Perception and Movement - How meaning is encoded into a landscape

Sonoran Desert [Author Image]

"Because the horizon bounds the viewer inside a wall-like vertical visual barrier, it commonly takes on deep cultural significance. As the point at which the land breaks visual contact with more distant places, in traditional societies the horizon often symbolizes the boundary between known and unknown, safe and dangerous, familiar and foreign" [Bernardini et al. 2013]

Unless asleep or unconscious, human visual perception is a continuous process. Light entering the eye was once believed to project onto something like a screen, but the eye is transparent, and light activates multiple layers of sensory arrays. Image data is compressed, broken into constituent parts, and then reassembled by a massively parallel neural architecture. [Scroggins 2022] Whether eye movements, head movements, or locomotion, vision is based on motion. If motion stops, the world disappears. Perception is a neurological process, but seeing is psychological, as the individual defines meaning to what is perceived. Seeing is embodied in the individual and their experiences, yet humans are social creatures, and meaning is embedded in the social fabric called culture. With some variation, we all perceive the same thing, but what we see may be different. How can we quantify this sameness of perception, and, if possible, use that quantification to understand meaning in cultures partially erased by time?

In 1979, the psychologist J. J. Gibson published The Ecological Approach to Visual Perception. [Gibson 1979] At one level, the book was a critique of cognitive theory in psychology. [Fodor & Pylyshyn 1981] However, after 47 years, cognitive psychology still thrives, as many of Gibson's ideas have now been accepted and folded into it. For our purposes, two of Gibson's ideas are important: Ground Theory and Affordances.

Ground Theory

"There is invariably beneath him a continuous terrain, and what he discriminates is the location of all points on this terrain rather than the specific distances of given points . . . .[T]he theory behind [this] should be a theory of continuous space with an underlying terrain in which the observer is himself located and in which he can move."

[Gibson, 1947, pp. 184–185]

The landscape space of human visual perception is a hemisphere divided into ground and sky. Ground is dominant. It is perceived as a continuous surface punctuated by objects placed in relation to this surface. One sees the sky or floor, walls, and ceiling while in a room, but objects are perceived in relation to this ground surface. This is a different geometry from a Cartesian one. Ground theory and especially ground prominence have empirical validation. [Warren 2019] [Sedgwick 2021]

Our psychological perception of landscape is encoded into a ground surface. Any 3D surface can be projected onto a 2D surface. This projection has a common name. It is called a map. This might sound strange because a map is a Cartesian object, and Gibson's theories are a rejection of using Cartesian space in perceptual psychology. Cartesian space is analytic, why it is called analytic geometry. If one can transform Cartesian space into perceptual space or perceptual into Cartesian, and still keep some quantitative measurement, then a theory becomes science.

Gibson realized this and saw the eye, instead of like a camera, as being more like an optic array that registered invariants in the perceptual geometry. Gibson and others thought that vanishing points were the invariants, but we don't see just vanishing points. In a breakthrough, I believe, from H. A. Sedgwick, what we also see is a line, a horizon line.

An infinite ground plane [Sedgwick 2021, p. 33, fig. 20]

An infinite ground plane is invariant wherever one looks, not only in terms of how high a person is above the ground but also in terms of what direction one looks. This invariance and making the line at infinity a point is the basis of projective geometry. Sedgwick explains it thus:

"The 360 spherical optic array, centered on the viewpoint of the observer, is best suited for a description of the array of vanishing points and vanishing lines associated with an environmental layout of edges and surfaces. The horizon is the ground plane’s vanishing line in the optic array. It surrounds the observer and could be described as the horizontal great circle of the spherical optic array.18 The location of the horizon in the optic array is invariant; if the observer moves in any direction—up, down, forward, backward, right, or left—the location of the horizon in the observer’s optic array is unchanged (Figure 20)."

[Sedgwick 2021, pp. 32-33]

Affordances

"... affordances are to be understood as possibilities for action provided to an animal by the environment ..."

[Rietveld & Kiverstein 2014, p. 347]

Affordance is a complex idea, with many different definitions and insights dependent on the context [Rietveld & Kiverstein 2014] [Wagman & Blau 2019], but for this article, affordances are defined as:

Behaviors made possible by the perception of the landscape and the dynamic sky.

By landscape, I mean the ground, including the open-air structures we call earthworks, and the horizon where ground meets sky. Earthworks are human-made parts of the landscape, as they control what can be seen on the ground and can become part of or all of the horizon. The two affordances that I am thinking about are visibility and prominence, and the two behaviors are wayfinding and the control of time.

