You are on page 1of 25

Prospect-refuge patterns in Frank Lloyd Wrights Prairie houses:

Using isovist fields to examine the evidence

Michael J. Ostwald, Michael Dawes


University of Newcastle, Australia

Pages: 136-159

The Journal of Space Syntax


ISSN: 2044-7507
Year:

2013

http://www.journalofspacesyntax.org/

volume: 4

issue: 1

Online Publication Date: 5 August 2013

J
S
S

J
O
S
S

Prospect-refuge patterns in Frank Lloyd Wrights Prairie


houses: Using isovist fields to examine the evidence

Michael J. Ostwald, Michael Dawes


University of Newcastle, Australia

It has been argued that the domestic works of Frank Lloyd Wright possess an innate phenomenological

appeal that is a direct result of a particular spatial and visual pattern in Wrights architecture. This argument, which is central to several architectural variants of prospect-refuge theory, suggests that a distinct
spatio-visual system exists which directly shapes the way people move through and experience Wrights
architecture. This paper uses a computational technique, isovist field analysis, to search for prospect-refuge

Keywords:
Isovist analysis,
visibility graph analysis,
spatial perception,
prospect-refuge theory,
Frank Lloyd Wright.

related spatio-visual patterns in paths through five of Frank Lloyd Wrights canonical Prairie houses. The
paper concludes that there is no clear evidence in these particular cases of the suggested pattern.

1. Introduction
The domestic designs of Frank Lloyd Wright, one

sponse felt by visitors. Moreover, through a detailed

of the 20th centurys most famous architects, have

discursive and diagrammatic analysis of Wrights

been repeatedly praised for their capacity to evoke

architecture, Hildebrand describes a particular set

positive experiential or phenomenological re-

of spatial and formal conditions that can be used

sponses (Laseau and Tice, 1992; Lind, 1994; Hale,

to both explain and potentially replicate this sub-

2000; Hienz, 2006). For example, Robert Twombly

conscious human reaction. The formal and spatial

(1979) observes that Wrights Prairie houses elicit a

features Hildebrand identifies could be broadly

deeply felt attraction that is strongly emotional and

categorised in two ways; the first associated with the

psychological (p.78). Multiple attempts have been

way a person passes through a house, from its entry

made to explain these qualities using a range of

to its living areas, and the second with the properties

theories including material ethics (Lin, 1991; Harries,

of the living area itself. For the first of these cases,

1997), spatial archetypes (Assefa, 2003) and genius

the entry or approach path, Hildebrand provides

loci (Norberg-Schulz, 1979; Hale, 2000). One of

detailed axonometric diagrams to identify the pre-

the more detailed and convincing arguments that

cise paths, through several of Wrights houses, that

explains the experience of Wrights architecture is

he argues are significant.

Grant Hildebrands (1991) application and refinement


of prospect-refuge theory.

Since its publication, Hildebrands variation of


prospect-refuge theory has been widely adopted

Prospect-refuge theory suggests that certain

in architecture as both an analytical and a design

spatial and formal configurations, which provide a

strategy (Weston, 2002; Roberts, 2003; Unwin,

balanced combination of outlook and enclosure,

2010). However, despite its evocative qualities,

respond to basic human psychological needs and

there is little evidence confirming the validity of this

thereby stimulate positive emotional reactions (Ap-

interpretation of Wrights architecture. Indeed, one

pleton, 1975; 1988). Hildebrand (1991) develops this

of the foundational features of Hildebrands argu-

theory to argue that Wrights domestic architecture

ment, that the combined visual and spatial proper-

possesses a distinct spatial and formal pattern that

ties of Wrights houses are sufficiently similar as to

is directly responsible for the strong emotional re-

constitute a distinct pattern, has never been tested.


136

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

137

The present research uses computational and

The present paper commences with a brief

mathematical means to test one critical facet of

background to isovist analysis and environmental

Hildebrands theory; the proposition that there are

preference theory; both of which have been exten-

strong similarities between the spatio-visual prop-

sively researched in the past, but have only rarely

erties of the approach paths found in Frank Lloyd

been connected (Franz et al., 2004; Wiener and

Wrights Prairie houses. The particular computa-

Franz, 2005; Stamps III, 2005). The paper then posits

tional method that is used in this paper is isovist field

a series of conditions which should be uncovered

analysis and the test case for this research is a set

if Hildebrands theory is valid. Thereafter the paper

of five Prairie houses: the Henderson house (1901),

describes the method that is used in the present

Heurtley house (1902), Cheney house (1903), Evans

research to analyse the spatio-visual characteristics

house (1908) and Robie house (1910). Hildebrand

of Hildebrands paths through the five houses. Using

describes and maps two of these in detail, and

new computer models of these houses, an isovist

identifies the Heurtley house as the perfect example

field is generated for each and then four major

of his argument, and the remainder are covered in

properties derived from this model are compared.

a more limited way.

The results for each of the five houses are examined

The method used in this paper, isovist analysis,

in detail, before the paper concludes with a discus-

is concerned with the geometric and mathematical

sion of evidence of the presence of a pattern in the

properties of the volume of space that is visible from

spatio-visual experience of the houses.

a single, fixed location. Whereas many arguments

The present research has several technical

about the relationship between space and vision,

and theoretical limitations; the former of which are

like Hildebrands, rely on qualitative propositions

described in detail in the methods section of the

and poetic descriptions, the isovist method provides

paper. In the latter case, the theoretical limitations,

a quantitative approach to one particular facet of

it must be acknowledged that Wrights houses pos-

the experience of space. In the past, isovists have

sess many features that cannot be examined using

been used to provide insights into spatial cognition

isovists. For example, some of the key qualities of

(Meilinger et al., 2009), way-finding (Conroy, 2001),

Wrights houses have been traced to the presence

social structure (Markhede and Koch, 2007) and

of rich surface ornament, dramatic natural outlooks

spatial occupation (Ellard, 2009). Importantly in the

and prominent hearths. While past research has

present context, Stamps III (2005) suggests that

criticised the way in which Hildebrand has inter-

some of the measures derived from isovist geometry

preted such features (Seamon, 1992), their impact

appear to reflect the basic elements of prospect-

is outside the scope of the present research.

