ArticlePDF Available

Regional Ways of Seeing: A Big-Data Approach for Measuring Ancient Visualscapes


Abstract and Figures

Archaeologists have long acknowledged the significance of mountains in siting Greek cult. Mountains were where the gods preferred to make contact and there people constructed sanctuaries to inspire intervention. Greece is a land full of mountains, but we lack insight on the ancient Greeks’ view—what visible and topographic characteristics made particular mountains ideal places for worship over others, and whether worshiper preferences ever changed. This article describes a data collection and analysis methodology for landscapes where visualscape was a significant factor in situating culturally significant activities. Using a big-data approach, four geospatial analyses are applied to every cultic place in the Peloponnesian regions of the Argolid and Messenia, spanning 2800–146 BC. The fully described methodology combines a number of experiences—looking out, looking toward, and climbing up—and measures how these change through time. The result is an active historic model of Greek religious landscape, describing how individuals moved, saw, and integrated the built and natural world in different ways. Applied elsewhere, and even on nonreligious locales, this is a replicable mode for treating the natural landscape as an artifact of human decision: as a space impacting the siting of meaningful locales through history.
Content may be subject to copyright.
Regional Ways of Seeing: A Big-Data
Approach for Measuring Ancient Visualscapes
Natalie M. Susmann
Archaeologists have long acknowledged the signicance of mountains in siting Greek cult. Mountains were where the gods preferred to
make contact and there people constructed sanctuaries to inspire intervention. Greece is a land full of mountains, but we lack insight on the
ancient Greeksviewwhat visible and topographic characteristics made particular mountains ideal places for worship over others, and
whether worshiper preferences ever changed. This article describes a data collection and analysis methodology for landscapes where
visualscape was a signicant factor in situating culturally signicant activities. Using a big-data approach, four geospatial analyses are
applied to every cultic place in the Peloponnesian regions of the Argolid and Messenia, spanning 2800146 BC. The fully described
methodology combines a number of experienceslooking out, looking toward, and climbing upand measures how these change
through time. The result is an active historic model of Greek religious landscape, describing how individuals moved, saw, and integrated the
built and natural world in different ways. Applied elsewhere, and even on nonreligious locales, this is a replicable mode for treating the
natural landscape as an artifact of human decision: as a space impacting the siting of meaningful locales through history.
Keywords: visualscape, viewshed, prominence, phenomenology, big data, Greece, mountain, vision, open access
Alrededor del mundo, las montañas provocan la creación de lugares religiosos en el paisaje como resultado de sus efectos visuales; los
adoradores se inspiran en las cuencas de las montañas y su prominencia física. Los arqueólogos debaten cómo mejor modelar este pro-
ceso: si a través de enfoques geoespaciales o fenomenológicos, o la integración de los dos. Las metodologías detalladas que involucran
múltiples enfoques y grandes conjuntos de datos son raras. Este documento proporciona un marco detallado para integrar análisis
geoespaciales precisos con la recopilación de datos fenomenológicos, con el n de determinar qué características topográcas y visuales
conguraron los paisajes religiosos. Presenta un estudio de caso sobre las regiones del Peloponeso de Argólida y Mesenia, y explora
lugares de culto (es decir, religiosos) utilizados por los griegos entre el 2,800 y 146 aC. Esta metodología, completamente descrita en esta
ponencia, combina varias experiencias (observación general, observación hacia el objeto, y la escalada en sí) y compara las preferencias
cambiantes a lo largo del tiempo. El resultado es un modelo dinámico del paisaje religioso griego, que articula la movilidad de los antiguos
griegos, y de cómo estos vieron e integraron de manera diferente el mundo construido y natural a lo largo del tiempo. Aplicado en otros
lugares, e incluso en lugares no religiosos, este es un modo replicable para tratar el paisaje natural como un artefacto de la decisión
humana: como un espacio que impacta la ubicación de lugares signicativos a lo largo de la historia.
Palabras clave: paisaje visual, cuenca visual, prominencia, fenomenología, grande datos, Grecia, montaña, vision, accesso abierto
Humans are affected by the physical landscape: the physical
experience of travel and the sights they encounter along the
way (Ashmore and Knapp 1999). These encounters provoke the
process of placemakingwhere individuals choose to pause
and derive meaning (Ashmore 2014; Tuan 1977:179). Challenging
terrain, prominence, panoramic views, lines of sight, and even
invisibility can therefore provoke the founding of culturally signif-
icant places of various functions: power, reverence, defense,
and religious signicance (Bradley 2000; Eliade 1961;
Llobera 2007:5153; Scully 1962,1991; Turner 1979; Van Dyke et al.
This article explores how to best measure the impact of moun-
tainous landscapes on the siting of religious places. The spatial
relationship will differ depending on the mountains religious
function (Bernardini et al. 2013; Boivin 2004;Grove1999; Reinhard
and Ceruti 2010; Williams and Nash 2006). Humans can sacralize
mountains, as happens with Hopi and Navajo tribes in the
southwestern United States (Bernardini et al. 2013:3951; Jett
1995:41). In the case of the Hopi, their low-lying villages main-
tained direct lines of sight with whichever mountains were sacred
at that time (Bernardini et al. 2013). In the Andes, complexes were
Advances in Archaeological Practice, 2020, pp. 118
Copyright 2020 © Society for American Archaeology
constructed on mountain summits, as well as oriented toward
higher peaks, and subsequently considered symbols of religious
and political power (Williams and Nash 2006). Reverence for
mountains could manifest even where these landforms were
unavailable. In Mesopotamia, for example, monumental ziggurats
acted as articial barriers; their height reinforced a division
between common people on earth and their physically elevated,
deied king (Eliade 1961:42; Leick 2002; Scully 1991:2324).
These examples model the attachment formed between humans
and what they seethe visualscape(Llobera 2003; Richards-
Rissetto 2017:4)and archaeologists have long sought precise,
analytical methodologies for its characterization.
There is much
to consider: the physical and visual effects that observers become
culturally attuned to noticing and with which they associate
meaning, where observers stand during that process, and how
shifting cultural practices might alter visual experience.
Archaeologists studying visualscapes largely rely on one of two
approaches: digital analyses using geospatial and 3D modeling
software or phenomenology. Used separately, either avenue will
pose methodological challenges. Some scholars criticize digital
landscape models for presenting a Cartesian or Gods-eye per-
spective (Brück 2005:54; Thomas 1993:25, 2004:198201;
Wickstead 2009:250). Other scholars contend that phenomen-
ology will inaccurately substitute an archaeologists experiences
for the ancient worshipers (De Reu et al. 2011:3435). Many studies
have proposed offsetting the objectivity of digital landscape
models with the subjectivity afforded by phenomenology (Eve
2012; Fábrega-Álvarez and Parcero-Oubiña 2019; Fitzjohn 2007;
Given 2004; Rennell 2012; Seamon 2014; Thomas 1993,2008;Van
Dyke et al. 2016), but full methodological disclosure is rare.
We need a precise, methodical approach for determining how
visualscapes affected the siting of meaningful places and whether
these preferences ebbed and owed over time. By combining
archaeological and historical data and phenomenological obser-
vation, as well as precise geospatial analyses, this article provides
exactly that: a clearly articulated framework for gauging the
cultural salience of natural landscapes.
I designed this methodology with four goals in mind:
(1) Data collection and analyses should rely on phenomeno-
logical observations as well as geospatial technology in order
to offset weaknesses in either approach.
(2) Data collection and analyses should gauge characteristics and
perspectives that are culturally relevant to the study area.
(3) Data collection and analyses should account for contrasting
observer experiences: looking toward as well as from signi-
cant places.
(4) A signicant span of time should be covered to compare
worshiper trends against major cultural and political shifts.
The methodology is applied toward a dataset of cultic (i.e., reli-
gious) places in the Peloponnesian regions of the Argolid and
Messenia circa 2800146 BC. The resultant model identies spe-
cic periods when topographical and visual preferences change. It
articulates how the ancient Greeks saw and utilized their religious
landscape differently through their history. More broadly, this
article provides an open-access framework for gauging how
visualscapes impact placemaking processes. It can be applied to
any period, region, or place typereligious or otherwise.
There is no question that natural features inuenced the siting of
cult in the ancient Greek world. Modern scholarship largely draws
connections between certain deities and natural settings (e.g.,
Poseidon and the sea), but every Greek deity was, in fact, con-
nected in some way to natural phenomena (Cole 2000:472474).
The natural world was a space for divine intervention and godly
benets: rest, regeneration and healing, or divine intervention
(Håland 2009:119120; Larson 2010:10; Williams 1991:109).
Mountains were considered spaces of transition between the
immortal and mortal worlds. They were not gateways or entrances,
as with cultures such as the Maya (Lucero and Kinkella 2014:1314),
but intervening places, providing deities with easy access to the
mortal world (Langdon 2000:461). As a result, mortal Greeks were
somewhat fearful of mountains. The mountains were frequently
visited by the gods, which meant that a mortals everyday
activitiestraveling, hunting, or herdingcould potentially go
awry (Buxton 1992:79, 1994:8586).
Signicant agency was imparted to ancient Greek gods; appearing
on a mountain was always their choice. That said, mortals who
sought divine intervention would try to entice them. Consequently,
the ancient Greeks have left us with a series of archaeologically
detectable actions for understanding the mountainworshiper
relationship. Worshipers selected a mountain, dened a temenos
(i.e., a sacred boundary), kept that space clean, and there regularly
deposited votives and made sacrices (Cole 2004:3536).