Visibility

"Like Gibson's, the essence of the approach here is to ask about the information available about (or better, specific to) the environment (defined as a surrounding collection of surfaces) at a point in space. This entails seeking a description of the environment that is specific to a given position or path 'through' it. An observer's perception is thus circumscribed, if not determined, by the environment-as-presented at the point of observation. A cumulative understanding of the form of the environment is arrived at by perceiving variants and invariants in the transformation of the information available caused by the observer's movement. It follows that since points of potential observation are contiguous so then is the information available spread throughout space in a field-like way. The environment is then describable in an alternative fashion: fused with space, as it were, one might speak of the visual world as a field of light-borne information in which the observer is immersed and which he samples in accordance with his intentions and curiosities."

[Benedikt 1978, p. 48]

A ground surface can be projected onto a 2D plane. Elevation can be encoded as a value associated with a picture element or pixel, which forms a grid over the plane. This DEM or Digital Elevation Model encodes the complete landscape. At some point in the landscape, what can be seen from that point forms an isovist [Benedikt 1978] or viewshed, which was first termed by Clifford Tandy in 1967 [Tandy 1967], which is some part of the 2D plane.

The two terms are interchangeable but have a different focus. Isovist comes from architecture and how an individual perceives and navigates the inside or outside of a building, or city streets. The idea is that the isovist is a direct representation of the viewer's perception, a concept in line with Gibson's. An isovist analysis produces a single complex polygon, since nothing beyond an object's initial line of sight is included. A viewshed is an analytical method in GIS that calculates whether someone at one point can see some object at another. Viewsheds measure visibility in a landscape, extending to the limits of visibility. The resulting structure can include many complex polygons. [Lllobera 2003]

Early isovist theory measured area, perimeter, occlusivity, variance, skewness, and circularity of the resulting polygon. Visibility is a dynamic process, so several vantage points are needed. Also, the resulting raster is represented by a binary number, 1 being visible, 0 not visible. This led to the idea of a cumulative viewshed, a set sampled from the whole landscape or the open parts of a building. These sets of points can be a grid, some random distribution, or points deemed important. Common points for each viewshed are added together to form a visibility landscape. This is coded as an integer, the higher values representing points with the most common views.

With a total viewshed, every point on the DEM is used. The resulting points each encode the area of the viewshed at that point. Total viewsheds can be used as a null to test against the results of cumulative viewsheds, especially if the control points constitute some sort of scientific question. [Llobera et al, 2004] Total viewsheds are computationally expensive, although there have been breakthroughs in faster algorithms. [Tabik et al. 2015] [Tabik 2016]

Isovists, viewsheds, cumulative viewsheds, and total viewsheds form a class of digital structures called visualscapes. They:

"... describe all possible ways in which the structure of visual space may be defined, broken down and represented within GIS independent of the context where it is being applied."

[Lllobera 2003, p. 26]

Space Syntax is a way to analyse the behaviours of people and the spaces they move through, particularly built structures. [Hillier & Hanson 1984] Used in city planning [Batty 2001], this results in a network of relationships between individuals and the space they inhabit. It was named Space Syntax because Hillier likened it to a language in that the relationships were hierarchical. This metaphor has been replaced by one of networks and network theory. Network theory has a rich and ever-expanding set of local and global properties that can serve as measures.

Like cumulative viewsheds, multiple isovists came under the term isovist fields or convex isovists. Again, this is related to the inside of a building or a built environment, not the whole landscape. This, however, was critiqued as still a local property of space:

"... despite their sensitiveness to the shape of spaces isovists provide essentially ‘local’ measures of configuration, whilst the lessons of space syntax research suggest that it is the global properties of spatial configuration that are important in determining the functional consequences of design."

[Turner & Penn 1999]

Turner & Penn create a graph (an older term for a network) from a set of points given two criteria:

• If two points share the same isovist, then they are connected
  

• If two points are on isovists that intersect.

  
  

"... human landscapes cannot be approached as a collection of objects, a candy box filled with features neatly separated from each other, but rather as a continuum of physical and perceptual variations. We need to evaluate the degree of meaningfulness of such variations, i.e. the likelihood that certain regularities would become recognised as specific landscape entities. It is only through such integrated approach, attentive to both the salience of topographic features and cultural practices of landscape inhabitation, that landscape archaeologists may address the semiotics of space, i.e. cultural/social strategies of geographical meaning‐making ..."