refuge theory. However, further examination fails to

Moreover, prospect-refuge theory is fundamentally

identify an ideal set of measures for prospect-refuge

about human emotional or psychological reactions

conditions (Dawes and Ostwald, 2013). While previous

to environments, objects, forms and spaces. These

studies of Wrights domestic architecture have

reactions can be influenced by a wide range of

used mathematical and computational means to

personal factors including age, cultural background

examine formal, geometric and stylistic similarities

and education (Nasar, 1984; Kaplan and Herbert,

(Koning, 1981; Laseau and Tice, 1992; Ostwald and

1988). These reactions are also contingent on

Vaughan, 2010; Vaughan and Ostwald, 2011), no

particular tectonic, spatial and stylistic conditions

previous study has investigated the spatio-visual

(Stamps III, 2006). None of these can be considered

experience of these houses using isovists.

using the present method, which is optimised for

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

examining one facet of Hildebrands version of

The isovist polygon perimeter will consist of surface

prospect-refuge theory, the presence or absence

edges, view limit edges, and occluding edges.

of a spatio-visual pattern in the passage through

Occluding edges extend beyond a surface vertex,

Wrights Prairie architecture.

such as beyond a window, and represent portions


of the isovist perimeter beyond which objects are

2. Background to Isovist Analysis

just out of sight (Figure 1).

The origins of isovist analysis can be traced to

Once the isovist polygon has been generated, a

the optic array model of thinking about sensory

range of mathematical values can be derived from

perception that arose from pioneering pilot training

it that reflect, for example, the volume of space that

programs in the 1940s (Gibson, 1947). Spatio-visual

is visible from the observation point, the furthest

research prior to that time was often focused on

a person can see, or how much of their vision is

understanding perception through an analysis of

constrained by surfaces. Past research has used

the subtended angle of an object in the field of view;

this approach to examine the spatio-visual geometry

that is, the angle occupied by the objects silhouette.

of significant locations in buildings (Stavroulaki and

In contrast, the optic array model holds that all light

Peponis, 2005; Markhede and Koch, 2007; Peponis

entering the eye carries valuable environmental

and Bellal, 2010). Isovist analysis is also favoured

information (Gibson, 1966; Gibson, 1979). Due to

for documenting the experiential properties of art

its origins in flight training, this model is ideal for

galleries and the display of valued objects within

examining the experience of movement through

a spatial structure dedicated to this task (Tzortzi,

space and, significantly, several of Gibsons (1979)

2004; Peponis et al., 2004; Psarra, 2005, 2009;

examples are suggestive of the changing experi-

Zamani and Peponis, 2013). While some of these

ence of a person following a path through a building.

studies track occupants moving through a building,

Michael Benedikt (1979) developed the concept of

or finding their way around space, none examines

the optic array into the more geometrically defined

and compares a systematically defined set of

isovist and, working with Larry Davis (Davis and

paths in the way proposed in the present paper.

Benedikt, 1979), provided a stable and repeatable

For example, circulation routes through a gallery

method for isovist generation. Benedikt (1979) de-

may serve a variety of purposes although they are

fined the isovist as the set of all points visible from

typically optimised for allowing a visitor to move

a single vantage point in space with respect to an

from one display to the next without getting lost. In

environment (p.47). The method has since been

contrast, the present research considers paths that

optimised for computational analysis including the

are less functional and are potentially more attuned

development of several statistical measures that are

to psychological or emotional needs.

derived from the ray properties used to generate


the isovist (Batty, 2001; Christenson, 2010).

For an analysis of the visual and spatial properties of a complete building, or a path through space,

In its most basic form, constructing an isovist

an isovist field is commonly needed (Ostwald

requires extending radial lines from an observa-

and Dawes, 2013). An isovist field is the set of all

tion point to the surfaces of an environment, or a

isovists, generated from a defined spread of loca-

predefined view-length limit. These lines are drawn

tions (typically a grid of points) in a building. This

at a regular angular increment, often one degree,

approach allows researchers to plot the changing

and the end of each line is linked to the ends of their

spatio-visual properties of the building as a scalar

immediate neighbours to form the isovist polygon.

map or to sequentially chart the measured values


138

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

Figure 1:
(a) Radial lines
traced from
observation point.
Occluding radials
shown as solid
lines, Radial lines
shown dotted.
(b) Isovist polygon
defined by surfaces
and occluding
radials.

as a graph. While the three dimensional analysis

istics or qualities that support these conditions

of individual isovists is gradually becoming viable

(Kaplan and Kaplan, 1983; Kaplan and Kaplan,

(Yang et al., 2007; Morello and Ratti, 2009), isovist

1989; Kaplan 1987). However, despite evidence of

field analysis using two dimensional data is more

genetic predisposition, research also suggests that

widely accepted and it is the method used in the

environmental preference is a complex amalgam

present paper.

of factors, with innate preferences forming a foundation which is then overlain by both socio-cultural

139

3. Background to prospect-refuge theory

and personal experience factors (Falk and Balling,

Habitat theory suggests that animals, including

2009, p.489). In order to examine and explain the

humans, are biologically motivated to seek envi-

complex impact of these factors, Appleton (1975)

ronmental conditions that are conducive to their

proposed prospect-refuge theory.

survival (Appleton, 1975). Environmental and spatial

Prospect-refuge theory conceptualises envi-

psychologists have also demonstrated that humans

ronments that are conducive to survival in terms

are drawn to intuitively, unconsciously and virtually

of the predator and prey. The central idea being

instantaneously appraise spaces for character-

that, in order for a predator to hunt successfully it

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

must be able to track its prey while remaining hid-

been largely focussed on psychological testing,

den. Conversely, hunted species require similar

rather than on a concomitant analysis of the prop-

environmental conditions: the ability to identify

erties of the environment itself. This is because

and avoid predators while remaining hidden. Thus,

many aspects of prospect-refuge theory rely on

prospect in Appletons (1975) theory refers to an

the interpretation of environmental symbols. For

unimpeded opportunity to see (p.73) and refuge

example, a dark forest in a distant view may evoke

to an opportunity to hide (ibid.) and, in the right

a sense of fear, while a waterfall, in exactly the

combination, these two spatial properties define

same location and filling exactly the same volume

an environment that is conducive to survival and

of space, may evoke a sense of excitement. In order

thereby yields environmentally derived pleasure. In

to overcome this problem Kaplan (1987) proposes

its architectural variant, Hildebrand (1991; 1999) ar-

that the primary features of an environment should

gues that Wrights living rooms provide a perfect mix

be extracted to construct a rough conceptual model

of outlook and enclosure - an ideal space for survival

of the three-dimensional space represented by the

at the end of complex approach path. Twombly

scene (p.22). In this way, for a reasonable analysis

(1979) similarly argues that Wrights Prairie houses

to occur, the environment should be converted into

elicit a positive emotional response by emphasising

a pure geometric system that is devoid of complex

in literal and symbolic ways, the security, privacy

symbolic or semiotic connotations. In the present

[and] shelter [that ] people found important in a

research, the geometry of Wrights architecture is

period of [] conflict (p.78).

