A brief history of Eleanor Gibson's visual cliff research
York University, Toronto
Eleanor J. Gibson is best known in psychology for her classic "visual cliff" studies with babies. This study often features in introductory textbooks, usually accompanied by a picture of a baby against a checkered surface, peering cautiously over the edge of the visual cliff while its mother beckons encouragingly from the other side of the apparatus. (cf. Myers, 2001). It is this image most psychologists associate with Eleanor Gibson.
Summaries of the visual cliff study are prone to gross simplification or factual error, but even in careful accounts the study is decontextualized, becoming a timeless experimental entity, one more eternal psychological truth. Presenting the visual cliff experiment as a single unitary, ahistorical object means ignoring the experiment's place in an ongoing disciplinary dialogue and its original experimental logic. Seen in retrospect the experiment loses its groundbreaking quality—the rationale for the study is assumed to be consistent with the results obtained and these results themselves seem commonsensical. This retroactive consistency is characteristic of origin myths (Harris, 1979), and while the impact of visual cliff myth is perhaps not be as wide reaching as Gordon Allport's pronouncements on the origins of social psychology (Samelson, 1974), it is appropriate to consider it within the origin myth framework. Like Watson's classic Little Albert study, the distortions of the visual cliff study foster a "false sense of continuity" (Harris, 1979, pg. 157) within disciplinary history, and prepare the way for convenient pedagogical morals to be drawn.
Yet the lore of the visual cliff is not entirely wrong. While it was hardly the pinnacle of Gibson's career, the visual cliff work deserves its "classic" status. It introduced a valuable experimental apparatus that was adopted and adapted by many researchers for a variety of purposes and it also provided clear evidence for infants' early accurate perception of their environment—something that was hotly contested at the time. Gibson herself seems to have recognized its importance, referring to the study frequently in later books and articles to underline her points, although her reasons for thinking the experiment important differed significantly from popular interpretations. She was inclined to see the study as illustrating her views on perceptual learning, whereas popular accounts usually linked it to the nature-nurture debate.
The visual cliff experiments were not formulated in a void, but were part of a loosely-related contiguous line of research which Gibson conducted throughout her career. But they were also the result of the constraints Gibson was under at Cornell, where anti-nepotism rules limited her ability to experiment, and therefore influenced her choice of what research interests to pursue. The visual cliff research was the result of her partnering with a colleague with lab space and animals—a creative response to those limitations (Gibson, 2002). Therefore, in order to understand the context of the visual cliff, we must backtrack to Gibson's biography.
Gibson's love of psychology and her love of James Gibson arose about the same time and were not entirely unrelated. Eleanor met James at a graduation garden party at Smith College where she, a junior, was assigned to serve punch and he, a young professor, was assigned to greet parents. A sudden rain caused them to seek shelter in the same area—the consequence of which was that the next day Eleanor rushed back to campus to change her fall course schedule to include James' advanced experimental psychology class before catching her train home for the summer! But Eleanor was a psychology major before this, and had her own interests—she wanted to map the "laws of learning" (Gibson, 1979, ix), particularly in comparative psychology. She was attracted by the hard science promise of experimental psychology, and after completing a master's at Smith College and marrying her advisor, none other than James Gibson, she began her Ph.D. at Yale. In 1936, a woman being admitted to Yale was still considered a concession and Gibson soon found herself limited by expectations in the male-oriented program (Gibson 2002, pg. 20). Unlike the other women in the program, Gibson wasn't content with the child psychology or mental testing work offered by the two advisors who traditionally accepted women, and approached Robert Yerkes, hoping to work in his chimpanzee lab. Upon requesting his advisorship, Eleanor was promptly shown the door, Yerkes proclaiming "I have no women in my laboratory" (Gibson 2002, pg. 21).
Foiled in her attempt to do animal psychology, but still hoping to do experimental work, Gibson approached Clark Hull. Hull agreed to advise Gibson, provided that she use his strict behaviorist methods and theory (Gibson, 1991). Gibson chose a dissertation topic consistent with Hull's theory, but which also related to her interests in learning. The topic was stimulus generalization and differentiation, the perceptual aspect of conditioning. Gibson's notion of differentiation was functional, so her true views had to be "disguised" by behaviorist vocabulary in her dissertation, in order for Hull to approve (Gibson & Levin, 1979, pg. 243).