Greek archaeologists have long focused on situating ancient
accounts on the tangible landscape. A signicant body of schol-
arship focuses on pinpointing where on the natural landscape
Greek worshipers preferred to practice (Barnett 2007; de Boer and
Hale 2008; Driessen 2003; Faro 2008; Goodison 2004; Hannah
2013; Hitchcock 2007; Retallack 2008; Scully 1962; Van de Moortel
2006,2011; Williamson 1993). Much of this work focuses on
mountain top worship. Geospatial analyses measure the wor-
shipers viewshed while standing inside the temenos (Briault 2007;
Hannah 2013; Nixon 2009; Peateld 1983; Soetens 2009; Soetens
et al. 2001; Soetens, Sarris, Topouzi et al. 2002; Soetens, Sarris,
Vansteenhuyse et al. 2002; Topouzi et al. 2000; Williamson 1993).
These studies often measure whether from there, observers could
see another culturally signicant place (e.g., a sanctuary; Doxtater
2009; Nixon 2009).
In order to effectively model an ancient visualscape, archaeolo-
gists need to scrutinize whether their chosen analytical
approaches have modern and discipline-specic biases.
Archaeologists exploring sacred landscapes in Greece have his-
torically favored sanctuaries located on the highest, steepest
peaks. Such studies are arguably impacted by modern topo-
graphical denitions, which, although largely lacking standard-
ization, do tend to emphasize elevation and slope as dening
factors (Blyth et al. 2002:12). The ancient Greeks were certainly
attuned to elevation and slope, but they also cared about prom-
inencethat is, noticeability (Belis 2015:41; Buxton 1992:2;
Natalie M. Susmann
2Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
Langdon 2000:462). A landform does not need either a high ele-
vation or a steep slope in order to be noticeable, and this might
explain why some landforms in Greecebeing well below the
300 m markare described as mountains in ancient texts (Belis
2015:40; Buxton 1992:2; Langdon 2000:462).
Previous archaeological studies suggest that views afforded from
mountains and hills could have impacted the founding of cultic
places on the ancient Greek landscapebut so too could the views
toward them. By comparing these contrasting observer perspec-
tives within a single dataset, we can determine the physical char-
acteristics that inspired the cultic placemaking process on the
ancient Greek landscape. My study asks: Were Greek cultic places
chosen according to landscape prominence (i.e., noticeability) or
visible extent? Did these preferences vary regionally and over time?
Dening the Study Area
This study is devoted to two Peloponnesian regions: the Argolid
(2,819.38 km
) and Messenia (4,060.21 km
;Figure 1). The Argolid
is in the northeast Peloponnese, bound on the eastern side by the
Arcadian mountain range. Messenia encompasses the modern
towns of Triphylia, Pylia, Messini, and Kalamata, with the Taygetos
Mountains on the east, and Mounts Nomia and Elaeum to the
north. Adhering to strict boundary lines such as these does pre-
sent a conundrum for landscape archaeologists (Cherry 1983:385
386). On the one hand, it provides a manageable space within
which to collect data; on the other hand, these lines inaccurately
presume that people adhere to a singular notion of boundary.
Questions about vision can further complicate archaeological
boundaries. In the observers mind, the physical space of their
political or social territory can be different based on their visible
extent, impacting their sense of spatial ownership and belonging
(Bernardini et al. 2013:3948). For research on visualscape, this
means considering expanding a study area beyond sociopolitical
boundaries. In this study, for example, the boundaries of the
Argolid and Messenia were extended because observers could
have easily seenand, in turn, utilizedsanctuaries beyond those
lines (e.g., at Bassae and Olympia, as well Skilloundia [Cooper
1996:42, 45; Larson 2007:27; Morgan 1990:8990]). The boundaries
of either regionand the cultural identities contained within
themalso change through history. The case study applies the
largest known ancient boundaries for either region in order to
understand whether and how historic locales were integrated into
the lives of contemporary people.
FIGURE 1. The survey recorded all cultic places used between 2800146 BC in the Argolid (2,819.38 km
) and Messenia
(4,060.21 km
Regional Ways of Seeing
2020 |Advances in Archaeological Practice | A Journal of the Society for American Archaeology 3
FIGURE 2. A cascading typology was applied to all constructed and natural places, allowing for each place to be both broadly and
specically dened.
Natalie M. Susmann
4Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
Survey Targets
A cultic place is dened as any individual built or natural space to
which material remains for repeated religious activity can be attrib-
uted. These activities can be unique to a particular time or region
(Antonaccio 1994:398).
For each of the study areas, I located all
places where previous researchers had discovered material evidence
for cultic activity (e.g., burning res, intentional deposits of objects,
sacrice, or construction of related buildings). Because the study
focuses on placemaking as it pertained to group worship, house-
hold cultic contexts were excluded. Restrictions set by the Greek
Ministry of Culture and Sports meant only published cases could be
used. In sum, 401 places were identied.
The data-recording strategy captures every individual act of cult.
Instead of representing an entire sanctuary as one cultic place
(one database record), every locale inside its boundary dedicated
to cultic acts (altar, temple, pit, etc.) is treated as a discrete place
as well. Recording data in this way makes it easier to identify
changing temenos sizes and functions over time and to measure
how these alterations impacted the worshipers visualscape.
Cultic places located in other environments (e.g., coasts, valleys,
plains) are not excluded from the study. The reasons for doing
so are threefold. First, as previously described, there are sig-
nicant differences between the ancient and modern Greek
denition of mountain.Analyzing all cultic placesregardless
of topographical situationaccounts for the archaeologist
potentially seeing the landscape differently. Second, analyzing
all cultic places encourages inquiries about how the ancient
Greeks saw and used their landscape differently across space
and through time. Despite the clear importance of mountains in
ancient Greek worship, there were still many sanctuaries
founded on low-lying, seemingly unnoticeable locations. Many
potential inuencers could have come into play: social compe-
tition and settlement proximity (de Polignac 1994,1995)and
certain cults requiring specialized natural environments (e.g.,
Poseidon worship occurring near water), as well as changing
worshiper preferences.
With all these factors in mind, there is a substantial body of evi-
dence pointing to the importance of mountainous terrain, and by
exploring all cases of worshipregardless of their natural envir-
onmentwe can cross-examine many inuencers and determine
whether the cultic placemaking process focused on meeting
multiple needs. Finally, recording all cultic places responsibly
produces a more expansive dataset, leaving room for future ana-
lytical improvements and research enquiries.
FIGURE 3. Visual data were collected both within cultic place boundaries and around them. For example, data collection (pink
points) occurred at various distances around the Sanctuary of Asklepios at Epidaurus and the Sanctuary of Apollo Maleatas in
order to understand the surrounding visualscape.
Regional Ways of Seeing
2020 |Advances in Archaeological Practice | A Journal of the Society for American Archaeology 5
Data Recording
I assigned a unique Place IDto each cultic place (every built
altar, open space altar, megaron, temple, tomb, terrace, etc.) and
stored them in an Access database (Supplemental Database;
Supplemental Text). Because this study treats the natural envir-
onment as an artifact, I recorded each cultic places associated
natural feature as a separate entry in the database as well. Each
natural and constructed places type was categorized via a cas-
cading typology: every place was dened broadly (e.g., con-
structed) as well as specically (e.g., open-air altar), thereby
ensuring that the data is exible and functional, even for analyses
beyond the scope of this work (Figure 2).
The database is designed to link places to one another. Linking
relationships are functional (e.g., a temple is located inside of a
sanctuary) or spatial (e.g., the Sanctuary of Apollo at Bassae is
visible from the Sanctuary of Nomian Pan).
Additionally, every
cultic place is linked to its natural feature through a spatial rela-
tionship (e.g., the Sanctuary of Apollo at Bassae is located on
Kotylion Mountain).
I recorded additional attributes for every constructed place,
such as ancient and modern bibliography, summary of
excavation and survey data, and previous theories regarding
motives for its placement. The database includes a timeline
table that records the following phases (if any) for every place:
(1) cultic use prior to construction of a permanent structure, (2)
construction of the cultic place, and (3) cultic reuse following
abandonment, thereby indicating a change in practice (e.g., a
Mycenaean tholos tomb being converted into an altar during
the Geometric period).
Field Survey
The goals of the eld survey were twofold. First, substantial visual
data were needed to ground-truth the visualscape analyses.
Second, I wanted to enact a phenomenological investigation of
these cultic places and experience the environment rsthand
(Susmann 2019b:111, 121139).
Observing ourselves and
others in the modern landscape opens our minds to the potential
pathways, views, and perceptions of past users (Barrett and Ko
2009:276; Given 2004:168; Tilley 1994:7375; Van Dyke et al.
2016). Phenomenology therefore helps us build a place-based
narrative of the [worshipers] journey and to explore (through
the analogue of our own bodies) the kind of perceptionsand
emotionsattendant on moving through such spaces
(Wescoat 2017:69).
FIGURE 4. Visual data were collected inside the cultic places boundaries (i.e., here, at the Sanctuary of Apollo Maleatas).
Locations of data collection (pink points) were determined by place type (e.g., temple, open-air altar, pit, etc.), and they model
where ancient observers would feasibly stand. Adjustments were occasionally made because of permit restrictions.
