Notes
What's Wrong with Using Photographs of the Environment in Environmental Perception Research?
Harry Heft, Emeritus Professor, Denison University
It is not reality that photographs make immediately accessible, but images. . . . [P]hotographic images tend to subtract feeling from something we experience at first hand and the feelings they do arouse are, largely, not those we have in real life. (Sontag, 1977, pp. 167-168)
Abstract
For decades, researchers of environmental perception have relied to a great extent on photographs of landscapes and urbanscapes in their efforts to understand the basis for environmental preference and to establish whether natural environments can play a psychologically restorative role for individuals experiencing stress and attentional dysregulation. In doing so, it has been assumed that photographs of environments are sufficiently similar to direct experience of those same environments in situ that they can be utilized as surrogates for the latter without undue concern; and indeed, several early comparative assessments seemed to support such a view. However, a close look at these assessments casts doubt on this conclusion, as do empirical findings from several experimental studies. Careful consideration of camera technology also makes it clear that photographs are the products of mechanized, constrained processes that are configured to achieve particular ends through the implementation of particular practices. Resulting photographs are static representations, frozen in time; and at best, they approximate what an environment looks like from a fixed, and often a selected viewing position. In contrast from the perspective of ecological psychology, visual perception is recognized as a dynamic perception-action process by means of which events as well as structure in the context of change are perceived. Research from this perspective has shown that many informative environmental properties are only perceivable from a moving point of observation and/or when features are in motion. In the absence of appropriate comparative studies, these considerations raise the possibility that much of the extant research literature looking into environmental perception bear mostly on picture perception rather than how the environment is experienced in the course of the direct experience of everyday life.
Introduction
Investigators in environmental psychology and design research are typically faced with a common decision from the outset of their work: how is the environment of interest to be included in the research? The most obvious choice is to carry out research within those environments, either by engaging individuals already in those locales or by transporting participants to them. But the most obvious choices do not necessarily turn out to be the most practicable. Conducting research within the environments of interest – that is, e.g., within landscapes and cityscapes as well as within homes, schools, and offices – present researchers with multiple challenges. To name just one is the challenge of maintaining adequate control of conditions in those environments so that apposite inferences can be drawn from the results. However, in all but highly constrained circumstances (e.g., research laboratories), environmental conditions can vary considerably owing to a wide range of changing circumstances, from varying lighting and weather conditions over time and to changing patterns of human activity.
The principal way in which such difficulties have been dealt with in the research literature to date has been to employ photographic surrogates of the environments of interest. Because photographs appear to ‘capture’ the visual appearance of some environment at an instant in time, perforce they allow researchers to control for changing conditions. Photographs are also advantageous for other reasons: they allow researchers to sample across a variety of environments, and their use in research eliminates the logistics of transporting participants to various, and sometimes remote, locations. In other words, there is much to recommend them.
However, the use of photographs in research must be predicated on the assumption that these surrogates of environments are experienced by perceivers in approximately the same manner as the environments they represent if perceivers were to experience those environments first-hand. To the extent that photographic surrogates are not experienced as environments are first-hand, their validity for research is in question, at least to some undetermined degree.
In the early years of this research field, several investigators recognized the need to assess the adequacy of photographs particularly in research investigating how environments are experienced. As a result, a handful of studies were conducted to evaluate this question (see reviews by Daniels, 1990; Zube, Simcox, & Law, 1987). By the start of the 1990’s, the consensus among researchers seemed to be that photographs were indeed adequate environmental surrogates (e.g., Kaplan & Kaplan, 1989; Stamps, 1990; Zube, et al., 1987). What followed were several decades of environmental perception research relying often exclusively on photographs, slides, or images on computer screens.
In this brief paper, I will show that the empirical justification for the use of photographs in environmental perception research is far from firmly established. Furthermore, photographs due to their inherently static nature necessarily omit numerous salient qualities of environments as they are immediately experienced. I will also argue that the theoretical grounds for using photographs in environmental perception is based on long-standing traditions in the study of visual perception that are too limited in light of more recent scientific advances and in relation to how individuals engage environments directly. Obviously, there is a great deal at stake in raising these issues, both for the assessment of the existing research literature, and more importantly, for the future course of research.