After receiving her Ph.D. in 1938, Gibson began working as an assistant professor at Smith where her husband still worked. Gibson was kept busy with full time teaching and, other than publications based on her dissertation, published no new research during this period. In 1940, Gibson had her first child, took a semester off, but soon returned to work after hiring a widow to look after her son. WWII interrupted this happy arrangement. James Gibson was recruited by the Air Force to create perceptual tests to select pilots, and this meant relocating to Texas and then California. Eleanor spent the duration of the war at home caring for their children (a daughter was born in 1943) while James collaborated with fellow psychologists in vision research. Although Eleanor did not have any official role in James' military research, she gained some satisfaction from their evening discussions of James' work and occasional social events with the other psychologist which gave her the opportunity "talk shop" (Gibson, 1991). James' studies of pilots' perception of flying offered him insight into the importance of movement for vision (or, more technically, the role of optic flow in the visual field) which was to prove influential in developing his ecological theory of perception. These views influenced Eleanor as well and would be the basis for some of her later research. The following quote from the report James wrote for the government on the vision research, which Eleanor later quoted in her autobiography, shows how important his work would be for Eleanor's visual cliff research:
The problem of three-dimensional vision, or distance perception, is basically a problem of a continuous surface which is seen to extend away from the observer. All spaces in which we live include at least one surface, the ground or terrain. If there was no surface, there would be no visual world, strictly speaking. Whether we stand on it or fly over it, the ground is the basis of visual space perception both literally and figuratively. (Gibson, 2002, pg. 41)
In James Gibson's emerging views, the ground or surface was essential for perception, an emphasis which Eleanor would affirm in her visual cliff studies.
When the war ended the Gibsons returned to Smith and resumed teaching. Eleanor discovered that Smith's psychological labs had been a casualty of the war, and would not be quickly reinstated (Gibson, 2002). Not long after returning to Smith, James received an offer from Cornell. Unlike Smith, Cornell had anti-nepotism rules, which meant that Eleanor could not also be on the faculty. The Gibsons decided to accept, despite this drawback, planning that Eleanor would seek out her own research opportunities. For the next 17 years, until she was finally made a professor in 1966 after Cornell stopped funding her husband's work, Gibson carried out her research by combining work as a research assistant with government grants and partnering with Cornell faculty (Gibson, 2002).
The first of these endeavors was working as an assistant for two years at the Behavior Farm, a laboratory of Cornell professor Howard Liddell. Liddell was a staunch behaviorist engaged in the classical conditioning of goats using shock in an attempt to induce experimental neurosis. Gibson had her doubts about the value of this project, but used the experimental set up to study her own interests by comparing the goatsÕ reactions when a shock was avoidable with when it was not. Since goats were bred on the farm, Gibson also set up an observational study of development and imprinting, with twin pairs of kids being divided into experimental and control animals. But this study came to a premature end when Gibson returned to the farm one weekend only to discover that some of her subjects had been given away as Easter presents (Gibson, 2002).
After this failed attempt at research at the Behavior Farm, Gibson received military funding to investigate training of the judgment of distances. These studies took place outdoors, on a Cornell athletic field, rather than in a lab, using Air Force recruits as subjects. The effect of training on judgment was rather modest, but they did show that the subjects were good at perceiving distances, even if the experiment did not discover how they made those judgments. In An Odyssey in Learning and Perception (Gibson, 1991), Eleanor noted that these distance perception studies were in line with more traditional ideas of perception, whereas her work just a few years later that focused on perspective transformations in distance judgments had been influenced by James' developing theorizing. According to Eleanor, "James Gibson was my tutor as regards perception, but although we argued about experiments . . . we never really disagreed, as we did sometimes about learning" and "By the 1950s he was well on the way formulating a new theory about how we perceive" (Gibson, 1991, pg. 203).
This was Eleanor Gibson's situation just prior to the visual cliff studies: she had compensated for her lack of lab space by patching together some experimental situations, and was fine-tuning her views of perceptual learning in coordination with James' developing theory. Richard Walk, a newly arrived psychology professor at Cornell who taught learning and had his own rat lab, provided Eleanor with the next opportunity for research, the visual cliff.
Before moving on to the origins of the visual cliff studies, it is important to first examine some of the stories that have grown up around the famous experiment. As mentioned in the beginning of this paper, the image associated with Gibson is of a baby on the visual cliff. This is despite the fact that rats, not human babies, were the first organisms tested on the visual cliff. The various animals tested prior to babies come last in accounts of the visual cliff, if they are mentioned at all. In these accounts it appears that Gibson and Walk set out to study the depth perception of babies in order to determine whether the ability was innate. In some cases, Walk and Gibson are said to be nativists, who designed the study attempting to find evidence to support their position (cf. Hock, 2005).