[Čučković 2025, p. 2]

Landscape chambers are a landscape archaeology approach to isovist - Space Syntax spatial analysis. This method turns a landscape into visually coherent zones and creates a way to measure the degree of coherence of each zone. The coherent zones are created by a cumulative viewshed, and the measure comes from the network created by this cumulative viewshed. The values for a cumulative viewshed is changed slightly from a frequency to a percentage based on the highest frequency being 1. This allows comparisons across regions. Below is an example of how the network delineates zones of interest:

Landscape chambers created by network measures. [Čučković 2025, p. 6, fig. 1]

This is done using the following measure:  [Čučković 2025, p. 6, fig. 1]

The links and nodes of a network can be partitioned into a set of subgroups. A modularity index is the sum of the number of links in the subgroup over the total links (L) plus the number of connections in a subgroup (d) over twice the network links (2L), this quantity squared.

This is a landscape-only measure, but is believed to represent the information human cognition takes from the landscape:

"... the purpose of the method is not to simply segment the landscape in visually distinct chambers, but rather to characterise such spaces, i.e. to measure the level of their visual distinctiveness and coherence."

 [Čučković 2025, p. 20]

Notice the series of transformations that take place. The human 'view' of a 3D landscape, that of sky, horizon, and ground, is mapped onto a set of 2D visualscapes, maps in that they encode both spatial and attribute information. This essentially removes the third dimension. A second transformation turns some part of these visualscapes into a network. This removes all spatial information; it is no longer a Cartesian geometry, as Gibson claimed visual perception to be.

Prominence

Another approach to landscape theory is phenomenology, exemplified by the anthropologist Christopher Tilley's 1994 book A Phenomenology of Landscape Places, Paths and Monuments. [Tilley 1994] He argues for a humanistic view of space. Human space has context; it is subjective, relational, empowered, and temporal.

"A centred and meaningful space involves specific sets of linkages between the physical space of the non-humanly created world, somatic states of the body, the mental space of cognition and representation and the space of movement, encounter and interaction between persons and between persons and the human and non-human environment."

[Tilley 1994, p. 10]

Once a space has been given cultural meaning, it becomes a place. This place is formalized through narrative, song, ritual, or dance. Thus, knowledge of this meaning is passed through generations. [Basso 1996]

Topographic Prominence [Llobera 2001, p. 1007, fig. 1]

The image above shows the idea of topological prominence, developed by the British archaeologist Llobera. [Llobera 2001] This is a local measure of what he deems as an affordance in the landscape. He defines it thus:

"Topographic Prominence is here described as a function of height differential between an individual and his/her surroundings as apprehended from the individual’s point of view.

More precisely, it is defined as the percentage of locations that lie below the individual’s location (terrain altitude plus individual’s height) within a certain radius."

[Llobera 2001, p. 1007]

Notice in the image above that the topological prominence of the person is high over a small distance because they are standing on a small hill, but drops drastically once the distance reaches a higher cliff. So the scale factor is important. I was once on a tour of Kotzebue, Alaska, an Inuit village north of the Arctic Circle. At a final ceremony, they brought out a huge round walrus skin; everyone grabbed a section, and we tossed a baby up in the air. Why? The Arctic is mostly treeless, and out on the sea ice, it is flat like a table. Tossing a person up in the air provides an expanded view, allowing one to spot game or see features for orientation that are not possible at human height. Height is an affordance that the environment cannot provide, so the behavior is to somehow increase height.

Notice again that topological prominence is not a visual measure but a measure of the possibility of visibility at a given distance.

Wesley Bernardini was originally at the University of Redlands in Redlands, California, where the commercial GIS giant ESRI has its headquarters. He now works for the Hopi Nation in northern Arizona. The Hopi are a unique indigenous culture in that they have occupied their mesas continuously for three thousand years and have avoided the conquests and expulsions that resulted from contact. This makes them a much-studied group; hundreds of books and papers have been written about them. Through his contact with the Hopi and his study of their ethnography, he is researching an important factor of their landscape, the horizon or skyline.