The assumption that photographs are sufficiently adequate for research in environmental perception will be examined in two ways. First, we will consider the evidence in support of this assumption, followed by evidence that raises questions about it. Second, apart from such empirical considerations, there are important conceptual ones. Photographs would seem to an obvious, and indeed, non-controversial choice for research because cameras seem to operate much like mammalian vision does. I will challenge that claim by locating its origins in an early historical era in the study of vision that precedes by roughly 300 years the 19th century revolution in scientific thinking driven by the theory of evolution in the life sciences and field theory in physics. I will then briefly point to some features of an alternative approach to visual perception, the ecological approach, which is a product of those later developments. That alternative suggests that different methods of investigation are needed as alternatives to the now nearly reflexive choice of photographs.
The Evidence for and Against
Many of the initial studies comparing the experience of viewing photographic surrogates and first-hand experience were summarized in the chapter contributed by Pitt and Zube to the Handbook of Environmental Psychology edited by Stokols and Altnam (1987), as well as the review paper by Zube, Simcox, and Law (1987). By contrast, it is noteworthy that none of the chapters in the two later appearing ‘handbooks” of environmental psychology were solicited to address this question in any substantive manner (Bechtel & Churchman, 2002; Clayton, 2012). I can only assume that either the issue was taken to be settled by then or as not warranting any further discussion. Pitt and Zube (1987) earlier summarized their survey of the research by stating: “the reliability and validity of photographic techniques (both slides and photos) as landscape simulation devices is well documented” (p. 1024). Stamps (1990, 2010) reached the same conclusion based on two meta-analyses of existing research, and yet a closer look at these publications yields a less clear-cut assessment.
The studies cited in the reviews by Pitt and Zube (1987) and Zube, et al. (1987) are a rather mixed bag, which in most cases do not compare photographic surrogates to on-site evaluations, but instead compare different types of surrogates (e.g., black & white versus color); photographs of different target sites (e.g., forests, landscapes, air quality); and assessments among different populations of raters (e.g., forestry professionals versus lay public). Finding a correspondence in each case is taken as convergent validity to warrant the use of photographs as surrogates for first-hand experience of environments. In a more selective, if partial review, Daniel (1990) reports “correlations between photo-based and direct ‘on-site’ assessments have been .80 or greater” (p. 634), that is, just under 70% of the variance of responses. That promising conclusion is severely undercut by a discerning, and yet often overlooked study by Hull and Stewart (1992).
These investigators systematically obtained on-site landscape ratings from hikers at specific sites during different phases of their route (i.e., up-hill and down-hill) through a national forest. These ratings were compared to ratings by the same sample of hikers of color photographs taken at those same sites three months and then nine months after their on-site experience. Working with the same sample on each occasion was a crucial step because it allowed Hull and Stewart to examine within participants’ assessments under both viewing conditions and over time. In contrast, research assessing landscape evaluations under different conditions (e.g., on-site versus photographs) typically employ different samples in each case. That procedure only makes possible comparisons or correlations among sites based on group ratings; even though it is conceivable that group means between groups may be comparable, and correlations high, even as the stability of individual assessments vary across conditions. If so, group-based assessments may mask varying responses across conditions ‘within individuals,’ raising questions about their reliability and the validity of photographs. That is indeed what Hull and Stewart found.
To be specific, while 20% of the 90 participants’ ratings on-site and later with respect to the photographs were correlated appreciably (i.e., >.80), for 38% of the participants (or 34 of 90), the correlations were not statistically significant. Looked at another way, the average individual correlations between on-site and photograph evaluation within individuals was .60 (or 36% of the variance accounted for) which just exceeds the .05 significance level. They conclude: “This suggests that the validity and/or reliability of photo-based assessments of landscape scenic beauty is in doubt for a sizeable percentage of [participants]” (p. 108). This judgment runs counter to the more sanguine assessments offered in other studies. We will return to other aspects of their data below.
As for the meta-analyses, Stamps (1990) reports a .86 combined correlation after including in the analysis eleven studies in which ratings of photographs and on-site evaluations of settings were compared. In a later meta-analysis, Stamps (2010)) conducted a meta-analysis of studies that compared the use of static displays or photographs and the use of dynamic displays such as video. The latter could be judged on their face, at least, as more like direct experience than photographs. In this second meta-analysis, he again reports a correlation across studies of a similar magnitude (.82). Based on these analyses, Stamps concludes that assessments of photographs and on-site assessments are sufficiently similar, and that static displays and dynamic displays are experienced in similar ways, that the use of photographs in environmental perception research can be employed without undue concern.