Although there are multiple plausible reasons for this phenomenon (such as textbook writers' interest in illustrating the nature vs. nurture debate), some of this myth is of Walk and Gibson's own making. Both the iconic pictures and the emphasis on the human participants come from a 1960 article called The "visual cliff" that Gibson and Walk published in the Scientific American. Gibson and Walk highlight the role of babies in the experiment; the article begins "Human infants at the creeping and toddling stage are notoriously prone to falls from more or less high places" (Gibson & Walk, 1960, pg. 67). Although Gibson and Walk do not take a nativist position in the article, they do frame their research using the nature versus nurture debate, so the common mistaken impression is understandable.
What Gibson and Walk were actually attacking was the old idea that visual information has to be supplemented by association— the visual stimulation of falling is associated with the experience of falling down and thus the child learns to avoid heights. Gibson was not attempting to prove that the perception of height was an innate quality but simply that the environment provides the baby with sufficient information to perceive a drop off. In Gibson's view, development, and locomotion in particular, plays a role in developing the baby's ability to perceive depth, so she is not a strict nativist. Gibson argued that depth perception is an early ability and that babies do not gain it by trial and error but by learning to perceive, by learning to differentiate surfaces as they move around their environment. Gibson and Walk's conclusion in the paper is that "a seeing animal will be able to discriminate depth when its locomotion is adequate, even when locomotion begins at birth" (1960, pg. 71), not that this ability is innate.
A related myth has to do with where the idea of the experiment arose. David Myer's popular textbook provides a typical account: "Gibson's inspiration for these experiments occurred while she was picnicking on the rim of the Grand Canyon. She wondered: would a toddler peering over the rim perceive the dangerous drop off and draw back?" (2001, pg. 215). This story, which generally includes the Gibson children frolicking on the edge the cliff, only makes sense when combined with the impression that the study was primarily about babies and secondarily about animals. The anecdote has also been used to cement the importance of the nativist-empiricist debate in her work—one retelling has James Gibson the nativist proposing "a simple but ruthless experiment" with their infant to his horrified empiricist wife (Sutherland, 1992, pg. 45).
This story is very loosely grounded in truth; the Gibsons stopped at the Grand Canyon in 1946, on their way from California to Massachusetts after the war, but at this point the Gibson children weren't infants but ages three and six (Caudle, 1990). What really happened, according to Gibson herself, was that she was nervous about the children's proximity to the edge, but "My husband reminded me that they could see the depth as well as I could, and I believed him." (2002, pg. 47). According to Gibson, this story was remembered, not by her, but by Walk, when they were trying to come up with additional discrimination tasks for the rats (Gibson, 1991). However, in her later biography Gibson says, "Contrary to a popular myth, this occasion was not the inspiration for my later research on the visual cliff" (Gibson, 2002, pg. 47). It is not clear from Gibson's conflicting accounts whether Walk never mentioned the Grand Canyon incident or whether she simply does not consider it to have been influential in the idea's generation.
Another incident, also not the inspiration of the visual cliff studies, but which Gibson says was "good preparation" for it, took place on the Behaviorist Farm (Gibson, 1991, pg. 103). Gibson was washing the goats intended for her twin study as they were born, and had just finished washing the first kid when the second started to emerge. As Gibson hurriedly looked about for a place to keep the washed goat clean, the watching farm manager suggested placing it on a high camera stand nearby. To Gibson's surprise, the newborn stood calmly on the stand until Gibson could deal with it, a reminder that goats are prepared to deal with heights from birth (Gibson, 1991).
In truth, the origins of the visual cliff were far more mundane, having more to do with rats and pragmatics than with babies and theory. Gibson was interested in comparative research, and partnering with Walk offered her the opportunity to work with rats. Walk and Gibson successfully applied for a grant from the Natural Science Foundation and set to work designing a study that would look at the role of environment in development.
In their first experiment, based on Donald Hebb's work, hooded rats in an "enriched" condition were exposed from birth to metal shapes on their cage walls. At three months they were tested against control rats in a discrimination task using the shapes, and it appeared that prior exposure to the shapes had given them an advantage (Gibson & Walk, 1956). But a series of nine follow up studies failed to replicate this result, and it appeared that simple passive exposure to shapes was not enough to create a difference in discrimination ability. Some of the later experiments compared light-reared and dark reared rats, and found that the 90-day-old dark-reared rats performed just as well on the discrimination task as the light-reared rats that had spent the 90 days being exposed to the shapes (Gibson, Walk & Tighe, 1959).