well defined and therefore an ideal case for analysis

As acceptance of prospect-refuge theory grew,


additional measures derived from research into

using geometric, mathematical and computational


means.

environmental preference began to be connected


to it. Probably the most persuasive of these ad-

4. Evidence for a Pattern

ditions have been complexity and mystery. Scott

Hildebrand (1991) proposes that, in his early Prairie

(1993) states that [c]omplexity refers to the amount

house plans, Wright developed a particular spatial

of visual information offered by an environment

configuration to a canonical state (p.19). This

(p.25) and Kaplan (1988) identifies that a scene

so-called Wright pattern (ibid.) comprises 13 ele-

high in mystery is one in which [a person] could

ments, of which 10 are consistently present in all of

learn more if [they] were to proceed farther into the

the Prairie houses, and all 13 in the Heurtley house.

scene (p.50). In a sense, complexity and mystery

However, several of these elements are not, as previ-

are closely related to what might be called the dis-

ously noted, related to combinations of spatial and

coverability of a space. Hildebrand (1991)

incor-

visual characteristics. For example, deep eaves,

porates and partially merges these characteristics

low ceilings, and horizontal exterior massing are all

in his analysis of Wrights work, arguing that the

features that could be measured, but they are not

path from the entry to the living room is an integral

within the scope of the present research. Similarly,

part of the psychological experience of architecture

the relationship between an elevated exterior ter-

because it evokes a sense of mystery and complex-

race and a prominent distant view is also difficult

ity that draws a person though a staged process

to analyse in any consistent way. However, many of

of discovery.

the other characteristics of the Wright pattern are

One of the traditional challenges for prospect-

explicitly concerned with the combined visual and

refuge theory is that past quantitative research has

spatial experience of the path through the house. If


140

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

then, in these limited terms, the Wright pattern does

the building at a height of 1.65 meters above the

exist, the following results might be anticipated from

ground or floor level. This height allows for views

a detailed study of the Prairie houses.

over some fixed elements, such as the bookcases

A gradual shift from being refuge-dominant

in the Cheney house, but not others, such as the

(smaller, enclosed spaces) to being more balanced in terms of prospect and refuge (larger
spaces, more outlook). In the present paper,
isovist area (see discussion in the results section) is used to assess this part of the pattern.
A gradual increase in mystery and complexity
experienced while moving along the circuitous
path (Hildebrand 1991, p.25). At the end of the
path, mystery and complexity should reach a
peak value because the major space is framed
by peripheral interior views to contiguous
spaces seen beyond architectural screening
devices (ibid.). In the present paper, occluded
length (see discussion in the results section) is
used to assess this part of the pattern.
A visitor will be drawn along the path by the

Robie house fireplace. The 1.65m height approximates the eye level of Wright, who had a habit of
calibrating the vertical dimension of his designs to
his own body (Hildebrand, 1991).
The scale of the 100mm isovist grid coupled with
the size of the buildings resulted in over 200,000
observation locations being generated, with each
one providing 11 useful measures; a set of results
which is too great to report in a paper. The data
presented in this research is therefore limited to that
associated with points located along the centreline
of the paths at 500mm increments, approximating
the length of Wrights step. The choice of 500 mm
increments also reduces the amount of manual
data processing required to identify measures from
observation points along the paths, while retaining
an accurate depiction of the spatial experience.

combined phenomenal power of space, form

The five paths chosen for analysis each com-

and vision. There is presumed to be both visual

mence from the front door and trace the shortest

pull and spatial force along the path, much of

route to the living room fireplace, before continuing

which is caused by glimpses of major spaces

to the centre of the living room; a location identified

(ibid., p.75). In the present paper, two values,

by drawing a diagonal line between the rooms

drift direction and drift magnitude (see discus-

corners. If a person must pass through the centre

sion in the results section) are used to assess

of the living room prior to reaching the hearth, the

this part of the pattern.

final step is not included. Hildebrand explicitly


maps multiple paths according to these rules and

5. Method

implies that the path must not commence from

The isovist fields used in this paper were generated

the servants entrance and it must avoid service

from floor plans derived from new three-dimensional

areas, such as kitchens. Furthermore, in accord-

AD models which were prepared in accordance with

ance with Hildebrands diagrams and analysis, all

Wrights final construction drawings (Futagawa and

of the paths follow straight lines of movement with

Pfeiffer, 1987). The floor plans of these houses were

90-degree turns.

analysed in UCL Depthmap (version 10.14.00b)

141

Due to unreliable and inconsistent information

(Turner, 2004) using a grid resolution of 100mm.

about the surrounding context, the primary analysis

All of the houses were analysed simultaneously by

is limited to habitable space. Therefore, the method

placing scaled drawings of each in a single file.

treats all external windows as opaque, and all exter-

Each floor plan depicts a horizontal slice through

nal doors, and doors to dedicated storage areas,

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

as opaque and closed. Storage and service areas

several houses. This procedure takes advantage of

forming an access route to other habitable spaces

the fact that isovist analysis measures only the local

are assumed to possess open doors, while all tim-

properties of the plans. Each 500mm step along the

ber screens are treated as solid, opaque elements.

path yields a discrete set of measures of the isovist

Multi-level buildings present a particular challenge

polygon. Any portions of the floor plan that do not

for isovist field analysis due to the complexity of

define the shape of the polygon are irrelevant to

linking floors across levels in UCL Depthmap. In

that position on the path. Therefore, between steps

order to solve this problem, the creation of modi-

along the path it is possible to substitute portions of

fied sections of floor plans, focused on the stair

one plan, for example the ground floor, for portions

and immediately visible spaces were required for

of another plan, such as the first floor. This proce-

Figure 2:
Example of hybrid
floor plan construction for multi-floor
isovist analysis.