However, meta-analyses are only an effective way of amassing findings across multiple studies as long as those studies are sufficiently similar as to their research settings, methods, materials, populations, etc. When meta-analyses are carried out with respect to laboratory-based studies, the chances that setting conditions are comparable across studies is far greater than when studies are carried out in situ. Judging solely from the titles of the studies included in Stamps (1990), sufficient commonality is difficult to establish with much confidence. In Stamps (2010), 32 sites were included in the meta-analysis, 17 which are labeled broadly ‘landscapes’, while the remaining include rooms, museums, plazas, hallways, and outdoor paths, and the dynamic medium in the studies included virtual reality, videos, and one film. Accordingly, in spite of the statistical outcomes of these meta-analyses, the assortment of studies included in each undermine a clear-cut conclusion
Let us consider other evidence that lead us to question the comparability of photographs and either onsite experience or dynamic presentations of landscapes. Landscape photographs are inherently static or ‘frozen’ in time. On occasion, as just noted, techniques for presenting representations of environments dynamically have been employed with the intention of better simulating first-hand experience. Dynamic forms of media include film, videotape, and computer graphics (e.g., immersive or virtual reality). To my knowledge, only two studies have compared reactions to photographs (‘stills’) and to dynamic displays, and in both of those cases videotape was employed.
In a study that was primarily designed to assess qualities relating to path information utilized in way-finding, Heft and Nasar (2000) asked participants to evaluate portions of a path where turns occurred, and these evaluations were made with respect to either a still image midway through the turn (i.e. a single video frame) presented for six seconds or a video clip including that frame and of the same duration. Groups of participants rated either the static or dynamic displays on several dimensions. Most pertinent to the topic at hand was the unexpected finding that static displays received significantly higher mean preference ratings than did the dynamic displays. Participants reported ‘liking to look at’ the still images more than the brief dynamic displays. To reassess this finding, Heft and Poe (2005) presented static (a video frame) and dynamic displays of a set of landscapes to participants and asked them to evaluate the degree to which they ‘would like to spend time at those sites’ and how much they liked ‘looking at’ the displays. Although the ratings for spending time did not vary across static and dynamic presentation modes, the preference ratings did, replicating the previous finding of participants assigning higher preference ratings to the static displays. These sets of findings suggest that a distinction must be made between these two types of perceptual experiences: those from a stationary observation point and those from a moving point of observation (see below).
Notably, we also see just such a differentiation of experience type in Hull and Stewart (1992) where 11 out of the 12 sites that were evaluated received more positive ratings for relative beauty when presented as photographs rather than on-site.
Finally, some recent data show that individuals look at photographs and at environments first hand in different ways. Sun, et al. (2018) recorded individuals’ eye movements for a 60 second interval when they were either looking at a Japanese garden from a single viewing position at the periphery of the garden, through a large glass window, or when looking at a projected slide of the garden taken from that same location. They found that there were more ‘points of fixation’ when participants looked at the slide than when they viewed the garden directly or through the glass pane; and that the ‘duration of eye fixations’ were longer when looking at the slide than in the other two conditions. At a minimum, these behavioral data are consistent with the preference data reported above suggesting that perceivers engage photographs in a different manner than they do when experiencing the environment first-hand.
Why assume that Photographs Are Adequate Surrogates for First-Hand Experience?
On its face, this question probably seems to be a foolish one even to pose. Quite obviously this is so because photographs indeed do look like the environments that they represent. They have face validity, short and simple. And yet the data just presented suggest that individuals do not experience photographs of the environment in quite the same way as they do either when they experience the environment first-hand or when they experience a dynamic presentation of the environment. In light of that, the question then is why choose photographs for research over either more dynamic representational formats or direct experience? The most straightforward answer – because they are so readily obtained and displayed, and most easily utilized in research – is surely true; and yet that answer masks the more substantive grounds for why they tend to be utilized. There are two seemingly sound justifications for utilizing photographs in environmental perception research: (1) photographs are straight-forward duplicates of what an environment looks like; and (2) the eye works like a camera in that it operates to capture a static image of the environment and its features. I maintain that neither of these claims is warranted. I will take up the first claim here, and will engage the second claim in the following section of the paper.