It was this "serendipitous" (Gibson, 1991, pg. 141) inclusion of dark-reared rats in the study that resulted in the creation of the visual cliff. Raising large groups of rats in the dark was labor intensive, so Walk and Gibson decided to put the rats to good use by testing the last group of dark-reared rats on a second task. The idea of a visual cliff grew out of a study by Karl Lashley (Lashley & Russell, 1934) which showed that dark-reared rats took the same amount of time to gage the distance between a jumping stand and a platform as light-reared rats, indicating that they were not dependent on experience to perceive distance. Walk and Gibson decided to create an artificial cliff, expecting that the dark-reared rats would be more likely to walk indiscriminately on the 'deep' and 'shallow' sides, since they had presumably failed to develop depth perception in the dark (Gibson, 1991). This set up had the advantage of correcting a flaw in Lashley's study— it would eliminate the training period in light which was necessary for the jumping task (Lashley & Russell, 1934).
Gibson and Thomas Tighe, their research assistant, quickly constructed a cliff apparatus with items that they found around the lab. In contrast to later more sophisticated models, it was simply a sheet of glass, backed by checked wall paper and held above the ground with some clamps and rods. The deep side was 53 inches deep whereas the shallow side was 3 inches deep—in terms of the distance of the wallpaper from the glass surface. The deep side was separated from the shallow side by a center board three inches high, where the rats were to be placed at the start of the experiment, and then observed for five minutes (Walk, Gibson & Tighe, 1957). The center board was the result of some preliminary tests with light-reared rats, in which it seemed that the rats weren't paying sufficient attention to the surface beneath them with simply a flat glass surface. The board made the rats look at the surface and also provided a convenient system of measurement: the researchers would record the side of the board a rat chose.
To the researchers' surprise, the dark-reared rats acted the same as their light-reared controls and consistently descended on the shallow side, avoiding the deep side. At this point the experimenters began to be worried, wondering if there was something that was biasing the rats towards the shallow side, a draft or odors, perhaps (Gibson, 1991). They checked this by adding wallpaper just beneath the deep side, so that the optical depths of the two sides were identical. After that change the rats explored both sides, often crossing the board to the other side multiple times. The consistency of the effect had been completely unexpected; Gibson sums up the collective sentiment by quoting Tighe as saying "I wouldn't have believed it if I hadn't seen it" (Gibson, 1991, pg. 141). They published their surprising results in Science (Walk, Gibson & Tighe, 1957)
They soon built a more permanent visual cliff apparatus and tested it on less than one day-old chicks, finding that chicks with no prior experience completely avoided the deep side of the visual cliff. The same was true of newborn kids and lambs, tested on a larger apparatus (Gibson & Walk, 1960). Since the animals avoided the deep side, Gibson also tried setting the lambs or kids on the deep side the visual cliff, which resulted in stereotyped defensive behavior (freezing, refusing to put its feet down, etc.). Further experiments at the Behaviorist Farm resulted in a modification of the apparatus so that the wallpaper's distance from the glass was easily adjustable. The researchers found that the goats would freeze if the wallpaper dropped more than a foot from the glass (Gibson & Walk, 1960) and that, despite their experience of the glass's support, the animals never became comfortable on the deep side of the visual cliff.
It was not until after this animal research that Walk and Gibson finally pursued the idea of recruiting human subjects. This appears to have been because they did not have easy access to infants—Gibson had to recruit crawling babies by placing an ad in the newspaper, promising to pay three dollars (Szokolszky, 2003). Despite her husband's concern that mothers would be too suspicious of psychologists to volunteer, Eleanor had no problem recruiting sufficient babies for their first experiment. Thirty-six babies ranging from six months to 14 months were tested on the visual cliff, with their mothers alternating standing just beyond the shallow and the deep side of the cliff, spinning a pinwheel and encouraging the baby to crawl to them. Of the 27 babies who left the board, all crawled on the shallow side, only three also crawled onto the glass of the deep side (Gibson & Walk, 1960). Some babies cried or backed away from their mother when encouraged to cross onto the deep side. Gibson noted their predominating dependence on vision—several patted the glass, but despite the tactile reassurance refused to move onto the glass (Gibson & Walk, 1960).