142

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

dure allowed for the documentation of a seamless

a more intuitive understanding of the spatio-visual

transition of isovist measures between separate

characteristics of the building.

levels but excludes the use of global measures

While an isovist analysis of the external spaces

(such as visual integration) which are more difficult

surrounding each house is outside the scope of

to reliably generate using these methods.

the present research, Hildebrand argues that the

As an example of this method of constructing

exterior paths (or driveways) that must be taken to

isovists across levels, consider the path through

reach the front door are also significant. Thus, these

the Robie house (Figure 2). At the start of the path,

routes are also recorded in the present research and

isovist measures are derived from the standard

a simple discussion of path length and number and

ground floor plan (Plan A) until a point where the

direction of turns is provided.

fireplace is no longer visible and the viewer is close


to the turn in the path towards the base of the stairs.

6. Results

At this point the isovist will yield identical proper-

Isovist field analysis produces 11 useful measures

ties if constructed on either Plan A or a hybrid Plan

of the spatio-visual geometry present at each step

B, which includes the base of the stairs. From this

along the selected path. While there is debate about

position along the path to the top of the first run of

the precise spatial qualities that each measure rep-

stairs, isovist properties are only extracted from Plan

resents (Stamps III, 2005), the four most significant

B as this depicts the ground floor and the staircase.

measures, in terms of their capacity to directly rep-

The top of the first run of stairs (isovist on Plan B) is

resent prospect-refuge characteristics, are isovist

the final position from which the entire ground floor

area, occluded length, drift magnitude and drift

is visible; advancing 500mm along the path (to the

direction (Dawes and Ostwald, 2013).

isovist on Plan C) brings the first floor into view, while

Isovist area is the volume of space visible from a

hiding portions of the ground floor, requiring a hybrid


plan to depict the space visible on both levels. Advancing another 500mm along the path (isovist on
Plan D) will hide the ground floor completely from
view; isovist properties are now derived only from
the first floor plan.
Because the paths through Wrights buildings
are considered a continuous spatial experience,
from initial entry to arrival at the centre of the living
area, they allow for the creation of graphs depicting
the changing isovist data from every step along the
path. In the reporting of these results, the data points
are logged sequentially such that the horizontal
axis of each graph in this paper represents the

to the measure of prospect (large areas) and to


a lesser degree, refuge (small areas). The visible
area also reflects the potential volume of visual
information available, and thus the potential
complexity of the view.
Occluded length is the portion of the isovist
perimeter representing the obscured or hidden
parts of the building - a property that relates
closely to the concept of mystery and, to a lesser
extent, complexity (Kaplan, 1987).
Drift magnitude is the distance between the
viewers location and the centre of the isovist

distance the observer has travelled along the path.

polygon. This measure describes the strength of

The left side of the graph corresponds to the front

the visual pull; or force of movement (Hoffman,

door and the right hand side to the end of the path

1995, p.X) that draws a visitor through Wrights

in the living room. Presentation of the data in this

architecture.

manner, as opposed to being tabulated, allows for


143

particular location. It provides a close correlation

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

Drift direction is the angle of a line connecting


the observation point to the centre of the isovist

type of phenomenological account that is common


in Wrightian scholarship.

polygon and measures the direction of the visual


pull. This measure directly reflects Hildebrands

Henderson House (1901)

(1991) notion of being drawn through space and

The Henderson house in Elmhurst, Illinois, repre-

the concept that Hoffman (1995, p.X) labels as

sents an early prototype of the Prairie style. The

the visual dynamic in Wrights architecture.

house exhibits a number of Hildebrands list of


characteristic Wrightian phenomenological features,

These four measures, when considered in

but fewer than are found in some later works. In

combination, can be used to confirm or challenge

this instance the living, dining and reception areas

Hildebrands suggestion that Wrights paths follow

share a raised ceiling and the fireplace is located

a discernible, prospect-refuge related, spatio-visual

opposite a glazed wall, beyond which a terrace

pattern. Two charts per house document these four

completes the truncated cruciform plan. The exter-

measures. In the first of these, the Y-axis records

nal approach path to the house follows a straight

isovist area and occluded length, with the former

line from the street, passing under the wide eaves

measured in square metres and the latter in metres.

of the hipped roof and rising to the terrace and front

The second chart for each house records drift mag-

door. This terrace functions as a traditional porch,

nitude, in meters, and drift direction, as the angular

unlike those in the other Prairie houses, which Wright

difference between the direction of travel and the

often isolated from the approach path. The route

direction of visual pull. As such, a result of +90 de-

into and through the house is complicated by a pair

grees represents a visual pull perpendicular and to

of dogleg manoeuvres that are required to pass

the left hand side of the direction of travel. A result

through the front door and entry hall. Once inside,

of -170 degrees represents a visual pull to the right

the route to the centre of the living room requires

hand side that is almost opposite to the direction

only three direction changes (Figure 3).

of travel. A drift magnitude of less than two metres

The data derived from the Henderson house

could be regarded as being barely noticeable,

show that upon emerging from the entry hall (A),

lengths of between two and five metres are notice-

the visitor experiences a minor but sudden increase

able and will draw the eye, and anything over five

in prospect (Figure 4) and an intense visual pull

could be considered as significant. The X-axis, in

directing attention left and toward the living room

both types of chart, records progress along a path

(Figure 5). Proceeding along the path toward the

from the front door to the centre of the living room

living room threshold (B), both prospect and mystery

in 500mm increments. Specific important locations

indicators gradually rise as more of the house is

along the path are also keyed to an axonometric

both revealed and suggested. The most mysterious

diagram of each house, and to a description of the

position coincides with a low visual pull, the centre

experience of a person attuned to the properties

of attention and the intersection of paths leading

of the isovists. In this description, to better allow

to the living room, kitchen and entry. Prospect is

for comparison with past claims about the impact

highest only after crossing the threshold to enter the

of Wrights architecture, the words prospect, mys-

living room (B), where prospect, mystery and visual

tery and pull are generally used in place of their

pull decrease as the hearth is passed (C), and the

mathematical counterparts. Therefore, this descrip-

visitor moves toward the centre of the room. Along

tion offers a mathematically informed version of the

almost the entire length of this path, the drift offset


144

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

Figure 3:
Henderson house,

axonometric showing
movement path.

