2P(c)

B: You veered to the left into the garage.
Comment.- C's memory of the garage is probably evoked, not by an environmental cue, but by an associative cue. He verifies his expectation by listening for the echo. Because he veers as he passes the entrance to the garage, he hits the garage wall with his cane.

C: And now we've got the wall, so I'll go up here and hit this crossing here.|
3P(a) 4M(s)P(a.ve)P(e.ve.ca) 5M( dr)
Comment. C is referring to the garage wall that is parallel to the sidewalk, and on his left. He maintains his position on the sidewalk by monitoring the sound reflected from the garage wall. He remembers this situation and expects the garage wall to come to an end shortly. He verifies this expectation by listening, and by finding the end of the wall with his cane. Having found the end of the wall with his cane, he knows where he is in relation to the curb he will encounter shortly, and this knowledge makes cognitive anticipation possible. His course, as he turns out toward the street, is a remembered course, and he is not guided, at least for a few feet, by perceptual feedback.

B: How did you determine where to turn, where to cross?

C: Being familiar, I waited until I got to the end of that wall and then came out.

B: Did you detect the end of the wall with your cane tip, or did you hear it?

C: I was pretty much using the cane. However, I did hear it, too.

C: Now this street, I like to cross it via aggressiveness.
6P(a)

F: How did you decide when to cross?

C: There wasn't much happening, and I know that this is a know-about sort of thing, where if you get out there and are seen, you'll probably make it alive.
Comment.- Another indication of a risk-taking policy. C bases his crossing decision on auditory information, but he is willing to settle for less than certainty.

C: I also did make a kind of outward arc to try to hit this intersection, I mean the curb, toward the outside of it.
7M(dr)

C: I like to do that. Sometimes there's curb cuts, but I hate to get tangled up in whatever might be on the inside part; poles and whatnot.
Comment: By dead reckoning, C determines a course which should bring him to the part of the curb he wants to hit. By making an outside arc, C increases the risk of missing the sidewalk on the other side, and walking in the parallel street, but he reduces the risk of getting tangled up in the poles and other objects he might encounter if he were off a little in the opposite direction.

F: What was the name of the street you just crossed?

C: I just crossed Webster Street. Now I am going to face west and head off down Clay Street.

C:And there's a pole. and so I've got to get over toward the inside part of the sidewalk.
8I[P(c)M(g)]
Comment: C finds the pole with his cane. Finding the pole evokes the remembered generalization that poles of this type are usually near curbs. The inductive inference of which this generalization is the result took into account numerous occasions in the past on which poles of this type were close to the curb. It follows that the pole he has just found is near the curb. Because he is near the pole, he must also be near the curb, and he decides to move to the right, away from the street, in order to avoid, as he walks, the poles and other articles of street furniture that are usually located close to curbs. 7he generalization that C applies to the present situation is probably the product of inductive reasoning that took place sometime in the past. C probably stored the generalization in memory and recalled it on this occasion. In my scoring I am not distinguishing between remembered inferences and inferences that occur in the present. However, this might be a useful distinction.

C: Now the cane is telling me that we have some dirt and so on the left.
9P(c)

C: My arm brushed a pole.
10P(d)

C: I get some auditory information off the dripping / off these buildings.
11P(a)

C: Now Clay Street has these drops off to the right, so that you not only have a building there, but you have the sense that the sidewalk falls away, and sometimes that adds quite a bit of echo, and it also I well, not quite a bit of echo, but some / it’s weird. It’s an unusual sounding thing.
12P(a)
Comment: From what C has said, it is not clear to me just what situation he is describing, but in any case, he is monitoring the sound reflected from the buildings on his tight to maintain his position on the sidewalk.

C: Obviously we have hammering coming up on the right.
13P(a)
Comment: At this point, the sound of hammering on metal is loud enough to mask other useful sounds.

C: And there’s an obstacle on the right, which I detected simultaneously with the cane and heard it. It is pallets or something.
14P(a) 15P(C)

B: Yes, that’s construction.

C: And now the rain introduces a lot more noise.
16P(a)
Comment: 7he sound made by falling rain is an effective masking sound that may hide sounds C would like to hear.

C: Oops, I’m hung up by something.
17P(c)
Comment: C finds some construction paraphernalia with his cane.

C: Now we’re off again. This is not the easiest sidewalk to know whether you’re centered. It never has been in all the years I’ve come here.

C: Motorcycle.
18P(c) 19P(a)

B: Did you find it with your cane?

C: The cane first, I think I heard it.
Comment: C means here that he thinks he detected its presence by echolocation. The motorcycle was parked, and its engine was not running.

C: Now we’re passing the first break in the real estate to the right, where you could actually hear rain falling between the buildings, at least that’s what I’m calling it.
2OP(a)

B: What cues are you using to stay roughly in the middle of the path?

C: There's a lot of information coming off this right side.

B: Echoes?

C: Yeah, from the cane, and from / you know this whole business of being able to hear it sort of, but not because it's doing anything, almost just its inertness seems to say something. And it's even faceted, you know. It's not a constant thing. There's ins and outs.
21P(a)
Comment: C is using the sound reflected from the indented building line to maintain his position on the sidewalk.