Natalie M. Susmann
6Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
The survey was accomplished in two phases. For each study area, I
spent approximately two months locating these places remotely in
Google Earth; survey and excavation publications often exclude or
generalize geographic coordinates. Next, I traveled to Greece,
located each cultic place in person, and recorded data for each.
Visual data captured what could be seen, based on multiple
vantage points. I photographed panoramas at varying distances
and directions, capturing views looking toward the cultic place,
and at the cultic place, looking at the surrounding landscape. I
recorded videos to convey the experience of walking through and
around each places boundaries. Then, I marked every video and
photograph with a GPS point (Figures 3 and 4). Video narratives
and journals recorded my experience of the place: difculty of
access, extent and contents of the viewshed, level of isolation
(e.g., if other visitors were present), and my impression of how
others utilized the place.
The Greeks regarded mountains as sources for godly contact, and
I hypothesize that this association inuenced where cultic activity
was located on the landscape. Historical and archaeological data
suggest that the Greeks identied mountains based on notice-
ability and/or vast viewsheds. I decided to test how frequently the
survey data reected these preferences: whether most places were
situated to be noticeable or afford worshipers panoramic views of
their landscape. By ordering the quantitative results in chrono-
logical orderwhen cultic activity occurred at each placeI
modeled how worshiper preferences changed over time.
The selected analyses fall into three categories (Table 1):
(1) Viewshed, or what can be seen from the cultic place. This was
measured via CVSL (Cumulative Viewshed of Location).
(2) Visual prominence, or whether the place was noticeable to
passers by based on characteristics like elevation, slope, and
shape against the horizon. This was measured via Cumulative
Viewshed of Observer (CVSO and CVSO 5K) and Visual
Prominence of Observer (VPO).
(3) Topographic prominence, or whether the place was notice-
able as a result of rugged terrain. This was measured via
Topographic Prominence of Location (TPL).
This section briey describes the factors that informed the design
of these analyses: representing place targets via polygons and not
centroids, number, and placement of observer points, as well as
maximum visible distance radii.
Boundaries and Observers. For the sake of easy database
recording, I represented each cultic places location as a single set
of GPS coordinates (Places.shp). I created a second shapele
(PlaceBounds.shp) to represent cultic placesuse areas. This shape-
le was utilized in the visualscape analyses. If a single cultic
place had multiple extents over time, I recorded each phase as a
separate polygon and linked it to respective records in the Access
databases timeline (Figure 5).
The analyses measure a range of visual effects, the majority of
which are founded on viewshed principles: measuring what an
observer can see, based on the individuals stature, the elevation,
and the elevations of the surrounding landscape. Each analysis
necessitated a customized set of observer points, and
PlaceBounds.shp dictated how these were placed (Tables 1 and 2).
For CVSL, which measured a places viewshed, I situated observer
TABLE 1. Visualscape Analyses Summary.
Method Description
Height (m)
Distance (m)
Size (pixels) Required Dataset
(visible extent)
CVSL (Cumulative
Viewshed of
Cumulative visible extent
from the cultic place
1.65 DEM,
Visual Prominence
CVSO (Cumulative
Viewshed of
Cumulative visual
prominence of each cultic
place, based on # of times
1.65 DEM,
Viewshed of
Observer 5K)
Cumulative visual
prominence of each cultic
place, based on # of times
seen within a 5 km radius
of observers
1.65 5,000 DEM,
VPO (Visual
Prominence of
Visual prominence of each
cultic place based on its
landforms shape against
the horizon
1.65 DEM,
TPL (Topographic
Prominence of
Topographic ruggedness of
each cultic place, based on
ruggedness and steepness
3 × 3 DEM,
Regional Ways of Seeing
2020 |Advances in Archaeological Practice | A Journal of the Society for American Archaeology 7
points (Observer_CVSL.shp) with respect to these spatial bound-
aries. For the three visual prominence analyses (CVSO, CVSO 5K,
and VPO), I placed random observers across the study area
(Observer_CVSO.shp and Observer_VPO.shp). In these cases,
PlaceBounds.shp demarcated ineligible placement areas: obser-
vers were not permitted to standon top of structures.
Maximum Visible Distance. Vision-focused geospatial tools
include an optional outer visible radius setting. If the radius is
applied, the observers ability to see is limited by that distance.
Archaeologists tend to favor this setting, and they apply radii
somewhere between 3,440 and 5,000 m in order to represent the
limitations of 20/20 human vision (Gillings 2015:4). Applying this
radius does not guarantee output accuracy. Many potential
impediments existsuch as vegetation, light, weather, and poor
visionbut archaeologists cannot easily account for them. This is
where visual eld data comes in handy.
My methodology does not apply maximum visible distance radii
(Table 1) for two reasons. First, increased elevation extends an
individuals vision well past the commonly applied 5,000 m
threshold, and this was observed with most places recorded in this
study. Second, 3,4405,000 m distance thresholds are based on an
individuals ability to see a 1 m wide object (Ogburn 2006:409
410). Recognition of a signicant locale is based on not only the
sight of the structure but also the landform. Landforms measured
in this study are certainly wider than 1 m.
Geospatial Approaches
Visual Prominence. Visual prominence, or a places noticeability,
was measured in two different ways (Table 1). Cumulative
Viewshed of Observer (CVSO) assesses how often an individual
would see a cultic place. This was calculated via the summation of
several thousand viewsheds across the study area. Visual
Prominence of Observer (VPO) focuses on horizon lineswhich
landforms were the most distinctive based on shape.
Cumulative Viewshed of Observer (CVSO). Cumulative Viewshed
of the Observer (CVSO and CVSO 5K) gauges noticeability based
FIGURE 5. Spatial extent affected how analytical outputs were measured. If a places extent changed over time (green polygons),
I made sets of observers per phase (black points) and then ran analyses per each observer point set. Note that the observer points
shown here are only for the Argive Heraion sanctuarys temenos. I recorded cultic structures located inside the temenos (i.e., the
altar, Old Temple Terrace, Old Temple of Hera, and New Temple of Hera) as individual cultic places, which means they have
separate observer point les (not shown) and analyses outputs.
Natalie M. Susmann
8Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
on how easily observers across the study area could see a certain
location (Table 1). These analyses use a reverse viewshed: obser-
vers are represented as vector points, spread across the entire
study area, and viewsheds are calculated for each. The cumulative
output represents the number of times every location on the
landscape was seen. Reverse viewshed accuracy depends on the
number of applied observer points. Ideally, a total viewshed
approach should be used: observers are placed on every pixel of
the Digital Elevation Model (DEM). Accuracy aside, total view-
sheds are unrealistic for large study areas. The high number of
observers means decreased processing speeds and virtually
unmanageable output data sizes (Llobera 2003:33). Cumulative
viewsheds are the next-best approach. Observers are regularly
spaced across the DEM, but in much smaller numbers, thereby
making the calculation computationally feasible.
Within each study area, I placed observers on a grid, each sepa-
rated by 200 m (Table 2). I calculated viewsheds for every observer
using Boston Universitys Shared Computing Cluster and
r.Viewshed (in GRASS GIS). Per study area, I merged the viewsheds
into a single raster output. Each pixel quanties the number of
times that 30 m
area was seen by the regional observers. In other
words, the raster output quanties an observers concatenated
view, as that person moves through space (Llobera 2003:33;
Figure 6a).
Although regional differences are apparent from the raster, I decided
to translate the data into comparable scores representing the num-
ber of times each cultic place was seen. To do so, I vectorized the
output and intersected it with a shapele recording each places
boundary. Two scores were calculated (Figure 6b). The rst is CVSO,
or the numberof times that placewas seen, with no maximum visible
radius set. The second is CVSO 5K. CVSO 5K normalizes the CVSO
based on the number of observers within a 5 km radius of the cultic
place. It measures prominence on a localized level, and when
compared against CVSO, it determines whether a places promin-
ence was dependent upon observer proximity.
Visual Prominence of Observer (VPO). VPO measures visible
prominence using an alternative perspective. Prominence is based
on shape: whether a landform is steep, isolated, and pointed
(Bernardini et al. 2013:3948; Figure 7). My approach is based on
the methodology pioneered by Bernardini and colleagues
(2013:3947), which measures local prominence, or noticeability
from xed locations of interest.
I adjusted their methodology to instead measure global visual
prominencethat is, to compare noticeability across the land-
scape, with no xed observer location in mind. To do so, I placed
gridded observers across the study area, each separated by
2,000 m (Tables 1 and 2)
, and I measured the extent and shape of
each observers 360-degree horizon line. This was done using
ArcGISs Skyline tool. Bernardini and colleagues(2013) custo-
mized Python scripts simplied the horizon line to its steepest,
pointiest, and most isolated peaks, and gave each a comparative
score (Barnash 2011; Bernardini et al. 2013:39483951). I merged
results from every observer, producing a vector shapele record-
ing prominent areas in each study areathat is, prominent areas
based on Bernardini and colleagues (2013): characteristics of
slope, shape, and location (Figure 8). Any score greater than 0 is
considered visually prominent. In order to relate these prominent
landforms to cultic activity, I intersected the polygonal boundaries
with cultic place boundaries (i.e., PlaceBounds.shp). Cultic places
were, in turn, assigned that score.