Figure 4:
Henderson house,

isovist area and


occlusivity.

145

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

Figure 5:
Henderson house,

isovist drift properties.

is less than 90 degrees, indicating that visual pull

external approach path to the house shows the early

is drawing the occupant toward the living room.

signs of deliberate complexity through the inclusion

The only exception occurs where the visitor finally

of two, redundant, 90-degree turns. Internally the

reaches the centre of the living room and a visual

route is even more complex; three 90-degree turns

pull of negligible strength directs attention back to

separate the front door and the base of the stairs,

the hearth. High visual pull at the start of the path,

two further turns are encountered while ascending

coupled with increasing levels of information and

the stairs and two more before reaching the hearth.

mystery suggest that the experience of the Hender-

A final turn is required to reach the centre of the

son house is that of a journey of discovery; albeit a

living room (Figure 6).

minor one. Upon reaching the living room centre,

Like the Henderson house, entry to the Heurtley

further movement yields little additional information

house is via an enclosed hall that obscures the rest

and the low levels of visual pull and mystery discour-

of the building. The isovist field data (Figures 7 and

age such movement.

8) demonstrate that as the occupant moves beyond


this point (A), both prospect and mystery rise until

Heurtley house (1902)

reaching the stairs (B), where both values suddenly

The Heurtley house in Oak Park is the first of the

fall again as vision is restricted. Walking toward the

Prairie houses to display all 13 of Hildebrands major

stairs also requires the occupant to move away from

prospect-refuge features. This design departs from

the playroom, resisting the strongest visual pull

contemporary planning conventions by locating the

experienced along the path. Visual pull remains

living and dining rooms on the first floor in order to

directed behind the occupant while ascending the

capture elevated views of the neighbourhood. The

stairs. It is only after reaching the landing (C) that


146

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

Figure 6:
Heurtley house,

axonometric showing
movement path.

147

prospect and mystery values rise again. This cor-

attention toward the dining room, living room, and

responds to a sharp increase in visual pull, drawing

finally back to the dining room. Once more, the liv-

the occupant toward the living and dining rooms

ing room centre is the location of highest prospect;

(D). Prospect continues to increase in steps while

however, this is coupled with a significant degree of

approaching the living room centre and mystery

mystery and mild visual pull toward the dining room,

peaks immediately before reaching the hearth (E).

suggesting that this is a more dynamic space than

Along the latter part of this route, visual pull directs

the living room of the Henderson house.

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

Figure 7:
Heurtley house,

isovist area and


occlusivity.

Figure 8:
Heurtley house,

isovist drift properties.

Cheney House (1903)

walkways), masking (the screening of view from

In the Cheney house, Wright offers a deliberately

the street), and convolution (p.38). Traversing the

complicated approach path. Hildebrand (1991)

correct exterior entry path requires a person to as-

argues that in this design the glass and overhangs

cend several groups of stairs, and two unnecessary

may suggest penetrability, but the house protects

90-degree turns (Figure 9, locations marked 1 and 2

its actual access through ambiguity (the dual

in the axonometric diagram of the movement path).


148

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

Figure 9:
Cheney house,

axonometric showing
movement path.

149

This elaborate exterior route leads the occupant to

high enough to obscure the view of an occupant

the entry of the house and, once inside, another

of Wrights stature.

turn (A) is all that separates the entrance from the

The Cheney house possesses the shortest of

living room threshold (B), the hearth (C) and the

all the internal paths in this study and the most

centre of the room (Figure 9). The absence of a

elaborate external path. The isovist analysis shows

second story in the house allowed Wright to raise

that a person moving toward the living room experi-

the ceilings of the library, living and dining rooms,

ences steadily increasing levels of prospect (Figure

and the hearth is again opposite a glazed wall. The

10). In contrast, mystery tends to decrease from

bookcases defining the living room and adjacent

the beginning of the dogleg (A) until reaching a

to the fireplace, approximately 1.5m tall, are not

position halfway between the living room threshold

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

Figure 10:
Cheney house,

isovist area and


occlusivity.

Figure 11:
Cheney house,

isovist drift properties.

(B) and the hearth (C). This peak corresponds to a

room except when crossing the threshold to this

glimpsed view of the fifth bedroom over the book-

room (B); here, the direction of pull aligns with the

case adjacent to the fire. Mystery then declines until

direction of travel. The end of the path offers high

the occupant approaches the living room centre.

prospect and mystery, with little visual pull sug-

Visual pull is high and forward close to the front

gesting an experience that is more static than that

door, drawing the occupant past the library and

of the Heurtley house and with more mystery than

toward the living room (Figure 11). The direction of

the Henderson house.

visual pull remains fixed on the centre of the living


150

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

Evans house (1908)


Wright designed the Evans residence as a fireproof
house for $5000 (MacCormac, 2005, p.143). The
house sits atop a low hill and its approach is via a
simple journey up a long driveway (Figure 12, path
location marked 1 in the axonometric) to the carport. Here, under the protection of the roof (Figure
12, location 2), a single 90-degree turn and a flight
of stairs bring a visitor to the front door. Inside the
entry, with its dropped ceiling, the living room and

Figure 12:
Evans house,

axonometric showing
movement path.

151

dining room form a semi-continuous space with the


fireplace located at the centre of the house, again
opposite glazing. A single turn (B) takes the occupant to the hearth and thereafter one additional
turn (C) is needed to reach the centre of the room
(Figure 12).
Entry to the Evans house is via a semi-enclosed
foyer where visual pull directs a persons attention
to the living room, then toward the kitchen hall as
this gradually enters into view. The first appearance

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

of this corridor (A) causes a spike in mystery and

occupant steps past the centre of the visible area,

contributes to the rapid increase in prospect that

pulling attention behind the visitor, but with negligi-

peaks at the living room threshold (B) (Figure 13).

ble strength (Figure 14). The living room centre is

Upon entering the living room, visual pull focuses

very much like that of the Henderson house, offer-

attention on the centre of the living room while

ing high prospect and little mystery or visual pull to

prospect decreases near the hearth (C). Like the

encourage further movement.