C: OK, we're arriving at Fillmore. I'm going to start paying attention to the traffic.
22M(s)P(a.ve)P(f.ve.ca)

B: How did you know that you had arrived here?

C: This is a familiar walk for me, and also, you had a couple of things happening. The building was ending on the right, the slope was flattening out, which it does at intersections, especially in the city, and I'm always trying to begin to think about what the traffic pattern is before I ever even get to the street, so if somebody's making a right turn, I start sort of taking on a defensive pose.
Comment: C is familiar with this section of the route. He expects the intersection shortly and has started listening to the traffic. The disappearance of the building line and the leveling of the sidewalk verify his expectation. 7his verification also allows him to fix his position in a remembered space, making cognitive anticipation possible. he now has an approximate memorial representation of his present position in relation to the curb he will be reaching shortly. He also notes that the leveling of the walk at the intersection is typical of San Francisco, which suggests a generalization that could apply to the present situation, but on this occasion, he is clearly remembering and expecting the leveling of a specific sidewalk.

C: And here we go. Now we have water running in the gutter in the drains.
23P(a.pa)

C: And I got some good parallel traffic.
24P(a)
Comment: 7he sound of water running in the gutter makes perceptual anticipation possible. Because he can interpret the sound of running water, he has perceived the location of the curb. C has turned and will be walking south. Therefore, the parallel traffic he hears is on Filmore, and because he hears it, he decides he can cross Clay.

C: Something's going here in front of me that I was not aware of, and there's one to my left, but I don't think he's going to back up and he was practically in our path here. I think he kind of was a little bit.
25p(a)

B: Did you detect that car from the sound that it made?

C: Yeah, I heard him start up, but the guy out in the middle of the intersection really did surprise me.

F: Did you cross at the wrong time?

C: I guess you could say I did, because I went behind a guy who was trying to get out into Fillmore. If I had been aware of him, I would have probably allowed him to make his move. Part of what can happen in this situation is that all this talking and being conscious of it is actually a distraction.

F: Right.
Comment.- Was C crossing Clay against the light, or were the cars in his path exercising there god-given right to turn tight on red?

Recommendations for Mobility Instructors

In the early days, mobility instruction emphasized specific movements of the body and specific manipulations of the cane. Not enough attention was paid to the spatial facts that could be acquired by direct perception or retrieved from memory, and how they might be used to guide the behavior of the pedestrian. I believe that more attention is now being paid to the role of information about space, but there is still room for improvement.

Many blind pedestrians could improve their performance by remembering to remember. Consider the bad pedestrian who is shown the way from the elevator to his hotel room. If he remembers the number and direction of the turns he made and the side of the corridor on which his room is located, he will be able to return to the elevator easily and without assistance. I have accompanied a blind pedestrian who, after one learning opportunity, followed, without assistance and without hesitation, a long and complicated route from the entrance to a shopping mall to a particular shop. He remembered accurately the pattern of path segments that connected the entrance and the shop. Having been alerted by observation of this performance, I paid attention to his performance on subsequent occasions, and found that he routinely remembered such information, and that this remembering had become so habitual, it occurred outside of awareness. Most blind pedestrians are not as good at remembering what they have done. They could be given practice in remembering, and this practice could be continued until they no longer forget to remember.

Mobility instructors often point out the utility of landmarks to their students, but I am not sure that they give them enough practice in the discovery and use of landmarks. Do they acquaint their students with the variety of spatial features that can be perceived, and the perceptual systems that can be used? Do they explain the role of landmarks in orientation? Are their students made sensitive to the information they must have to begin a journey, the information they must acquire to continue that journey, and the information they must acquire in order to know when they have reached the goal?

Mobility instructors often provide their students with an experience called a "drop off. " The student is driven to some point in a space with which he or she has had some experience, but the exact location of the drop off is not known. The student's task is to gather the information needed to determine the present location, and then decide on a route that will take him or her to a location indicated by the instructor. I regard this as good training in the use of inference to solve a spatial problem, but does it go far enough? Could a student be taken to a number of drop-offs and be given initial guidance in gathering the information needed to solve the problem, additional drop-offs in which no guidance was given unless requested, and a few final drop-offs in which no assistance was given?

Students are told to observe traffic patterns in order to gather information on which to base decisions about when to cross the street, but do they receive systematic practice in gathering information of this kind? Are they ever taken to an intersection, required to stand there for an extended time and report each time it is safe to cross and why? Does the instructor discuss with them the evidence on which they have based their decisions?

Students are told to pay attention to the sound of parallel traffic in order to walk along a sidewalk without getting too close to the street, but do instructors give them much opportunity to practice this skill? Are they taken to places in the city where sidewalks are very wide so that, if they walk in the middle of the sidewalk, the sound of parallel traffic provides the most dependable information for walking a straight course?

Do instructors have high expectations for their students? I believe that, in many cases, instructors should expect much more of their students than they do, should motivate their students to expect much more of themselves, and should make more demands of their students.