Photographs would seem to be instances of the direct transfer of structure in the environment onto a two-dimensional surface -- a matter of sheer physical optics. This view seems obviously so because “[t]he photographer does not intervene in the process of representation. The world delivers itself to the film in the form of the image of nature and the act of exposure allows the world to trace or impress itself in a purely natural way” (Snyder, 1980, pp. 504-505). In this respect, the assumption that photographs merely duplicate ‘what the eye sees’ would seem fully vindicated. It is not difficult to show that this claim reflects a misunderstanding about the nature of photography.
First, consider a commonplace occurrence that should seem quite strange upon reflection. If photographs are mere straight-forward copies, then how can it be that a photograph of a person sometimes appears to be a poor resemblance of that person? How could a mere copy of a person not look like that person? The same can said about photographs of landscapes. This is one reason why researchers nearly always select which photographs are to be included in their research; and selection indicates that the photographs employed in research are not mere impartial copies of the environment as it appears.
Compare this selective approach to the strategy proposed by Brunswik (1954) at the time when photographs were beginning to be used regularly in environmental perception research. Brunswik famously called for “representative design” in all facets of psychological research. This means that not only is it necessary to randomly sample research participants in order to have a sample that is representative of the population, but also that if we hope to understand environmental perception in a representative manner that the sites where photographs are taken needed to be randomly sampled as well. I am not proposing that we follow Brunswik’s methodology. For one thing, our goal may be to examine particular kinds of environments in research, just as we might want to include particular participant types. But still, Brunswik’s proposal brings into sharp relief the extent to which the photographs that are used in most psychological research are selected rather than representative as a matter of course.
Second, to claim that a photograph merely captures the world as it would be experienced by a perceiver positioned at the same position is to overlook the phenomenological differences between the appearance of photographs and immediate visual experience (Snyder, 1980). For one thing, visual experience is not bounded as photographs are. Even in the rare instance of stationary fixation, those features that are perceived beyond the foveated area of vision gradually diminish in clarity rather than simply stop at a rigid boundary. The periphery of vision, however temporary, is always blurred owing to the distribution of photoreceptors that make up the retina. In contrast, light sensitive film and light sensors in digital cameras are equally sensitive to light from one edge of the frame to the other (without some photographic manipulation.) In short, “[o]nly in the most exotic circumstances would one mistake a photograph for the objects photographed . . . The flatness of the photographs, their frames . . . are virtually always obvious and unmistakable . . . Photographs look like what they are: photographs. (Walton, 1984, as quoted in Carlson, 2000, p. 58).
Third, and perhaps most critically, an assertion that photography merely involves transferring patterns from the world onto a projection surface (e.g., film) overlooks the complicated history of cameras, and the many technological changes that have accompanied their development. A camera in a modern sense is quite a bit more than a pinhole camera that merely allows for the projection of light rays through an aperture. Without a proper lens, these devices are not effective for photography. Lenses for cameras are manufactured with precision and chosen for a wide variety of effects. Further a range of camera settings are built into the workings of the device, with a given level of each setting chosen either by the user or, more often, established by default in its manufacture. Such settings set parameters on exposure speed, aperture size, lighting, focal point, and more. Although we may treat a photograph as a facsimile of “what we would have seen had we been there ourselves,” Snyder and Allen (1975) aptly explain why that claim could only be accurate if it were “qualified to the point of absurdity” (p. 152). They continue (circa the 1970’s):
A photograph shows us "what would have seen" at a certain moment in time, from a certain vantage point if we kept our head immobile and closed one eye and if we saw the equivalent of a 150-mm or 24-mm lens and if we saw things in Agfacolor or in Tri-X developed in D-76 and printed on Kodabromide#3 paper. By the time all the conditions are added up, the original position has been reversed: instead of saying that the camera shows what our eyes would see, we are now positing the rather unilluminating proposition that, if our vision worked like photography, then we would see things the way a camera does (p. 152, emphasis added)
What modern cameras clearly show is that although photographs are indeed structured by patterns of light emanating from the environment, they are very much a contrivance and artifact of the technology involved. In other words, the appearance of realism that photographs offer is created rather than being a mere copy of how things appear.