Henderson house, visual pull decreases until the

Figure 13:
Evans house,

isovist area and


occlusivity.

Figure 14:
Evans house,

isovist drift properties.

152

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

Robie house (1910)


The Robie house exterior approach is relatively
straightforward, and like that of the Evans house,
likely due to the exigencies of vehicular access.
However, unlike the Evans house, the internal path
of the Robie house is complex (Figure 15). It commences with a long straight path linking the main
entrance to the stairs. Thereafter the stairs form a
vertical transition, spiralling up through 360 degrees
to the first floor to cross the lower level path. A further
straight path leads to the living room, where another
turn is required to reach its centre, before a final turn

Figure 15:
Robie house,

axonometric showing
movement path.

153

is taken to approach the hearth. There is no direct


route from the stairs to the hearth, so the occupant
must move through the centre of the living room to
reach it (this is the reverse of the pattern in the other
four houses). Interestingly, this path through the
Robie house makes exclusive use of left hand turns.
When entering the Robie house, the isovist
data suggests that the occupant immediately experiences an extraordinarily high, and thereafter
quickly diminishing visual pull, drawing one to the
base of the stairs (A) and past the point of maximum
prospect. Mystery remains stable until the occupant

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

reaches the stairs, before dropping dramatically

prospect and mystery to fall sharply, until emerging

(Figures 16 and 17). Ascending the stairs requires

onto the first floor (C). At the top of the stairs, the

the person to move away from the centre of visible

strength of visual pull (left bias) is negligible while

space, causing a rearward visual pull of increas-

suggesting a relatively neutral decision point with a

ing strength until capturing a final glimpse of the

high degree of prospect. This prospect diminishes

lower level while traversing the landing (B). The

through the living room threshold (D), increasing

constrained experience of the stairs causes both

as the living room enters sight and before views

Figure 16:
Robie house,

isovist area and


occlusivity.

Figure 17:
Robie house,

isovist drift properties.

154

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

back down the hall are obscured, after turning (E)

few, if any, clear trends. Moreover, those instances

toward the living room centre (F). Here visual pull

where trends could be said to occur, can readily

reaches its lowest level and moving toward the

be explained as by-products of the way in which

hearth requires resisting an increasing pull toward

Hildebrand constructed his paths; a process repli-

the centre of the living room. Again, the centre of the

cated in this paper.

living room provides high prospect and little mystery


or visual pull to encourage further movement.

Starting with prospect related spatio-visual


measures, four of the houses possess a lower level

The Robie house has a similar vertical transition

of prospect at the point of entry and a higher level

to the Heurtley house; however, unlike the Heurtley,

at the centre of the living room (Figure 18). While

the Robie utilises visual pull and prospect to deliver

this result superficially supports the existence of

the occupant directly to the stairs. The stairs remain

Hildebrands pattern, this trend would be antici-

a significant transition point after which visual pull

pated in any family house large enough to have a

remains centred on the staircase, providing no

separate entry and living area. In essence, any route

guidance and confusing the journey. It is only upon

that starts in a smaller space and ends in a larger

entering the living room, a threshold of several steps,

one is likely to show an increase in prospect over

that visual pull again begins to direct the visitor

the journey. Therefore, this result could readily be

forward to the centre of the room. In summary, the

explained as an artefact of Hildebrands method and

journey through the Robie house is one of strong

not as something specific to Wrights architecture.

guidance through the lower level and weak guid-

There is no overarching pattern in the results

ance through the upper floor.

for mystery found in the five paths. Staircase zones


tend to possess the lowest prospect and the least

Figure 18:
Linear trend lines
(least squares
method) of the isovist
area values of each
house. Solid lines
show the relative
length of each path,
dotted lines represent
a continuation of the
trend beyond the end
of a path.

155

7. Discussion

mystery because such tight spaces not only restrict

When the data developed from the isovist field

vision, but also restrict a sense (suggested through

analysis for each house is compared, there are

occlusivity) of what lies just out of sight. The paths

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

Figure 19:
Linear trend lines
(least squares
method) of the isovist
occlusivity values of
each house. Solid
lines show the relative
length of each path,
dotted lines represent
a continuation of the
trend beyond the end
of a path.

through the Robie and Cheney houses commence

the suggestion of a drift offset pattern in the data.

with high levels of mystery, while this value is low at

The final critical measure, drift magnitude, has a low

the entry of the remaining houses (Figure 19). The

level of correspondence across the houses, with a

Robie, Heurtley and Henderson houses show minor

minor tendency to peak at or near the entrance and

peaks in mystery corresponding to the threshold of

thereafter dropping along the length of the path.

the living room but few other similarities exist. Mys-

Staircases again cause these values to fluctuate,

tery is also low at the end of the journey through the

spiking in the Robie house and falling in the Heu-

Robie, Henderson and Evans houses, but high at a

rtley. Therefore, the final anticipated result for the

similar location in the Cheney and Heurtley houses.

existence of the Wright pattern is also, on balance,

Thus, only the results of two of the five houses con-

not reflected in the data.

form to the anticipated pattern.


The results for drift offset display a seemingly

8. Conclusion

higher degree of correlation between the houses.