I could give other examples, but perhaps the point I want to make is now clear enough. In general, I believe that learning how to manipulate the cane accurately, though important, is not sufficient, and that mobility students need much more practice in acquiring the spatial facts they need to perform the task in which they are engaged. They need practice in acquiring spatial facts by direct perception, and in remembering the spatial facts they perceive. Beyond this, they need practice in using the information they acquire. In general, they need to understand the nature of the task in which they are engaged, and the perceptual and cognitive processes that make the task possible.

Conclusion

The visual system is, by design, well suited for the acquisition of information about the spatial extension of spatial features and about their arrangement in space. Other perceptual systems do provide some information about space but, in respect to quantity, accuracy, dependability, distribution, and timeliness, they are inferior to the visual system. Nevertheless, the information they can provide is sufficient for effective mobility. However, it is important to understand that since the spatial information needed by blind pedestrians is different than the spatial information used by sighted pedestrians, the task performed by blind pedestrians is a different and more complex task than the task performed by sighted pedestrians.

Further improvement in the mobility of blind pedestrians can certainly be brought about by better training of the sort provided by O&M instructors, but the dramatic improvement envisioned by those who believe that technology will, in time, provide the solution to the mobility problem, will not be possible until some way is found to provide more, and more accurate information about both immediate and remote space. To cite one simple example, blind pedestrians would be greatly benefited by having more information about the surface over which they will soon walk, and they need that surface information before the surface that provides it is under the tip of the cane or the foot of the pedestrian. Yet, the electronic aids built so far provide very little information about the surfaces over which pedestrians walk.

As a first step, if technology can rise to the challenge of enlarging the immediate space that can be observed by blind pedestrians, and of providing more detailed, accurate, and timely information about that enlarged space, significant progress will be realized. If, in addition, a way can be found to provide more accurate and detailed information about a greatly increased number of the spatial features in remote space, we will have come a long way toward realizing the objective proposed by Leonard (1968). Though his reach exceeded his grasp, Leonard set as his objective the achievement of mobility for blind pedestrians that compares favorably with the mobility of sighted pedestrians. For now, the best way to improve the mobility of blind pedestrians is to teach them to make more efficient use of their perceptual and cognitive abilities. They will travel better when they are better at using their canes and their brains.

References

Armstrong, J. D. (1977). Mobility aids and the limitation of technological solutions. New Beacon, 61(127), 113-115.

Barth, J. L., & Foulke, E. (1979). Preview: A neglected variable in orientation and mobility. Journal of Visual Impairment & Blindness, 73, 41-48.

Brambring, M. (1982). Language and geographic orientation for the blind. In R. J. Jarvella & W. Klein (Eds.), Speech, place, and action (pp. 203-218). New York: John Wiley & Sons, Ltd.

Chew, S. L. (1986). The use of traffic sounds by blind pedestrians. In Proceedings of The Louisville Space Conference, April 13-14, 1984 (E. Foulke, Chairman), pp. 97-105. A limited edition publication of the College of Arts & Sciences, University of Louisville, Louisville, KY.

Crossman, E. R. (1960). The information capacity of the human motor system in pursuit tracking. Quarterly Journal of Experimental Psychology, 12, 1-16.

Foulke, E. (1982). Perception, cognition, and mobility of blind pedestrians. In M. Potegal (Ed.), Spatial orientation: Development and physiological foundations (pp. 55-76). New York: Academic Press.

Foulke, E., & Berlá, E. P. (1978). Visual impairment and the development of perceptual ability. In R. D. Walk & H. L. Pick, Jr. (Eds.), Perception and experience (pp. 213-240). New York: Plenum Press.

Gibson, J. J. (1966). The senses considered as perceptual systems. Boston: Houghton-Mifflin.

Guth, D. A. (1984). How do blind pedestrians use traffic sounds for street-crossing alignment. In Proceedings of 7he Louisville Space Conference, April 13-14, 1984 (E. Foulke, Chairman), pp. 107-119. A limited edition publication of the College of Arts & Sciences, University of Louisville, Louisville, KY.

Hershman, R. L., & Hillix, W. A. (1965). Data processing in typing: Typing rate as a function of kind of material and amount exposed. Human Factors, 7, 483-492.

Howard, I. P., & Templeton, W. B. (1966). Human spatial orientation. New York: Wiley.

Leonard, J. A. (1968). Towards a unified approach to the mobility of blind people. American Research Bulletin, 18, 1-21.

Levin, H., & Kaplan, E. L. (1969). Listening, reading and grammatical structure. In D. L. Horton & J. J. Jenkins (Eds.), Perception of language: Part I. Proceedings of a Symposium of the Learning Resources and Development Center (pp. 1-23). Pittsburgh: University of Pittsburgh.

McLean, J. R., & Hoffman, E. R. (1973). The effects of restricted preview on driver steering control and performance. Human Factors, 15, 421-430.

Poulton, E. C. (1952). The basis of perceptual anticipation in tracking. British Journal of Psychology, 43, 295-302.

Poulton, E. C. (1954). Eye-hand span in simple serial tasks. Journal of Experimental Psychology, 47, 403-410.