The Scientific American article on the visual cliff studies (Gibson & Walk, 1960) also described experiments on puppies, kittens, and turtles, though these experiments may have occurred after the studies of babies. A later article by Gibson alone (Gibson, 1963) mentioned experiments on pigs, adult chickens, and monkeys, all of which showed perception of the cliff, but it is not clear when these studies were conducted, nor by whom. Gibson might have pursued further research using the visual cliff, however, in 1959 Walk left Cornell and they agreed that only one of them should work on the cliff (Szokolszky, 2003). Walk continued research on the visual cliff, expanding the animals tested, testing on adult human subjects, the effect of monocular vision, and the effect of different patterns beneath the glass cliff. Gibson's interests turned towards human perceptual development, and she also spent the next several years researching reading. It was not until the 1980s that Gibson returned to experiments with infants similar to the visual cliff (e.g. Gibson et al., 1987), such as experiments giving infants the choice between crawling on a hard surface and or waterbed-like surface.
Caudle, F. M. (1990). Eleanor Jack Gibson (1910- ). In A. N. O'Connell and N. F. Russo (Eds.), Women in psychology: A bio-bibliographic sourcebook (pp. 104-116). New York: Greenwood.
Gibson, E. J. (1963). Perceptual development. In H.W. Stevenson, J. Kagan & C. Spiker, (Eds.), Child Psychology, (144-195). 62nd Yearbook National Society for Study of Education, Chicago: University of Chicago.
Gibson, E. J. (1991). An odyssey in learning and perception. Cambridge: MIT Press.
Gibson, E. J. (2002). Perceiving the affordances: A portrait of two psychologists. Mahwah, NJ: Lawrence Erlbaum Associates.
Gibson, E. J & Levin, H. (1969). Psychology of reading. Cambridge, MA: MIT Press.
Gibson, E. J., Riccio, G., Schmuckler, M. A., Stoffregen, T.A., Rosenberg, D. & Taormina, J. (1987). Detection of the traversability of surfaces by crawling and walking infants. Journal of Experimental Psychology: Human Perception and Performance, 13, 533-544.
Gibson, E. J. & Walk, R. D. (1956). The effect of prolonged exposure to visually presented patters on learning to discriminate them. Journal of Comparative and Physiological Psychology, 49, 239-242.
Gibson, E. J., & Walk, R. D. (1960). The "visual cliff." Scientific American, 202, 67-71.
Gibson, E. J., Walk, R. D., & Tighe, T. J. (1959). Enhancement and deprivation of visual stimulation during rearing as factors in visual discrimination learning. Journal of Comparative and Physiological Psychology, 52, 74-81.
Gibson, J. J. (1967). Autobiography. In E. G. Boring and G. Lindzey, (Eds.), History of Psychology in Autobiography, Vol. 5, (125-143). New York: Apple-Century-Crofts.
Gibson, J. J. (1979). Forward: A note on E. J. G. by J. J. G. In A. D. Pick (Ed.), Perception and its development: A tribute to Eleanor J. Gibson (ix-xii). Hillsdale, NJ: Lawrence Earlbaum Associates.
Harris, B. (1979). Whatever happened to Little Albert? American Psychologist, 34, 151-160.
Hock. R. R. (2005). Forty studies that changed psychology (5th ed.). Upper Saddle River, NJ: Pearson Prentice Hall.
Lashley, K. S. & Russell, J. T. (1934) The mechanism of vision. XI. A preliminary test of innate organization. Journal of Genetic Psychology, 45, 136-144.
Myers, D. G. (2001). Psychology (6th ed.). New York: Worth Publishers.
Samelson, F. (1974). History, origin myth, and ideology: "Discovery" of social psychology. Journal for the Theory of Social Behavior, 4, 217-231.
Szokolszky, A. (2003). An interview with Eleanor Gibson. Ecological Psychology. 15, 271-281.
Sutherland, S. (1992). Review: Mrs Gibson leaps in at the deep end. New Scientist, 1847, 45.
Walk, R. D. (1979). Depth perception and a laughing heaven. In A. D. Pick, (Ed.), Perception and its development: A tribute to Eleanor J. Gibson. Hillsdale, NJ: Erlbaum.
Walk, R. D., Gibson, E. J. & Tighe, T. J. (1957). Behavior of light- and dark-reared rats on a visual cliff. Science, 126, 80-81.