In the specific case of Hildebrands reading of

For example, the Henderson, Cheney and Evans

Wrights Prairie houses, and based on a detailed

houses all have paths that are dominated by forward

analysis of five major works from this set, there are

visual pull, as does the path through the Robie

insufficient similarities to support the claim that there

house, if its staircase and a few other minor spatial

are any underlying spatio-visual patterns. Two of the

disjunctions are ignored. Ironically, the exception to

four measures suggest a minor level of correlation

this rule, the Heurtley house, is the one that Hilde-

with the results anticipated in the Wright pattern,

brand argues most fully represents the Wrightian

but these are readily explained in other ways. The

pattern. The Heurtley house has many instances

other two results neither support the existence of

where the drift offset is over 90 degrees, meaning

the Wright pattern, nor demonstrate any consistency

the visual pull is against the direction of travel. Fur-

between the houses.

thermore, as was the case with the prospect data, it

Despite these results, it must be acknowledged

might be expected that any large house would have

that the key spatial experience of Wrights architec-

a strong drift from the smallest to the largest spaces;

ture may exist primarily in the third dimension and

a methodological anomaly that could account for

simply cannot be captured using two-dimensional


156

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

isovists. Hildebrand even argues that Wrights

About the authors:


Prof. Michael J. Ostwald

(michael.ostwald@newcastle.edu.au), is Dean of Architecture at the University


of Newcastle, Australia and
a Visiting Professor at RMIT
University (Melbourne). He
has a PhD in architectural
history and theory and a
higher doctorate (DSc) in
the mathematics of design.
He has lectured in Asia,
Europe and North America
and has written and published extensively on the
relationship between architecture, philosophy and
geometry. Michael Ostwald
is a member of the editorial
boards of ARQ, the Nexus
Network Journal (Architecture and Mathematics)
and Architectural Theory
Review.

Michael Dawes

(michael.dawes@newcastle.edu.au) is a graduate
architect who completed
undergraduate research
on Space Syntax methods,
and is currently a research
assistant on a major ARC
grant undertaking mathematical and computational
modeling and analysis of
twentieth century houses.

manipulation of ceiling height may have a strong


impact on the prospect-refuge characteristics of the
architecture. However, within the limits of the present
method, this research demonstrates that, in terms
of the four primary mathematical determinants of
prospect-refuge qualities in architecture, there is no
evidence for a foundational pattern in five of Wrights
most important Prairie house designs.

Acknowledgements
An ARC Fellowship (FT0991309) and an ARC Discovery
Grant (DP1094154) supported this research.

References
Appleton, J. (1975), The Experience of Landscape, London: John Wiley and Sons.
Assefa, E. M. (2003), Inside and outside in Wrights Fallingwater and Aaltos Villa Mairea. In: Environmental
& Architectural Phenomenology Newsletter, Vol. 14,
p.11-15.
Batty, M. (2001), Exploring isovist fields: Space and shape
in architectural and urban morphology. In: Environment and Planning B: Planning and Design, Vol. 28,
p.123-150.
Benedikt, M. L. (1979), To take hold of space: Isovists and
isovist view fields. In: Environment and Planning B:
Planning and Design, Vol. 6, p.47-65.
Christenson, M. (2010), Registering visual permeability in
architecture: Isovists and occlusion maps in AutoLISP.
In: Environment and Planning B: Planning and Design,
Vol. 37, p.1128-1136.
Conroy, R. (2001), Spatial Navigation in Immersive Virtual
Environments. Unpublished PhD Thesis, London: University of London.
Davis, L. S. and Benedikt, M. L. (1979), Computational
models of space: Isovists and isovist fields. In: Computer Graphics and Image Processing, Vol. 11, p.49-72.

157

The Journal of
Space Syntax

S
S

Volume 4 Issue 1

Dawes, M. J. and Ostwald, M., J. (2013), Using isovists

L. (ed.), Environmental aesthetics: theory, research

to analyse prospect-refuge theory: Measuring spatio-

and applications, Cambridge: Cambridge University

visual relations in Wrights Heurtley House. In: The


International Journal of the Constructed Environment.
Vol. 3 (1), p. 25-40.
Ellard, C. (2009), You are here, New York: Anchor Books.

Press, p.379-389.
Kaplan, R. and Kaplan, S. (1989), The experience of
nature: A psychological perspective, Cambridge:
Cambridge University Press.

Falk, J. H. and Balling, J. D. (2009), Evolutionary influence

Kaplan, S. and Kaplan, R. (1983), Cognition and environ-

on human landscape preference. In: Environment and

ment: Functioning in an uncertain world, New York:

Behavior, Vol. 42, p.479-493.

Preager.

Franz, G., Von Der Heyde, M. and Bulthoff, H. (2004),

Koning, H. and Eizenberg, J. (1981), The language of

Predicting experiential qualities of architecture by its

the prairie: Frank Lloyd Wrights prairie houses. In:

spatial properties. In: Proceedings of the 18th IAPS-

Environment and Planning B: Planning and Design,

Conference, Vienna, p.1-10.


Futagawa, Y. and Pfeiffer, B. (eds.) (1987), Frank Lloyd
Wright Monograph 1907-1913, Tokyo: A.D.A. Edita.
Gibson, J. J. (1947), Motion picture testing and research.

Vol. 8, p.295323.
Laseau, P. and Tice, J. (1992), Frank Lloyd Wright: Between Principle and Form, New York: Van Nostrand
Reinhold.

(Report No.7. A.A.F. Aviation Psycology Research

Lin, Y. (1991), Karsten Harries natural symbols and Frank

Reports). Washington D.C.: U.S. Government, Print-

Lloyd Wrights natural houses. In: Environmental & Ar-

ing Office.
Gibson, J. J. (1966), The Senses Considered as Perceptual
Systems, Boston, Houghton Mifflin Company.
Gibson, J. J. (1979), The Ecological Approach to Visual
Perception, Boston: Houghton Mifflin Company.
Hale, J. (2000), Building Ideas: An Introduction to Architectural Theory, New York: John Wiley and Sons.
Harries, K. (1997), The Ethical Function of Architecture,
Cambridge, MA: MIT Press.
Hienz, T. (2006), The Vision of Frank Lloyd Wright, Edison,
NJ: Chartwell Books.
Hildebrand, G. (1991), The Wright Space: Patterns &
Meaning in Frank Lloyd Wrights Houses, Seattle:
University of Washington Press.
Hoffman, D. (1995), Understanding Frank Lloyd Wrights
Architecture, New York: Dover Publications.

chitectural Phenomenology Newsletter, Vol. 2, p.13-15.