Cumulative Viewshed of Location (CVSL). CVSL represents the
extent of view from a cultic place. Most archaeological studies
measure binary viewshedsthat is, the visible extent available to
one or a few observers. This approach can be useful in identifying
TABLE 2. GIS Files Used for the Visualscape Analyses.
File Name Description
Associated Analyses/
Type Placement Rules
Places Locations of each surveyed
cultic place
Survey, data management,
and all analyses
Point Centroid of cultic place
Waypoints Locations of surveyed visual
data; links to Places via
unique identiers
Ground-truthing Point GPS locations where data was recorded
DEM 30 m SRTM Digital Elevation
All Raster
Observer_CVSL Observers standing in and
around each cultic places
(i.e., PlaceBounds)
CVSL (Cumulative Viewshed
of Location)
Point Location and # standardized per place type.
Placement ensures peripheral and foveal vision
within and around cultic places. PlaceBounds and
site plans used to avoid overlapping
contemporary structures. If extent changed over
time, multiple sets of observers are made.
Observer_CVSO Observers across
the landscape
CVSO (Cumulative
Viewshed of Observer)
Point One set of observers is created per study area.
Placed 200 m apart on all areas >49° in slope.
Cannot overlap with PlaceBounds.
Observer_VPO Observers across
the landscape
VPO (Visual Prominence of
Point One set of observers is created per study area.
Placed 200 m apart on all areas >49° in slope.
Cannot overlap with PlaceBounds.
Regional Ways of Seeing
2020 |Advances in Archaeological Practice | A Journal of the Society for American Archaeology 9
lines of sight (Danahy 2001)whether, for example, Minoan peak
sanctuaries were visible from one anotherbut the results are also
quite static. A binary viewshed models an unrealistic situation:
what a motionless, forward-facing observer sees (Llobera 2003:33).
CVSL addresses these limitations, recalling the logical principles
of the CVSO approach: increasing the number of observers will
produce a more accurate model of view (Table 2). By situating
observers throughout the space of the cultic place, CVSL accounts
for both peripheral and foveal visionit conveys the topographic
changes encountered as one moved around and looked about. It
models a sense of the complete scene(Danahy 2001:126),
better than a traditional binary approach does.
The methodology was easily implemented. I created a set of
observer points for each cultic place, paying mind to changing
boundaries over time, as well as adjacent contemporary structures
cultic or otherwise (Table 2;Figure 5). Using a viewshed tool
(e.g., Visibility in ArcGIS or one similar), I calculated the total vis-
ible area from each observer point. The resultant raster represents
what a worshiper could see while moving through and around
the cultic place.
Topographic Prominence of Location (TPL). TPL is the measurable
difference between the height of an observerand the surroundings.
It gauges noticeability based on a places physical ruggedness
(Llobera 2010:1007). TPL is measured by (1) identifying a pixel of
interest, (2) demarcating a neighborhood around that place, and (3)
measuring the height differences between the pixel of interest and
the pixels in its neighborhood. Although the resultant measure-
mentdifferences in altitudecontributes to our understanding of
landform noticeability, there are challenges associated with related
parameters. The researcher chooses the neighborhood size,
thereby impacting whether a locale is determined to be more
rugged than its surroundings. In other words, a locale can be
exceedingly prominent on a local scale but not regionally.
Due to these methodological challenges, topographic promi-
nence results were described in context with visual prominence:
whether slope contributed to visual prominence, and whether any
cultic places were situated on rugged but also unnoticeable areas.
In the case of the latter, topographic prominence could highlight
an alternativeand largely understudiedperspective: whether
Greek worshipers ever preferred difcult-to-access locales that
were largely invisible to passersby.
FIGURE 6. (a) CVSO (Cumulative Viewshed of Observer) measurement is performed by creating a grid of observers within the
regional study area. For each of these observers, a separate viewshed is calculated. The nal merged raster is a continuous dataset
comparing regional noticeability. (b) Each places polygonal boundary or boundaries are intersected with the raster, and the
following is calculated: (sum of # times seen/# pixels intersecting the boundary) = prominence.
Natalie M. Susmann
10 Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
There are several methods for measuring topographic promi-
nence (Christopherson 2003; De Reu et al. 2011,2013; Llobera
2003). I chose Focal Statistics in ArcGIS (Tables 1 and 2). As was the
case with previous analyses, I intersected the results with the cultic
place boundaries shapele in order to ascertain the average,
minimum, and maximum scores within the use area. This article
presents average scores.
Data Presentation. I performed each analysisCVSO, CVSO 5K,
VPO, CVSL, and TPLper cultic place, and if that places
boundary changed over time, I ran the analyses per each time
phase. I normalized the values between 0 and 1, and I graphed
them per century using JMP software. The resultant model
represents the average visualscape values per century, based on
changes on the landscape: new constructions or newly recycled
locales (Figure 9). It conveys the sorts of natural environments
preferred for worship, and the way these preferences changed
over time.
The results identify distinctive moments of time when ancient
Greeks in Messenia and the Argolid preferred specic types of
natural environments for worshipspecically, when they
developed cultic places at locales that were steep, that were
noticeable, or that provided a vast viewshed (Figure 9). Their
preferences can be contextualized against other historical and
archaeological data; we can surmise the motives for preferring
specic environs over others throughout their culture history.
That discussion is beyond the methodological scope of this
article and can be found in Susmann (2019b). Instead, I will
engage with both phases of the methodologyeldwork
and geospatial analysesand discuss the benets of my conuent
Geospatial Approaches Produce Objective,
Comparable Data
This article utilizes Geographic Information Systems (GIS) to
measure visual prominence, ruggedness, and visible extent.
Although archaeologists highlight phenomenology as an alter-
native (e.g. Rennell 2012), these approaches produce qualitative
data, making site-to-site comparisons extremely difcult. Also
problematic is their subjectivity. Cultural context and observer
identity will impact how a landscape is seen, experienced, and
perceived. Archaeologistshypotheses and historical knowledge
inevitably impact their descriptions.
My geospatial methodology produces objective, comparable
data. It facilitates exploration about whether viewshed or
FIGURE 7. VPO (Visual Prominence of the Observer) measures visual prominence based on landformscomparable elevation,
slope, and visual isolation: (a) these landforms yield VPO scores of 0 because they are not distinctive from one another; (b) these
landforms yield scores greater than 0 because they are isolated from one another and form distinctive shapes against the horizon.
Regional Ways of Seeing
2020 |Advances in Archaeological Practice | A Journal of the Society for American Archaeology 11
prominenceor bothwere culturally relevant in the cultic place-
making process. In the case of prominence, my methodology
also explores how the same place could have been noticed
differently. The Argive Heraion sanctuary, located in the central
Argive Plain, is an excellent example. Based on the CVSO analysis,
this sanctuary is noticeable. Its location midway up Mount Akraia
meant that it was visible at ranging distances across the Argive
Plain. The VPO analyses tell a different story. The Argive Heraion
yields a VPO score of 0; its mid-slope location made it unnotice-
able against the horizon (Figure 10). Consequently, the Argive
Heraion was noticeable, provided an observer knew where
to look. During eldwork, I did not notice such a distinction.
I anticipated this sanctuarys prominence as a result of previous
scholarswork, all of whom highlight a CVSO-way of looking at the
landscape (Cole 1986:115116; de Polignac 1995:33; Foley
1988:135; Kerényi 2015:117).
There are earlier sanctuaries in the region (e.g., Mount Arachnaion)
with high VPO scores and CVSO scores close to 0, and here is
where comparison against regional datasets, as well as critical
FIGURE 8. VPO (Visual Prominence of the Observer) analysis output is not continuous. Only cultic places that intersect the
polygonal boundaries of visually prominent landforms are considered prominent.
FIGURE 9. I normalized the visualscape analysesaverage scores and graphed them per century. These scores reect new
additions to the landscape or revisitations of older cultic places.
Natalie M. Susmann
12 Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
engagement with excavation and historical data, becomes neces-
sary (Susmann 2019b:140153). Comparing the Argive Heraions
CVSO and VPO results against other contemporary sanctuaries in
theArgivePlain(Figure 11), I concluded that Argive Plain inhabi-
tants valued accessibility alongside prominence at specic
moments, and these preferences are tied to cultural and political
shifts (Susmann 2019b:153). Prior to the seventh century BC, sanc-
tuaries such as Mount Arachnaionlocated at a noticeable
mountain peakwere preferred (Susmann 2019b:8083).
Fieldwork Can Identify Limitations in Analytical
Parameters and Data Collection Strategies
Fieldwork provided critical ground-truthing data. Environmental
factors such as vegetation, smoke, and light will impact what an
observer sees, but archaeologists cannot record geospatial data
for them at a regional level. As I collected data at each cultic
place, I observed whether my ability to move and see was
impeded or enhanced by certain variables.
To that end, eldwork is necessary for identifying whether certain
place types should not be included in the study. Greek archae-
ologists consider burials as cultic places (Antonaccio 1994:398).
My eldwork experiences drove me to exclude their results from
Figure 9.
As I surveyed tholos tombs, for example, I began to
recognize their extremely variable prominence. These are sub-
terranean tombs covered with earthen mounds, and my ability to
spot them depended on topography, vegetation, travel pathway,
and direction of gaze. While Mason (2007) has successfully applied
ancient roadway datasets to understand how travel routes
impacted tholoi prominence, the study is specically about the
citadel of Mycenae and its surrounding tombs. Similar road
datasets are not available for my entire study area.