A skyline [Bernardini 2013, p. 3948, fig. 1]

A skyline is actually a circle; each point on the circle represents an elevation. The actual radius of this circle can vary widely, as seen in the image above, yet it is perceived as constant. Bernardini gives various numbers for the limit of visual range. For wet, steamy Ohio, I would think the lower range of 12 km (7.46 miles) to 30 km (18.64 miles) should work. Records of airport visibility for central Ohio from the 1970s on show maximums from around 11 to 13 miles. Bernardini sees his method of prominence as a local analysis and Llobera's as global since it creates a surface. I suppose this is true, although the surface reflects a local scale. Bernardini uses a point-reduction algorithm that reduces the number of points in the skyline while preserving shape information, primarily the maximin slopes and heights. When his algorithm is applied iteratively, it creates a hierarchy of peaks. Thus, the skyline is reduced to a small set of prominent points, quantified by importance.

Iterative Point Reduction Algorithm [Bernardini 2013, p. 3950, fig. 5]

This is just one set of important points on the skyline. As seen below, the motions of the sun, moon, and stars add importance to the skyline, not necessarily connected to prominent features.

Hopi Sunwatcher with Horizon Calendar, Walpi, AZ  [Williamson 2015a, p. 644, fig. 45.1]

Approach

It can be seen that several quantitative approaches to landscape have been derived from theories based on human perception and the emphasis on a first-person view, phenomenology. Some of these approaches may provide insight into the geometric earthworks of Ohio. Not only are Digital Elevation Models of different elevations needed, but a fairly accurate location and at least a Potemkin 3D model of the earthwork are needed. Squier & Davis' maps are not the best, but sometimes all that we have. Total or continuous viewsheds are computationally expensive, so extensive experimentation is needed to find an optimal workflow. The criteria are to leave behind not only some insights, but also useful digital artifacts.

Conclusion

Earthworks differ from structures such as buildings in that they are open to the sky. The geometrical earthworks in Ohio can have additional major structures in ditches and secondary inner structures, such as mounds, buildings, fire pits, and circles and lines of posts. From a landscape perspective, I think the most appropriate perspective is that of a Japanese walled garden. From within a wall, all that is not hidden by it is considered a borrowed landscape; the builders define what is and is not to be seen. [Higuchi 1988] Walled gardens are built in accordance with the aesthetics of Japanese culture, including what is borrowed from the background. The original height of each earthwork is unclear or unknown. Sometimes it varies to present a constant height to the eye. What is borrowed from the landscape by Ohio earthworks is not clear, but possibilities can be tested experimentally.

The researchers Ray Hively and Robert Horn have postulated that the great earthworks at Newark, Ohio, and the relationship of the Observatory Circle and the Octagon to the long-term motions of the moon, are the result of the unique positioning of the hills and sightlines of the region. [Hively & Horn 2013] Sister sites like the octagon and circle at High Bank in Chillicothe, Ohio, represent another attempt. [Hively & Horn 1984] Earthworks have similar features but can also be very different. A question remains on to why this diversity. Is some of it a property of the landscape they were built into?

---

• Basso, Keith H. 1996. Wisdom Sits in Places: Landscape and Language Among the Western Apache. The University of New Mexico Press.
  

• Batty, Michael. 2001. "Exploring Isovist Fields: Space and Shape in Architectural and Urban Morphology." Environment and Planning B: Planning and Design 28: 123–50.
  

• Benedikt, M. L. 1979. "To Take Hold of Space: Isovists and Isovist Fields." Environment and Planning B 6: 47–65.
  

• Bernardini, Wesley, Alicia Barnash, Mark Kumler, and Martin Wong. 2013. "Quantifying Visual Prominence in Social Landscapes." Journal of Archaeological Science 40: 3946–54.
  

• Brughmans, Tom, and Ulrik Brandes. 2017. "Visibility Network Patterns and Methods for Studying Visual Relational Phenomena in Archeology." Frontiers in Digital Humanities 4: 17.
  

• Čučković, Zoran. 2025. "Landscape Chambers: Towards an Archaeology of the Cognitive Landscape." Journal of Archaeological Method and Theory 32: 20.
  

• Davis, Larry S., and Michael L. Benedikt. 1979. "Computational Models of Space: Isovists and Isovist Fields." Computer Graphics and Image Processing 11: 49–72.
  

• Fodor, J. A., and Z. W. Pylyshyn. 1981. “How Direct Is Visual Perception?: Some Reflections on Gibson’s ‘Ecological Approachment.’” Cognition 9.
  

• Gibson, James J. 1947. Motion Picture Testing and Research. Aviation Psychology Research No. 7. U.S. Government Printing Office.
  

• Gibson, James J. 1979. The Ecological Approach to Visual Perception. Classic Edition. New York: Psychology Press.
  