Lind, C. (1994), Frank Lloyd Wrights Prairie houses, California: Archetype Press.
MacCormac, R. (2005), Form an philosophy Froebels
Kindergarten training and Wrights early work. In:
McCarter, R. (ed.), On and By Frank Lloyd Wright: A
Primer of Architectural Principals, New York: Phaidon
Press Limited, p.124-143.
Markhede, H. and Koch, D. (2007), Positioning analysis:
Social structures in configurative modelling. In: Kubat,
A.S. et al. (eds.), Proceedings of the Sixth International
Space Syntax Symposium, Istanbul: I.T.U. Faculty of
Architecture, p. 069.1-069.14.
Meilinger, T., Franz, G. and Bulthoff, H. (2009), From
isovists via mental representations to behaviour: First
steps toward closing the causal chain. In: Environ-

Kaplan, S. (1987), Aesthetics, affect and cognition: Envi-

ment and Planning B: Planning and Design, Online

ronmental preference from an evolutionary perspec-

publication, p.1-16. Also in: Environment and Planning

tive. In: Environment and Behavior, Vol. 19, p.3-31.

B: Planning and Design (2012), Vol. 39 (1), p.48-62.

Kaplan, S. (1988), Perception and landscape: Conceptions

Morello, E. and Ratti, C. (2009), A digital image of the

and misconeptions. In: Naser J. L. (ed.), Environ-

city: 3D isovists in Lynchs urban analysis. In: Envi-

mental aesthetics: theory, research and applications,

ronment and Planning B: Planning and Design, Vol.

Cambridge: Cambridge University Press, p.45-54.

36, p.837-853.

Kaplan, R. and Herbert, E. J. (1988), Familiarity and

Nasar, J. L. (1984), Visual preferences in urban street

preference: A cross cultural analysis. In: Naser J.

scenes: A cross-cultural comparison between Japan

158

S
S

Prospect-refuge patterns
in Frank Lloyd Wrights Prairie houses
Ostwald, M. J. & Dawes, M.

and the United States. In: Journal of Cross-Cultural


Psycology, Vol.15, p.79-93.
Norberg-Schulz, C. (1979), Genius loci: towards a phenomenology of architecture, New York: Rizzoli.
Ostwald, M. J. and Dawes, M. J. (2013), Using isovists to

ability theory. In: Environment and Planning B: Planning and Design, Vol. 32, p.735-762.
Stamps III, A. E. (2006), Interior prospect and refuge.
In: Perceptual and Motor Skills, Vol. 103, p.643-653.

analyse architecture: Methodological considerations

Stavroulaki, G. and Peponis, J. (2005), Seen in a different

and new approaches. In: The International Journal

light. Icons in Byzantine museums and churches. In:

of the Constructed Environment, Vol. 3 (1), p. 85-106.

van Ness, A. (ed.), Proceedings of the Fifth Interna-

Ostwald, M. J. and Vaughan, J. (2010), The mathematics

tional Space Syntax Symposium, Delft: University of

of style in the architecture of Frank Lloyd Wright: A

Technology, Vol. 1, p.251-263.

computational, fractal analysis of formal complexity in

Turner, A. (2004), Depthmap 4. A Researchers Handbook.

fifteen domestic designs. In: Geller, P. S. (ed.), Built

London: Bartlett School of graduate studies, University

Environment: Design, Management and Applications,


New York: Nova Scientific, p.63-88.
Peponis, J. and Bellal, T. (2010), Fallingwater: The inter-

College London.
Twombly, R. C. (1979), Frank Lloyd Wright: His Life and
His Architecture, New York: John Wiley and Sons.

play between space and shape. In: Environment and

Tzortzi, K. (2004), Building and exhibition layout: Sainsbury

Planning B: Planning and Design, Vol. 37, p.982-1001.

Wing compared with Castelvecchio. In: Architectural

Peponis, J., Conroy-Dalton, R., Wineman, J. and Dalton, N.


(2004), Measuring the effects of layout upon visitors
spatial behaviors in open plan exhibition settings. In:
Environment and Planning B: Planning and Design,
Vol. 31, p.453-473.

Research Quarterly, Vol. 8, p.128-140.


Unwin, S. (2010), Twenty Buildings Every Architect Should
Understand, London: Routledge.
Vaughan, J. and Ostwald, M. J. (2011), The relationship
between the fractal dimension of plans and elevations

Psarra, S. (2005), Spatial culture, way-finding and the

in the architecture of Frank Lloyd Wright: Comparing

educational message. In: Macleod, S. (ed.), Reshap-

the prairie style, textile block and Usonian Periods.

ing Museum Space: Architecture, Design, Exhibitions,


London: Routledge, p.78-94
Psarra, S. (2009), Architecture and Narrative. The Formation of Space and Cultural Meaning, London:
Routledge.

In: Architectural Science Research, Vol. 4, p.21-44.


Weston, R. (2002), Utzon - Inspiration, Vision, Architecture,
Copenhagen: Blndal.
Wiener, J. and Franz, G. (2005), Isovists as a means to
predict spatial experience and behavior. In: Freksa,

Roberts, J. W. (2003), Prospect-refuge theory and Alvar

C., Knauff, M., Krieg-Brckner, B., Nebel, B. and

Aaltos experimental house at Muuratsalo: What is it

Barkowsky, T. (eds.), Spatial Cognition IV. Reason-

that we like about Aaltos site-related architecture, and

ing, Action, Interaction, Berlin / Heidelberg: Springer,

why?. In: Progress: The Proceedings of the Twentieth Annual Conference of the Society of Architectural
Historians, Sydney: SAHANZ, p.236-268.
Scott, S. (1993), Complexity and mystery as predictors
of interior preferences. In: Journal of Interior Design,
Vol. 19, p.25-33.

p.42-57.
Yang, P. P.-J., Putra, S. Y. and Li, W. (2007), Viewsphere:
A GIS-based 3D visibility analysis for urban design
evaluation. In: Environment and Planning B: Planning
and Design, Vol. 34, p.971-992.
Zamani, P. and Peponis, J. (2013), Co-visibility and peda-

Seamon, D. (1992), Review of The Wright Space: Pat-

gogy: Innovation and challenge at the High Museum of

terns and Meaning in Frank Lloyd Wrights Houses

Art. In: The Journal of Architecture, Vol. 15, p.853-879.

by Grant Hildebrand. In: Great Plains Research: A


Journal of Natural and Social Sciences, Paper 59, Vol.
2 (1), p.131-134.

159

Stamps III, A. E. (2005), Isovists, enclosure, and perme-

You might also like