My chosen geospatial approaches will yield misleading results
about burial prominence. Quantitative data do not convey how
minor topographic changes impact the observers ability to see
these small targets, and only eldwork can provide such context.
Consequently, although burials are considered Greek cultic
places, their functional relationship to the landscape is very
FIGURE 10. The Argive Heraion sanctuary is located midway up Mount Akraia. Although portions of Mount Akraia yield a VPO
(Visual Prominence of the Observer) score greater than 1, the sanctuarys boundaries do not intersect that area.
Regional Ways of Seeing
2020 |Advances in Archaeological Practice | A Journal of the Society for American Archaeology 13
different from places such as altars or temples. Burials require a
separate geospatial strategy, one that potentially measures
degree of invisibility.
Phenomenology Transforms Hyperclean
Archaeological Sites (Weddle 2013:154) into
Active Cultural Spaces
Phenomenology produces real-time data about human engage-
ment with the landscape. At any given site, I encountered tourists,
students, archaeologists, and state employees; I witnessed how
every persons pauses, gazes, and movements were a product of
their identity. These individuals animated my archaeological data
points and transformed them into active cultural spaces (Barrett
and Ko 2009:282).
My observations about modern people guided my interpretations
about the ancient cultic landscape. Phenomenology reminded me
of the precise subjects of my geospatial model. Consider
Panhellenic festivals and the diverse group of visitors they
attracted to sanctuaries every year. Devotees, athletes, politicians,
vendors, and spectators came to the same sanctuary. The way
each person engaged with ittheir sense of place (Ashmore and
Knapp 1999:21)was based on their identity, goals, and level of
piety. This geospatial model measures visual impact as pertains to
a worshiping observer. If an observer looked at the same land-
scape through a different cultural lens (e.g., political, social,
defensive, ancestral), their attention could shift toward other
places, topographies, and sensory experiences.
The geospatial analyses account for change over broad spans of
time. Phenomenology models how visualscapes can change
FIGURE 11. Visualscape scores can be represented per sanctuary and isolate smaller study areas for detailed comparisons.
Natalie M. Susmann
14 Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
within a few hours. I left my apartment on a Saturday morning to
travel to Epidaurus for data collection. It was a peaceful drive, and
I had clear views of the surrounding landscape. A Ministry of
Culture sign served as a landmark for my exit off the highway.
Later that evening, I returned to Epidaurus for a theater perfor-
mance. I followed the same pathway but my experience traveling
drastically changed. Trafc made the route noisy, slow, and lled
with exhaust fumes, and it altered the viewshed I had previously
encountered. It was difcult to see past these other cars. Instead
of a highway sign, the end of my journey was signaled by a grid-
locked off-ramp.
This was a paramountalbeit unplannedphenomenological
experience because it forced me to think more creatively about
immeasurable and time-sensitive impacts. The activities sur-
rounding cultic practiceprocessions, lighting smoky res, and
sacricing animalswould have temporarily increased a places
noticeability (Susmann 2019b:130131, 134136). The geospatial
analyses do account for change over time, but short-term human
impacts cannot be measured without other critical data: festival
dates and durations, crowd size and travel patterns, and the
number of sacral res and their spatial extent.
Many archaeologists have discussed the complementary rela-
tionship between phenomenological approaches and geospatial
analyses (Eve 2012; Fábrega-Álvarez and Parcero-Oubiña 2019;
Fitzjohn 2007; Given 2004; Rennell 2012; Seamon 2014; Thomas
1993,2008; Van Dyke et al. 2016). Geospatial technology produces
quantitative, comparable data about the landscape, but these are
problematically treated as nished product[s](Brück 2005:54;
Thomas 2008:7). Phenomenology explores valuable questions
about how humans shapeand are shaped bythe landscape.
Critics understandably express concern about imprecise approaches
and subjective interpretations (Brück 2005; De Reu et al. 2011;
Fleming 1999). There is a wide gap between pixels and people, and
archaeologists need clear pathways for bridging it. This reexive
methodology constantly interrogates either datasets limitations and
links. Phenomenology activates the geospatial data; quantitative
methods pave the way for systematic examination of cultural behavior.
Ashmores(2002:1173) life history framework was foundational in
developing this approach. Every places formation, decline, and
possible revisitation was accounted for. During both the eldwork
and geospatial phases, I considered how these distinctive phases
of use were physically represented in the landscape. I explored
whether and why individualspathways changed as a result,
thereby dictating geospatial parameters. Creating dynamic time-
lines underscores the complexity of visualscape. What we see is
born from individual and collective senses of place.
This is a precise, analytical framework for treating natural features
as archaeological artifacts. Scholars can use it to collect, record,
and organize archaeological and natural data. They can examine
how the landscape impacted the siting of meaningful placesbe
they for worship or otherwise. That said, interested parties should
be cautious in assuming that visibility analysesand particularly
those described herewill always produce meaningful results.
Gillings says it best: Whilst generating them has always been a
relatively trivial task, knowing what to do with them once generated
has proven far more difcult(2017:121). Indeed, visibility-based
analyses are only as powerful as the questions they answer. There
should be compelling cultural data to inspire their implementation.
This methodology was applied to the ancient Greek Peloponnese,
but its impact reaches farther. All datatypological, visual, and
methodologicalare being moved into an open-access user
interface (Susmann 2019a).
In addition to survey data, the inter-
face provides access to methodological tutorials and database
and interface codes, as well as my daily journals kept during the
eldwork and research phases.
I would like to thank Andrea Berlin, Mac Marston, Christopher
Roosevelt, Curtis Runnels, and Bonna Wescoat for their guidance.
I am indebted to staff at the American School of Classical Studies
at Athens (ASCSA) and the Greek Archaeological Service for their
cooperation. Data collection at the Sanctuary of Apollo Maleatas
was completed in compliance with permit number 6569, granted
by the Ephorate of Antiquities of Argolida. Data collection at
Bassae was completed under permit number 2427, and data col-
lection at Skilloundia was completed in compliance with permit
number 6795, both granted by the Ephorate of Antiquities of Ilia.
Data collection at all other sites did not require research permits,
conrmed by the ASCSA and Greek Archaeological Service. I am
grateful for the invaluable technical support offered by Wes
Bernardini, Mark Gillings, Brian Gregor, and Boston Universitys
Hariri Center for Computing and Computational Science &
Engineering. This research was completed with the following
research grants: Boston University Graduate Research Abroad
Fellowship, Boston University Graduate Student Organization
travel grant, GeoEye Foundation Award for the Application of
High-Resolution Digital Satellite Imagery, and Boston University
Digital Humanities Seminar Start-Up Grant.
Data Availability Statement
Spatial and visual data, methodological tutorials, eld notes, and
research journals from this project are freely available on an
open-source interface: The interface was
launched in mid-September 2019. Readers are also encouraged to
contact the author with any questions or data requests.
Supplemental Materials
For supplemental material accompanying this article, visit https://
Supplemental Database. Template of the Access database
developed to record data for the survey as well as geospatial
results. Field descriptions are included in each tables metadata;
tables are also populated with sample records.
Supplemental Text. Descriptions of database schema to be used
in conjunction with table metadata.
1. Visualscape was made popular by Marcos Llobera, who wanted to coin a
single phrase capturing the multitude of analytical approaches by which an
Regional Ways of Seeing
2020 |Advances in Archaeological Practice | A Journal of the Society for American Archaeology 15
archaeologist could describe all of the visual qualities for a place. See
Llobera 2001,2003,2004,2007; Llobera et al. 2010.
2. Behaviors include prayer, sacrice, construction, and intentional deposit of
objects, as well as burials. Excluding Bronze Age (ca. 28001100 BC) tombs
converted into post-eleventh-century BC sanctuaries, burial results are not
presented here.
3. See Linkstable in Supplemental Database.
4. See Place_Periodtable in Supplemental Database.
5. Experiential archaeologyis an equivalent term (Rennell 2012:511512).
6. See Susmann (2019a,2019b) for detailed descriptions of eld equipment.
7. Lake and Ortega (2013:223224) employed a 100 km total viewshed to
examine the visibility between 306 stone circles in Great Britain, amounting
to 29,624 viewsheds calculated in 18 days. Considering my much larger
study area, I applied a cumulative approach. I began with observers sepa-
rated by 1,000 m, using ArcGIS running on two Alienware Area-51 machines.
Power outages led me to contact Boston Universitys Shared Computing
Cluster (SCC), and I was able to decrease observer distance to 200 m apart.
For the Argolid, this was 70,378 observers, and for Messenia, 101,278
observers. Total CPU time was seven days and six hours, and 10 days and six
hours, respectively.
8. Owing to institutional CPU time restrictions, VPO was not completed using
the SCC. The minimum feasible radius for two Alienware Area-51 machines
was 2,000 m. Results are consistent with eld observations.
9. Certain Bronze Age (c. 28001050 BC) burials were reused as altars after the
tenth century BC. These burials were included in the analyses, but only as
pertains to their new cultic function.
10. Tutorials are publicly available. To protect site locations, researchers can
apply for data access.
Antonaccio, Carla M.
1994 Contesting the Past: Hero Cult, Tomb Cult, and Epic in Early Greece.
American Journal of Archaeology 98:389410.