With the advent of digital cameras in recent decades, any effort to describe a photograph as a mere realistic projection of light on a surface completely falls apart. In these cases, light does not cast an image but is immediately ‘broken up’ into pixels registered in the camera, and only later to be reconstituted into a viewable image.
These considerations lead to a rather startling possibility: that our evaluations of landscapes are quite likely influenced by our extensive exposure to landscape photographs. Snyder (1980) has proposed that “[o]ur judgments of similarity [between photographs and the world] are as much influenced by what we know of pictures as by what we know of nature.” (p. 502). It is this possibility that may best explain the degree to which photographs and landscapes are assessed as being similar to the degree that they are.
More generally, mistaking photographs for direct representations of the way the world appears points to a historical contention of this paper to be examined in the next section: namely, that beginning in the early Renaissance a highly intellectualized approach to visual experience began to take hold in Western Europe which assumes that visual experience to be essentially pictorial in character. The uncritical use of photographs in landscape perception research is symptomatic of that stance.
Two modes of environmental experience
Perceiver as spectator
Ever since the early 17th century when Kepler described his examination of an excised animal eye, it has been widely assumed among scientists and philosophers that visual perception begins with the projection of an image on the rear interior surface of the eye – the retina. During this same era, two other related developments emerged that further reinforced the claim that vision begins with a picture-like image: 1) the wide-spread use of the camera obscura in the visual arts and (2) the development of linear perspective.
As for the former, it had been known for quite some time prior to Kepler’s writings on the retinal image that if a small hole is made in one surface of an otherwise hollow enclosed chamber (e.g., a box) that the array of features in front of that aperture will be projected as an image on the opposite interior wall of the chamber (Crary, 1990). Devices that were intentionally built to produce these effects came to be called camera obscura. Renaissance artists such as Leonardo (1452-1519) and della Porta (1535-1615) explicitly compared the way the eye works to a camera obscura (Crary, 1990; Lindberg 1976). Over two centuries later, roughly in the 1840’s, the camera obscura was modified such that the projection of light rays could be fixed on light-sensitive media: the photographic camera was born. In reality, it is an elaboration of the Ur-design of the camera obscura; and it continues to be the case for contemporary psychology texts to portray the eye as operating like a camera.
If we begin with the claim that vision begins with a two-dimensional image projected onto the retina in the manner that the camera obscura operates, then most of the classic problems in visual perception ineluctably follow. These include the perception of three-dimensional space, the perception of distance, the perception of the relative distance of objects, and the perception of shape, qualities that are not represented on the retinal image. How then do we perceive these qualities? Theories of visual perception from Descartes’ time to the present have been shaped by attempts to solve these problems.
These same problems are among the ones that have long been faced by visual artists who attempt to represent the appearance of landscapes, public spaces, and interiors on a two-dimensional surface. A major step forward in depicting the appearance of a ‘three-dimensional world’ on a flat surface was the development of techniques of artificial linear perspective in the 16th century (as opposed to natural perspective, see Edgerton, 2009). This technique as well as the use of interposition, relative height in the field of view, knowledge of familiar size, and aerial perspective became technical tools for producing ‘realistic’ looking paintings and murals. By the 19th century, scientists such as Helmholtz began to reframe those artistic techniques as visual “cues” on the retina used by perceivers for transforming a two-dimensional retinal image into the three-dimensional mental experience. This conceptualization is currently the standard view in psychology.
Moreover, other aspects of the design of the camera obscura in addition to the projected image influenced the manner in which visual perception theory took shape. Some of these devices were small rooms that the artist could walk into; while more commonly, they were devices that the viewer could look into while standing in a stationary position or sitting. In all instances, the initial projection site of the still image was inside the camera obscura. If then vision is modeled on the camera obscura then the experience of seeing would also be conceptualized as a mechanical event that occurs inside the visual apparatus and is essentially private – that is, the experience of seeing becomes interiorized. This way of thinking about vision dovetails with Cartesian dualism. Crary (1990) writes: “First of all, the camera obscura performs an operation of individuation; that is, it necessarily defines an individual as isolated, enclosed, and autonomous within its dark confines” (pp. 38-39.) If vision is conceptualized along similar lines, then seeing too is judged to be based on ‘looking’ at a private image inside the body boundary, not at a publically shared world as such. Moreover, if we think about vision as beginning with an image captured on the retina, analogous to the image production in a camera obscura, then we are apt to consider the act of seeing in a manner that is entirely separate from the actions of the body. As Bourdieu (1997) puts it, “bodies [are] reduced to a pure gaze” (p. 22). The model of a camera obscura effectively separates the act of seeing from the body of the perceiver, just as the photographic camera sustains centuries later.