• Higuchi, Tadahiko. 1983. The Visual and Spatial Structure of Landscapes. Translated by Charles S. Terry. MIT Press. https://www.aaeportal.com/publications/-16537/the-visual-and-spatial-structure-of-landscapes.
  

• Hillier, Bill, and Julienne Hanson. 1984. The Social Logic of Space. Cambridge University Press. https://doi.org/10.1017/CBO9780511597237.
  

• Hively, Ray, and Robert Horn. 1984. “Hopewellian Geometry and Astronomy at High Bank.” Journal for the History of Astronomy 15 (7): S85–100. https://doi.org/10.1177/002182868401500701.
  

• Hively, Ray, and Robert Horn. 2013. “A New and Extended Case for Lunar (and Solar) Astronomy at the Newark Earthworks.” Midcontinental Journal of Archaeology 38 (1): 83–118.
  

• Llobera, Marcos. 2001. "Building Past Landscape Perception with GIS: Understanding Topographic Prominence." Journal of Archaeological Science 28: 1005–14.
  

• Llobera, Marcos. 2003. “Extending GIS-Based Visual Analysis: The Concept of Visualscapes.” International Journal of Geographical Information Science 17 (1): 25–48. https://doi.org/10.1080/713811741.
  

• Llobera, Marcos, David Wheatley, James Steele, Simon Cox, and Oz Parchment. 2004. “Calculating the Inherent Visual Structure of a Landscape (‘Total Viewshed’) Using High- Throughput Computing.” XXXII International Conference - Computer Applications in Archaeology 2004.
  

• Osiurak, François, Yves Rossetti, and Arnaud Badets. 2017. “What Is an Affordance? 40 Years Later.” Neuroscience & Biobehavioral Reviews 77 (June): 403–17. https://doi.org/10.1016/j.neubiorev.2017.04.014.
  

• Rietveld, Erik, and Julian Kiverstein. 2014. “A Rich Landscape of Affordances.” Ecological Psychology 26 (4): 325–52. https://doi.org/10.1080/10407413.2014.958035.
  

• Scroggins, Michael. 2022. “The Eye Is Not a Camera.” Slide. https://michaelscroggins.wordpress.com/explorations-in-stereoscopic-imaging/presentation-archive/.
  

• Sedgwick, H. A. 2021. “J. J. Gibson’s ‘Ground Theory of Space Perception.’” I-Perception 12 (3): 20416695211021111. https://doi.org/10.1177/20416695211021111.
  

• Tabik, Siham, Antonio R. Cervilla, Emilio Zapata, and Luis F. Romero. 2015. “Efficient Data Structure and Highly Scalable Algorithm for Total-Viewshed Computation.” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 8 (1): 304–10. https://doi.org/10.1109/JSTARS.2014.2326252.
  

• Tabik, Siham. 2016. “Total Viewshed.” GitHub. https://github.com/SihamTabik/Total_Viewshed/blob/master/readmeV2.txt.
  

• Tandy, CRV. 1967. “The Isovist Method of Landscape Survey.” Methods of Landscape Analysis.
  

• Tilley, Christopher. 1994. A Phenomenology of Landscape: Places, Paths and Monuments. Oxford: Berg.
  

• Turner, Alasdair, and Alan Penn. 1999. "Making Isovists Syntactic: Isovist Integration Analysis." Paper presented at the 2nd International Symposium on Space Syntax, Brasilia.
  

• Turner, Alasdar. 2026. Space Syntax. https://www.spacesyntax.online/#video-group-1.
  

• Wagman, Jeffrey B., and Julia J. C. Blau, eds. 2020. Perception as Information Detection: Reflections on Gibson’s Ecological Approach to Visual Perception. New York: Routledge.
  

• Warren, William H. 2019. “Perceiving Surface Layout: Ground Theory, Affordances, and the Objects of Perception.” In Perception as Information Detection. Routledge.
  

• Weitkamp, Gerd, Ron van Lammeren, and Arnold Bregt. 2014. "Validation of Isovist Variables as Predictors of Perceived Landscape Openness." Landscape and Urban Planning 125: 140–45.
  

• Wiener, Jan M., Gerald Franz, Nicole Rossmanith, Andreas Reichelt, Hanspeter A. Mallot, and Heinrich H. Bülthoff. 2007. "Isovist Analysis Captures Properties of Space Relevant for Locomotion and Experience." Perception 36: 1066–83.