Ashmore, Wendy
2002 Decisions and Dispositions: Socializing Spatial Archaeology. American
Anthropologist 104:11721183.
2014 On Ancient Placemaking. In Of Rocks and Water: Towards an
Archaeology of Place, edited by Ömür Harmanşah, pp. 4046. Oxbow
Books, Oxford.
Ashmore, Wendy, and A. Bernard Knapp
1999 Archaeologies of Landscape: Constructed, Conceptualized, Ideational.
In Archaeologies of Landscape: Contemporary Perspectives, edited by
Wendy Ashmore and A. Bernard Knapp, pp. 125. Blackwell Press, Malden,
Barnash, Alicia Nicole
2011 Hopi Horizon Analysis with GIS. Masters thesis, Geographic Information
Systems, University of Redlands, Redlands, California.
Barnett, Rod
2007 Sacred Groves: Sacrice and the Order of Nature in Ancient Greek
Landscapes. Landscape Journal 26:252270.
Barrett, John C., and Ilhong Ko
2009 A Phenomenology of Landscape: A Crisis in British Landscape
Archaeology? Journal of Social Archaeology 9:275294.
Belis, Alexis Marie
2015 Fire on the Mountain: A Comprehensive Study of Greek Mountaintop
Sanctuaries. PhD dissertation, Department of Art and Archaeology,
Princeton University, Princeton, New Jersey.
Bernardini, Wesley, Alicia Nicole Barnash, Mark Kumler, and Martin Wong
2013 Quantifying Visual Prominence in Social Landscapes. Journal of
Archaeological Science 40:39463954.
Blyth, Simon, Brian Groombridge, Igor Lysenko, Lera Miles, and Adrian Newton
2002 Mountain Watch: Environmental Change and Sustainable Development
in Mountains. UNEP World Conservation Monitoring Centre, Cambridge.
environmental-change-sustainable-development-in-mountains, accessed
February 20, 2020.
Boivin, Nicole
2004 Landscape and Cosmology in the South Indian Neolithic: New
Perspectives on the Deccan Ashmounds. Cambridge Archaeological
Journal 14:235257.
Bradley, Richard
2000 An Archaeology of Natural Places. Routledge, New York.
Briault, Camilla
2007 Making Mountains out of Molehills in the Bronze Age Aegean: Visibility,
Ritual Kits, and the Idea of a Peak Sanctuary. World Archaeology 39:
Brück, Joanna
2005 Experiencing the Past? The Development of a Phenomenological
Archaeology in British Prehistory. Archaeological Dialogues 12:4572.
Buxton, Richard G. A.
1992 Imaginary Greek Mountains. Journal of Hellenic Studies 122:115.
1994 Imaginary Greece: The Contexts of Mythology. Cambridge University
Press. Cambridge.
Cherry, John
1983 Frogs Round the Pond: Perspectives on Current Archaeological Survey
Projects in the Mediterranean Region. In Archaeological Survey in the
Mediterranean Area, edited by Donald R. Keller and David W. Rupp, pp.
375416. BAR International Series 155. Archaeopress, Oxford.
Christopherson, Gary L.
2003 Using ARC/GRID to Calculate Topographic Prominence in an
Archaeological Landscape. In Proceedings of Arc/INFO User Conference,
San Diego, California.
Cole, Susan Guettel
1986 The Uses of Water in Greek Sanctuaries: Early Greek Cult Practice.
Proceedings of the Fifth International Symposium at the Swedish Institute
at Athens 4:161165.
2000 Landscape of Artemis. Classical World 93:471481.
2004 Landscapes, Gender, and Ritual Space: The Ancient Greek Experience.
University of California Press, Berkeley.
Cooper, Frederick A.
1996 The Temple of Apollo Bassitas I: The Architecture Illustrations. Vol. I.
American School of Classical Studies, Athens, Greece.
Danahy, John W.
2001 Technology for Dynamic Viewing and Peripheral Vision in Landscape
Visualization. Landscape and Urban Planning 54:125137.
de Boer, Jelle Zeilinga, and John R. Hale
2008 The Geological Origins of the Oracle at Delphi. In The Archaeology of
Geological Catastrophes, edited by W. G. McGuire, D. R. Grifths, P.
L. Hancock, and I. S. Stewart, pp. 399412. Geological Society of London,
de Polignac, François
1994 Mediation, Competition, and Sovereignty: The Evolution of Rural
Sanctuaries in Geometric Greece. In Placing the Gods: Sanctuaries and
Sacred Space in Ancient Greece, edited by Susan E. Alcock and
Robin Osbourne, pp. 318. Oxford University Press, Oxford.
1995 Cults, Territory, and the Origins of the Greek City-State. University of
Chicago Press, Chicago.
De Reu, Jeroen, Jean Bourgeois, Machteld Bats, Ann Zwertvaegher, Vanessa
Gelorini, Philippe De Smedt, Wei Chu, Marc Antrop, Philippe De Maeyer,
Peter Finke, Marc Van Meirvenne, Jacques Verniers, and Philippe Crombé
2013 Application of the Topographic Position Index to Heterogeneous
Landscapes. Geomorphology 186:3949.
De Reu, Jeroen, Jean Bourgeois, Philippe De Smedt, Ann Zwertvaegher,
Marc Antrop, Machteld Bats, Philippe De Maeyer, Peter Finke, Marc Van
Meirvenne, Jacques Verniers, and Philippe Crombé
2011 Measuring the Relative Topographic Position of Archaeological Sites in
the Landscape: A Case Study on the Bronze Age Barrows in Northwest
Belgium. Journal of Archaeological Science 38:34353446.
Doxtater, Dennis
2009 Rethinking the Sacred Landscape: Minoan Palaces in a Georitual
Framework of Natural Features on Crete. Landscape Journal 28:121.
Driessen, Jan
2003 The Court Compounds of Minoan Crete: Royal Palaces or Ceremonial
Centers? Athena Review 3(3):5762.
Natalie M. Susmann
16 Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
Eliade, Mircea
1961 Images and Symbols: Studies in Religious Symbolism. Sheed & Ward,
New York.
Eve, Stuart
2012 Augmenting Phenomenology: Using Augmented Reality to Aid
Archaeological Phenomenology in the Landscape. Journal of
Archaeological Method and Theory 19:582600.
Fábrega-Álvarez, Pastor, and César Parcero-Oubiña
2019 Now You See Me: An Assessment of the Visual Recognition and Control
of Individuals in Archaeological Landscapes. Journal of Archaeological
Science 104:5674.
Faro, Elissa Z.
2008 Ritual Activity and Regional Dynamics: Towards a Reinterpretation of
Minoan Extra-Urban Ritual Space. PhD dissertation, Classical Art and
Archaeology, University of Michigan, Ann Arbor.
Fitzjohn, Matthew
2007 Viewing Places: GIS Applications for Examining the Perception of Space
in the Mountains of Sicily. World Archaeology 39:3650.
Fleming, Andrew
1999 Phenomenology and the Megaliths of Wales: A Dreaming Too Far?
Oxford Journal of Archaeology 18:119125.
Foley, Anne
1988 The Argolid 800600 B.C.: An Archaeological Survey, Together with an
Index of Sites from the Neolithic to the Roman Period.P.A
Göteborg, Sweden.
Gillings, Mark
2015 Mapping Invisibility: GIS Approaches to the Analysis of Hiding and
Seclusion. Journal of Archaeological Science 62:114.
2017 Mapping Liminality: Critical Frameworks for the GIS-based Modelling of
Visibility. Journal of Archaeological Science 84:121128. DOI:10.1016/j.jas.
Given, Michael
2004 From Density Counts to Ideational Landscapes: Intensive Survey,
Phenomenology, and the Sydney Cyprus Survey. In Archaeological
Landscapes: Current Issues, edited by Efe F. Athanassopoulos and
LuAnn Wandsnider, pp. 165182. University of Pennsylvania Museum of
Archaeology and Anthropology, Philadelphia.
Goodison, Lucy
2004 From Tholos Tomb to Throne Room: Some Considerations of Dawn
Light and Directionality in Minoan Buildings. British School at Athens
Studies 12:339350.
Grove, David C.
1999 Public Monuments and Sacred Mountains: Observations on
Three Formative Period Sacred Landscapes. In Social Patterns in
Pre-Classic Mesoamerica: A Symposium at Dumbarton Oaks,edited
by David C. Grove and Rosemary A. Joyce, pp. 255299. Dumbarton
Oaks, Washington, DC.
Håland, Evy Johanne
2009 Take, Skamandros, My Virginity: Ideas in Connection with Rites of
Passage in Greece, Modern and Ancient. In The Nature and Function of
Water, Baths, Bathing, and Hygiene from Antiquity through the
Renaissance, Vol. 11, edited by Cynthia Kosso and Ann Scott, pp. 109149.
Brill Publishers, Leiden, Netherlands.
Hannah, Robert
2013 Greek Temple Orientation: The Case of the Older Parthenon in Athens.
Nexus Network Journal 15:423443.
Hitchcock, Louise A.
2007 Naturalizing the Cultural: Architectionised Landscape as Ideology in
Minoan Crete. British School at Athens Studies 15:9197.
Jett, Stephen C.
1995 Navajo Sacred Places: Management and Interpretation of Mythic
History. Public Historian 17(2):3947.