Likewise, the development of linear perspective came at a cost in terms of its capturing everyday experience. The contemporary painter David Hockney commented in an interview: “Perspective make you think of deep space on a flat surface. But the trouble with perspective is that it has no movement at all” (Joyce, 1988, p. 30). Movement has been set aside in the interest of what is in effect an illusion of depth. Hockney continues: “But in wanting to make it solid [through perspective drawing] you must stop time, that’s the problem, the core of the problem, isn’t it?” (Joyce, 1988, pp. 37-39, emphasis added). It is “a fracturing of times with a single space” (Friedberg, 2006, p. 36). Photographic cameras have the same effect by controlling exposure time and stabilizing the device. In this regard, a profound effect of the development of linear perspective, abetted by the image produced within a camera obscura, is the conceptual separation of space and time in the sciences and the arts.
These events all came together in the writings of Descartes and his successors Locke and Berkeley among many others. While it is true that the latter two philosophers differed from Descartes on many substantive matters, they accepted Descartes’ starting point for the analysis of vision which is essentially the legacy of Kepler’s anatomical discovery of the retinal image. From these beginnings, it continues to be assumed in most quarters of psychology and related fields that vision begins with a two-dimensional image projected onto the retina. Linear, mechanistic models of vision that continue to dominate psychological thought – the latest versions being information-processing and computational models arising to prominence with the development of serial processing computers --- are direct successors of this tradition. Such models are not inevitably the most suitable accounts of how vision works, but rather the expression of one line of thought. Collectively, these cultural-historical occurrences, with their manifestations in modern psychology, provide much of the justification for the often taken-granted assumption that photographs are fully adequate simulations of how the environments appear to the perceiver.
Writing in the early decades of the 20th century, John Dewey (1929) pejoratively referred to this overall stance toward environmental experience as “the spectator view.” The perceiver, as conceptualized in this Cartesian tradition, is passive in the sense of being immobile and detached from the environment. Immobility is a product of a variety of factors: single point linear perspective, the requirement of stationary position to avoid ‘blurring’ the retinal/camera image, and the long Western tradition of separating the workings of the body from the operations of the soul/mind.
Perceiver as agent
Dewey’s most immediate influences, as well as for his like-minded colleagues James and Baldwin, were the two major revolutionary events of 19th century science: evolutionary theory and field theory, the latter eventuating in the theory of relativity. Let’s begin with evolutionary theory which demonstrated that rather than limiting our efforts to understand psychological matters, including visual perception, solely with a focus on humans, there is much to be gained by thinking about vision from a phylogenetic perspective – that is, from the point of view of the historical lineage of animal species. When we do so, we are led to consider not only what makes each species unique, but also what they share – and among the things that they share (with a few exceptions) is their animacy.
Animals, nearly by definition, move around in their habitats principally to gain access to needed resources. And yet, while it is the case that all but sessile animals (e.g., anemone) engage in self-initiated movement, it is not the case that all animals are sighted beyond some minimal sensitivity to light. Species currently occupying habitats that are lacking in much illumination, such as earthworms, moles, cave-dwelling bats, and aquatic animals living at great depths, have only limited sensitivity to light. It would seem that vision is especially advantageous from the point of view of natural selection for animate organisms who occupy illuminated econiches. These considerations would seem to indicate that self-movement is the primordial and quintessential animate function; and more than likely, vision is somewhat of a late-comer over evolutionary time. This conjecture suggests that most fundamentally vision functions in the service of action.
From this perception-action perspective, which characterizes Gibson’s (1979) ecological approach to perceiving, seeing from a fixed stationary position is only one mode of visual experience, and a far less common one at that, compared to seeing from a moving point of observation. Consistent with this claim is the extensive experimental literature demonstrating that moving in the course of perceiving, or reciprocally, perceiving rigid and non-rigid features in motion, facilitates the detection of visual information and supports actions (for a selective annotated bibliography, see Heft & Richardson, 2013). Collectively, this literature conclusively demonstrates that movement enhances the recognition of objects and surface properties as well as the accuracy of perceptual judgments as compared to stationary views. On these grounds and from an evolutionary-functional standpoint, visual perception from a fixed point of observation, rather than being the ideal case, is instead a limiting condition.
As for field theory, and its apotheosis as expressed in relativity theory, the classic Newtonian distinction between space and time collapses. A central part of relativity theory is that measurements of physical properties vary in relation to the context of the measuring observer. This means that there is no single privileged location or one privileged set of observation conditions from which to measure or otherwise describe reality; but multiple equally valid locations and conditions. Reality is understood to be a space/time manifold of process and change, not a static three-dimensional Newtonian spatial universe with time an orthogonal dimension.
The upshot of these revolutions in science for the study of visual perception is that seeing is best understood to be process of perception/action – that is, a process of an agent detecting information about the environment (and the self) by engaging the world through action -- rather than adopting a passive and detached stance as a spectator.
Experiencing the Environment as an Active Perceiver
There is little doubt that photographs can typically be reasonably good approximations of what a landscape looks from a stationary observation point. I say typically because just as is the case with photographs of a person, when a landscape is shot from ‘non-conventional angles’ they may be rendered unrecognizable in relation to first-hand experience. Even more significant, photographs necessarily exclude qualities of the environment that are only perceivable from a moving point of observation or when features are in motion. This is because a number of salient visual qualities in environments are only perceivable over time – that is, they are totally absent from static displays. These qualities include:
Optical flow: when an individual moves in the environment, either actively or being transported, a flow field of optic information is generated over time. The rate of optic flow is a source of information as to how fast the individual is moving (e.g., contrast the experience of walking and biking), and the direction of movement (i.e., the visible point of radial outflow specifies the target, the eventual the point of contact). See the papers, e.g., by Lee (1976) and Warren (2006).
Motion Parallax: when moving in any locale with fixed features that are located at different distances from the path of travel, those features are perceived to move laterally at different rates as a function of their distance from the path of travel and in the direction opposite it. Closer features appear to move at a faster rate relative to the background, while farther features move more slowly.
Dynamic occlusion: when moving relative to two or more objects that are approximately in the same line of sight, the closer object gradually occludes over time at its ‘leading edge’ the more distant object; and over time the more distant occluded object is gradually revealed at the trailing edge of the occluding object. Objects that are temporarily occluded over time are experienced as still existing, but temporarily out of sight (Gibson, 1979). Optical flow, motion parallax, and dynamic occlusion are visible and wide-spread throughout an environment with features over the course of action.
Mystery: The effect of the ‘promise’ of further information along a path of travel, as in the case of information revealed around a turn (Kaplan & Kaplan, 1993; that is, new information is revealed at an occluding edge. This is a special case of dynamic occlusion.
Surprise (temporal incongruity): the pattern of landscape features experienced over time commonly has degree of predictability due to the stochastic regularity in similarity of adjacent features. Abrupt changes in the pattern of features in the course of moving in the landscape can be experienced as ‘surprising.’ Those phases can be experienced positively, as in the case of a mountain peak abruptly coming into sight during a hike, or negatively, as in the case of unexpectedly finding oneself at a cliff edge.
Depth at an edge: A change in the relative height of a surface, (e.g., a descending step) is revealed by the dynamic occlusion produced by a visible sheering of the lower surface at the edge of the current (upper) surface of support. The visual cliff experiments are the classic case (Gibson & Walk, 1960).
It is important to emphasize that typically these dynamic qualities are generated by self-motion. What this means in most case is that accompanying the production of these visual effects is the immediate experience that “I am” producing these changes. That is, in addition to generating visible changes in the world, concurrently the individual experiences some degree of self-efficacy (Chawla & Heft, 2002). To cite only one example here, in Hull and Stewarts’ (1992) comparison of on-site and photographic assessments among the same individuals across 12 sites, when raters were walking downhill they tend to rate landscape views more positively than when they were walking uphill; and most notably, sites along downhill portions of the route were rated as more beautiful than those same sites when presented as photographs.