Kerényi, Karl
2015 Zeus and Hera: Archetypal Image of Father, Husband, and Wife.
Princeton University Press, Princeton, New Jersey.
Lake, Mark, and Damon Ortega
2013 Compute-Intensive GIS Visibility Analysis of the Settings of Prehistoric
Stone Circles. In Computational Approaches to Archaeological Spaces,
edited by Andrew Bevan and Mark Lake, pp. 213241. Left Coast Press,
Walnut Creek, California.
Langdon, Merle K.
2000 Mountains in Greek Religion. Classical World 93:461470.
Larson, Jennifer
2007 Ancient Greek Cults: A Guide. Routledge, New York.
2010 A Land Full of Gods: Nature Deities in Greek Religion. In A Companion
to Greek Religion, Blackwell Companions to the Ancient World, edited by
Daniel Ogden, pp. 5670. Wiley-Blackwell, Oxford.
Leick, Gwendolyn
2002 Mesopotamia: The Invention of the City. Allen Lane/Penguin, London.
Llobera, Marcos
2001 Building Past Landscape Perception With GIS: Understanding
Topographic Prominence. Journal of Archaeological Science 28:1005
1014. DOI:10.1006/jasc.2001.0720.
2003 Extending GIS-Based Visual Analysis: The Concept of Visualscapes.
International Journal of Geographical Information Science 17:2548.
2004 What You See Is What You Get? Visualscapes, Visual Genesis and
Hierarchy. In Digital Archaeology: Bridging Method and Theory, edited by
Thomas L. Evans and Patrick Daly, pp. 148167. Routledge, London.
2007 Reconstructing Visual Landscapes. World Archaeology 39(1):5169.
Llobera, Marcos, David Wheatley, James Steele, Simon Cox, and Oz Parchment
2010 Calculating the Inherent Visual Structure of a Landscape (Total
Viewshed) Using High Throughput Computing. In Beyond the Artifact:
Digital Interpretation of the Past: Proceedings of CAA2004, edited by
Franco Niccolucci and Sorin Hermon, pp. 146151. Archaeolingua,
Budapest, Hungary.
Lucero, Lisa J., and Andrew Kinkella
2014 A Place for Pilgrimage: The Ancient Maya Sacred Landscape of Cara
Blanca. In Of Rocks and Water: Towards an Archaeology of Place, edited by
Ömür Harmanşah, pp. 1339. Oxbow Books, Oxford.
Mason, David J.
2007 The Location of the Treasury of Atreus. Oxford Journal of Archaeology
Morgan, Catherine
1990 Athletes and Oracles: The Transformation of Olympia and Delphi in the
Eighth Century BC. Cambridge University Press, Cambridge.
Nixon, Lucia
2009 Investigating Minoan Sacred Landscapes. Hesperia Supplements
Ogburn, Dennis E.
2006 Assessing the Level of Visibility of Cultural Objects in Past Landscapes.
Journal of Archaeological Science 33:405413.
Peateld, Alan
1983 The Topography of Minoan Peak Sanctuaries. Annual of the British
School at Athens 78:273279.
Reinhard, Johan, and María Constanza Ceruti
2010 Inca Rituals and Sacred Mountains: A Study of the Worlds Highest
Archaeological Sites. Cotsen Institute of Archaeology Press, Los Angeles.
Rennell, Rebecca
2012 Landscape, Experience, and GIS: Exploring the Potential for Methodological
Dialogue. Journal of Archaeological Method and Theory 19:510525.
Retallack, Gregory J.
2008 Rocks, View, Soils, and Plants at the Temples of Ancient Greece.
Antiquity 82:640657.
Richards-Rissetto, Heather
2017 An Iterative 3D GIS Analysis of the Role of Visibility in Ancient Maya
Landscapes: A Case Study from Copan, Honduras. Digital Scholarship in
the Humanities 32:195212.
Scully, Vincent
1962 The Earth, the Temple, and the Gods: Greek Sacred Architecture. Yale
University Press, New Haven, Connecticut.
1991 The Sacred Mountain: In Mesopotamia, Egypt, and the Aegean. In
Architecture: The Natural and the Manmade, edited by Vincent Scully,
pp. 2337. St. Martins Press, New York.
Seamon, David
2014 Place Attachment and Phenomenology: The Synergistic Dynamism of
Place. In Place Attachment: Advances in Theory, Methods and
Regional Ways of Seeing
2020 |Advances in Archaeological Practice | A Journal of the Society for American Archaeology 17
Applications, edited by Lynne C. Manzo and Patrick Devine-Wright, pp. 11
22. Routledge, New York.
Soetens, Steven
2009 Juktas and Kophinas: Two Ritual Landscapes out of the Ordinary.
Hesperia Supplements 42:261268.
Soetens, Steven, Apostolos Sarris, and Sophia Topouzi
2001 Peak Sanctuaries in the Minoan Cultural Landscape. In 9th International
Congress of Cretan Studies, edited by Theocharis Detorakis and
Stella Papdaki-Oekland, pp. 213. Society of Cretan Historical Studies,
Heraklion, Crete, Greece.
Soetens, Steven, Apostolos Sarris, Sophia Topouzi, and Achilleas Tripolitsiotis
2002 GIS Modeling of the Minoan Peak Sanctuaries of East Crete. In
Computer Applications and Quantitative Methods in Archaeology:
Proceedings of the 29th Conference, Gotland, Sweden, edited by
Göran Burenhult and Johan Arvidsson, pp. 129138. Archaeopress,
Soetens, Steven, Apostolos Sarris, K. Vansteenhuyse, and S. Topouzi
2002 GIS Variations on a Cretan Theme: Minoan Peak Sanctuaries. In Metron:
Measuring the Aegean Bronze Age: Proceedings of the 9th International
Aegean Conference, Vol. 24, edited by Karen Polinger Foster and Robert
Lafneur, pp. 483488. University of Liège, Belgium.
Susmann, Natalie M.
2019a Greek Natural Projects Interface. Data and Methods Repository.
Boston., accesssed February 20, 2020.
2019b The Nature of Cult: Visualizing Continuity on the Greek Cultic
Landscape in the Argolid and Messenia (c. 2800146 BCE). PhD disserta-
tion, Department of Archaeology, Boston University, Boston.
Thomas, Julian
1993 The Politics of Vision and the Archaeologies of Landscape. In
Landscape: Politics and Perspectives, edited by Barbara Bender, pp. 1948.
Berg, Oxford.
2004 Archaeology and Modernity. Routledge, London New York.
2008 On the Ocularcentrism of Archaeology. In Archaeology and the Politics
of Vision in a Post-Modern Context, edited by Julian Thomas and Vítor
Oliveira Jorge, pp. 112. Cambridge Scholars Publishing, Cambridge.
Tilley, Christopher
1994 A Phenomenology of Landscape: Places, Paths, and Monuments. Berg,
Topouzi, Sophia, Steven Soetens, Anthoula Gkiourou, and Apostolos Sarri
2000 The Application of Viewshed Analysis in Greek Archaeological
Landscape. Paper presented at the 6th Annual Meeting of the European
Association of Archaeologists, Lisbon, Portugal.
Tuan, Yi-Fu
1977 Space and Place: The Perspective of Experience. University of Minnesota
Press, Minneapolis.
Turner, Harold W.
1979 From Temple to Meeting House: The Phenomenology and Theology of
Places of Worship. Mouton Publishers, Paris.
van de Moortel, A.
2006 A Reexamination of the Pottery from the Kamares Cave. In Pottery and
Society: The Impact of Recent Studies in Minoan Pottery: Gold Medal
Colloquium in Honor of Philip P. Betancourt, pp. 7393. Archaeological
Institute of America, Boston.
2011 The Phaistos Palace and the Kamares Cave: A Special Relationship. In
Our Cups Are Full: Pottery and Society in the Aegean Bronze Age: Papers
Presented to Jeremy B. Rutter on the Occassion of His 65th Birthday. edited
by Walter Gauß, Michael Lindblom, R. Angus K. Smith, and James C. Wright,
pp. 306318. BAR International Series 2227. Archaeopress, Oxford.
Van Dyke, Ruth M., R. Kyle Bocinsky, Thomas C. Windes, and Tucker J. Robinson
2016 Great Houses, Shrines, and High Places: Intervisibility in the Chacoan
World. American Antiquity 81:205230.
Weddle, Candace
2013 The Sensory Experience of Blood Sacrice in Roman Imperial Cult. In
Making Senses of the Past: Toward a Sensory Archaeology, edited by
Jo Day, pp. 137159. Center for Archaeological Investigations Occasional
Paper No. 40. Southern Illinois University Press, Carbondale.
Wescoat, Bonna D.
2017 The Sanctuary of the Great Gods, Samothrace. In Excavating Pilgrimage:
Archaeological Approaches to Sacred Travel and Movement in the Ancient
World, edited by Troels Myrup Kristensen and Wiebke Friese, pp. 6786.
Routledge Press, New York.
Wickstead, Helen
2009 The Uber Archaeologist. Journal of Social Archaeology 9:249271.
Williams, Mary Francis
1991 Landscape in the Argonautica of Apollonius Rodius. Peter Lang,
Frankfurt am Main, Germany.
Williams, Patrick Ryan, and Donna J. Nash
2006 Sighting Theapu: A GIS Analysis of Wari Imperialism and the Worship of
Mountain Peaks. World Archaeology 38:455468.