Concluding Comment
At a minimum, it appears that the experience of viewing landscapes, cityscape, or interior spaces directly or first-hand is not equivalent to viewing those sites by means of photographic surrogates. Considering the sheer number of published studies in environmental psychology that rely exclusively on photographs, this conclusion should give researchers pause both in interpreting existing research and designing future studies. This is not to say that we cannot learn a great deal from the existing literature. Instead, we need to exercise greater caution when extrapolating those findings to everyday experience. The field of environmental perception has been remiss to date in not more systematically assessing the assumption that photographs are sufficiently adequate surrogates of first-hand experience.
References
Bechtel, R.B., & Churchman, A. (Eds.). (2002). Handbook of environmental psychology. New York: John Wiley.
Bourdieu, P. (1997). Pascalian meditations. Palo Alto, CA: Stanford University Press.
Carlson, A. (2000). Aesthetics and the environment; The appreciation of nature, art and architecture. London: Routledge.
Chawla, L., & Heft H. (2002). Children’s competence and the ecology of communities: A functional approach to the evaluation of participation. Journal of Environmental Psychology, 22, 201-216.
Clayton, S.D. (Ed.). (2012). The Oxford handbook of environmental and conservation psychology. New York: Oxford University Press.
Crary, J. (1990). Techniques of the observer: On vision and modernity in the nineteenth century. Cambridge, MA, MIT Press.
Daniel, T. C. (1990). Measuring the quality of the natural environment: A psychophysical approach. American Psychologist, 45, 633-637.
Dewey, J. (1934). Art as experience. New York: Penguin.
Edgerton, S.Y. (2009). The mirror, the window, and the telescope: How Renaissance linear perspective changed our vision of the universe. Ithaca, NY: Cornell University Press.
Friedberg, A. (2006). The virtual window: From Alberti to Microsoft. Cambridge, MA: MIT Press.
Gibson, E.J., & Walk, R. (1960). The “visual cliff.” Scientific American, 202, 64-71.
Gibson, J.J. (1979). The ecological approach to visual perception. Boston: Houghton-Mifflin.
Heft, H., & Nasar, J.L. (2000). Evaluating environmental scenes using dynamic versus static displays. Environment & Behavior, 32, 301-322.
Heft, H., & Richardson, M. (2013). Ecological psychology: An annotated bibliography. In Dana S. Dunn (Ed.), Oxford Bibliographies in Psychology. New York: Oxford University Press,
Hull, R.B., & Stewart, W.P. (1992). Validity of photo-based scenic beauty judgments. Journal of Environmental Psychology, 12, 101-114.
Joyce, P. (1988). Hockney on photography: Conversations with Paul Joyce. New York: Harmony Books.
Kaplan , R., & Kaplan, S. (1989). The experience of nature: A psychological perspective. New York: Cambridge University Press.
Lindberg. D.C. ( 1976) Vision from Al-Kindi to Kepler. Chicago: University of Chicago Press.
Lee, D.N. (1976). A theory of visual control of braking based on information about time-to-collision. Perception, 5, 437-459.
Snyder, J. (1980). Picturing vision. Critical Inquiry, 6, 499-526.
Snyder , J., & Allen, N.W. (1975). Photography, vision, and representation. Critical Inquiry, 2, 143-169.
Sontag, S. (1973). On photography. New York: Penguin Press.
Stokols, D. & Altmam, I. (Eds.) (1987). The handbook of environmental psychology. New York: John Wiley
Stamps, A.E. (1990). Use of photographs to simulate environments: A meta-analysis. Perceptual and Motor Skills, 71, 907-913.
Stamps, A.E. (2010). Use of static and dynamic media to simulate environments: A meta analysis. Perceptual and Motor Skills, 1111, 355-364.
Sun, M., Herrup, K., Shi, B., Hamano, Y., Liu, C., & Goto, S. (2018). Changes in visual intreractio0n: Viewing a Japanese garden directly, through glass, or as a projected image. Journal of Environmental Psychology, 60, 116-121.
Pitt, D.G., & Zube, E.H. (1987). Management of natural environments. In D. Stokols. & I. Altmam, I. (Eds.), The handbook of environmental psychology. Vol 2. (pp. 1009-1042). New York: John Wiley
Warren, W.H. (2006). The dynamics of perception and action. Psychological Review, 113: 359—389.
Zube, E.H., Simcox, D.E., & Law, C.S. (1987). Perceptual landscape simulations: History and prospect. Landscape Journal, 6, 62-80.