Williamson, Christina
1993 Light in Dark Places: Changes in the Application of Natural Light in
Sacred Greek Architecture. PHAROS: Journal of the Netherlands Institute
at Athens 1:137.
Natalie M. Susmann History Section, Massachusetts Institute of Technology,
77 Massachusetts Avenue, Cambridge MA 02139, USA (
Natalie M. Susmann
18 Advances in Archaeological Practice | A Journal of the Society for American Archaeology | 2020
... Archaeologists have uncovered physical evidence of these beliefs. With diverse analytical approaches, scholars have explored how on the sacred landscape, the ancient Greeks underscored the auras they perceived (Scully 1962;Williamson 1993;Driessen 2003;Goodison 2004;Moortel 2006;Barnett 2007;Hitchcock 2007;De Boer and Hale 2008;Faro 2008;Retallack 2008;Moortel 2011;Hannah 2013;Belis 2015;Susmann 2019Susmann , 2020. In these investigations, the Greek island of Crete has held special interest. ...
... These sorts of explorations are Scully's (1962) seminal volume The Earth, The Temple, and The Gods ambitiously traced potential threads of continuity between the Greek mainland and Crete. More recently, scholars like Mason (2007), Psychoyos and Karatzikos (2015) Susmann (2019, 2020 have explored how visibility and viewshed shaped sacred landscapes on mainland Greece. In this paper, I trace changing human views and interrogate how behavior, setting, and wider historical circumstance combined to create new ways of seeing older places. ...
Full-text available
This paper explores the reciprocal relationship between landscape, human attention, and time. It presents two sacred landscapes: Epidaurus and Nemea, located in the northeastern Greek Peloponnese. In both landscapes, worshipers created sanctuaries on prominent mountains. Eventually, their attention shifted downhill where they built larger, monumental sanctuaries on the flat ground. I trace each mountains’ role as a sacred landmark; I question what other social functions they had – if at all – after the new sanctuaries were built. I consider a wide range of evidence: beginning with a comparative example from Mount Fuji in Japan, and moving onto the archaeological excavation and survey data, ancient testimonials, and modern tourist reviews about the Greek sanctuaries. I also use Geographic Information Systems to quantify each Greek sanctuaries’ visual impact in comparison to the surrounding topography. Woven together, these data reveal generations of sacral continuity. The Sanctuary of Asklepios and the Sanctuary of Zeus encircled new sacred temenoi, but worshipers’ collective memory guided their pathways and vision; the mountains remained sacred landmarks.
Full-text available
Visibility analysis has become extremely popular in landscape-oriented archaeology in recent decades and has become even more widespread with the popularization of GIS tools, which have multiplied the ways in which visual perception can be analysed and digitally modelled. Visibility has been used as a proxy for different archaeological approaches, from the analysis of subjective perception to the assessment of strategic control. While interest has most often been focused on how objects, features and sites are perceived, there has also been an interest in how visual control is exerted from archaeological sites or other places in the landscape. Within the latter approaches, the distances at which visual control can be exercised have usually been determined in a more or less arbitrary manner, without a clear and empirically informed reference on how things and, especially, people can be observed and recognized differently at a distance. In this paper, we present the results of a field experiment carried out to measure the distances at which individuals can be spotted, recognized and identified with the naked eye in favourable conditions. Based on these results, we introduce the concept of Individual Distance Viewshed (IDV) as a GIS-based representation of visual control. This will serve as a reference to better qualify potential visibility in landscape analysis. Finally, we illustrate the applicability of this approach with a case study which explores the relationship between visual control and mobility during the Iron Age in the NW of the Iberian Peninsula.
Full-text available
For several decades, Geographic Information Systems (GISs) have held center stage in archaeological studies of ancient landscapes. Recently, three-dimensional (3D) technologies such as airborne LiDAR and aerial photogrammetry are allowing us to acquire inordinate amounts of georeferenced 3D data to locate, map, and visualize archaeological sites within their surrounding landscapes. GIS offers locational precision, data overlay, and complex spatial analysis. Three-dimensionality adds a ground-based perspective lacking in two-dimensional GIS maps to provide archaeologists a sense of mass and space more closely attuned with human perception. This article uses comparative and iterative approaches ‘tacking back and forth’ between GIS and 3D visualization to explore the role of visibility in conveying sociopolitical and ideological messages at ancient Copan—today a UNESCO World Heritage Site in Honduras. A two-prong approach comprising computational and experiential components explores the potential role of visibility in sending messages that participate in the shaping of social interaction on a daily basis. The organization of built forms within the natural landscape created spatial configurations that sent visual messages targeting specific different groups, subsequently influencing how people negotiated their physical surroundings and the frequency and intensity of social interactions. The ancient Maya belief that sight played a key role in structuring everyday experiences because it triggered perception in the other senses thus serves to bridge the computational and experiential results in this case study.
This dissertation presents a systematic study of the relationship between sanctuaries, visualscapes, and the changing cultural valence of landscape in ancient Greek culture. The ancient Greeks situated their deities within the natural word; godly encounters were particularly expected on mountains. Despite significant archaeological and textual data confirming the important connection between sacred spaces and the natural world, no consistent, systematic framework for determining which natural environments were preferred exists; whether the Greeks sought elevated, noticeable, or rugged environments, and if changes happened over time and space. This dissertation answers those questions by offering a chronological model of ancient Greek worship in the natural world. Geospatial and 3D technology were used to measure the visual and topographic qualities of over 300 cultic places in the Peloponnesian regions of the Argolid and Messenia. These analyses convey that later Greeks shared Bronze Age preferences for worshiping in elevated, visually prominent locales that afforded vast viewsheds. Beginning in the 10th/9th c BCE, worshipers formed new sanctuaries at former Mycenaean cultic places -- but only the most noticeable ones. Over time, the conjoined aspects of vision and view changed in significance. By the 5th c. BCE, as worshippers moved to larger spaces downslope, the number of visitors to certain peak sanctuaries declined. Yet, visual prominence remained an integral part of worship. Historic peak sanctuaries remained important landmarks, guiding worshipers’ movements along a sacred way. At particular moments during the year, lower-lying sanctuaries became more noticeable; festival days meant more noise, smells, and smoke. And, despite the growing popularity of lower-lying locales, new sanctuaries like Bassae, Nomian Pan, and Skilloundia suggest that prominence and viewshed remained important for worship. By placing over 2500 years of Greek religion into the natural landscape, this dissertation presents new and important arguments about place-making; it describes how natural features remained highly influential in the siting of cult. Particular beliefs, practices, and structures changed over time, but the same sorts of physical environments remained integrated and revered. This dissertation thus provides a model for scholars to identify, measure, and describe natural features as artifacts of human behavior.
Since the widespread adoption of GIS by archaeologists in the early 1990s, analyses of visibility have steadily gained traction, becoming commonplace in landscape and regional analysis. This is in large part due to the routine way in which such products can be generated, bolstered by a raft of landscape-based studies that have placed varying degrees of emphasis upon human perception and direct bodily engagement in seeking to understand and explore the past. Despite this seeming popularity, two worrying trends stand out. The first is the lack of any coherent theoretical framework, applications preferring instead to seek justification in the very first wave of experiential landscape approaches that emerged in the early 1990s. Needless to say, the intervening 20 or so years have seen considerable development in the conceptual tools we draw upon in order to make sense of past landscapes, not to mention considerable finessing of the first-wave developments alluded to above. Second is the tendency to relegate viewshed analysis to certain types of predictable problem or question (i.e. viewshed analysis has become typecast). These trends have been compounded by a host of other issues. For example, whilst there have been refinements, tweaks and variations to the basic viewshed (and the frequency with which they are generated and combined), not to mention establishment of robust calibration criteria for controlling them and statistical approaches for assessing the patterns tendered, these have yet to be brought together in any coherent fashion and their veracity critically assessed. Likewise, a failure to establish an agreed vocabulary has resulted in a number of proverbial wheels being reinvented time and again. The argument presented here is that viewsheds have considerably more to offer archaeology but to realise this entails confronting these issues head on. That this is possible and desirable is illustrated through discussion of a new theoretical framework for visibility-studies that draws upon developments in assemblage theory and the author's own work on affordance and relationality. To demonstrate the value of this approach in encouraging different ways of thinking about what viewsheds are and how we might begin to draw creatively upon them, a case-study is described where viewsheds are folded into a detailed exploration of landscape liminality.
Phenomenological archaeologists and GIS scholars have turned much attention to visibility-who can see whom, and what can be seen-across ancient landscapes. Visible connections can be relatively easy to identify, but they present challenges to interpretation. Ancient peoples created intervisible connections among sites for purposes that included surveillance, defense, symbolism, shared identity, and communication. In the American Southwest, many high places are intervisible by virtue of the elevated topography and the open skies. The Chaco phenomenon, centered in northwestern New Mexico between A.D. 850 and 1140, presents an ideal situation for visibility research. In this study, we use GIS-generated viewsheds and viewnets to investigate intervisible connections among great houses, shrines, and related features across the Chacoan landscape. We demonstrate that a Chacoan shrine network, likely established during the mid-eleventh century, facilitated intervisibility between outlier communities and Chaco Canyon. It is most likely that the Chacoans created this network to enable meaningful connections for communication and identity. We conclude that the boundaries of the Chaco phenomenon are defined in some sense by intervisibility.