Point Line Plane

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/279819931
Point, line, plane : basic elements of formal composition in Bauhaus and shape
computation theories
Article · December 2011
CITATIONS READS
0 5,496
1 author:
Sotirios Kotsopoulos
National Technical University of Athens
44 PUBLICATIONS 116 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
The Connected Sustainable Home View project
All content following this page was uploaded by Sotirios Kotsopoulos on 10 October 2015.
The user has requested enhancement of the downloaded file.

1
Point, Line, Plane:

Basic Elements of Formal Composition in Bauhaus and Shape Computation Theories
By
Sotirios D. Kotsopoulos
Master of Architecture
Southern California Institute of Architecture, 1995
SUBMITTED TO THE DEPARTMENT OF ARCHITECTURE IN PARTIAL FULFILLMENT OF THE

REQUIRMENTS FOR THE DEGREE OF
MASTER OF SCIENCE IN ARCHITECTURE STUDIES

AT THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
JUNE 2000
c 2000 Sotirios D. Kotsopoulos. All rights reserved.
The author hereby grants to MIT the permission to reproduce and to distribute publicly paper
and electronic copies of this thesis document in whole or in part.
Signature of Author:
___________________________________________________________________________
Department of Architecture
May 18, 2000
Certified by:
___________________________________________________________________________
Terry Knight
Associate Professor of Design and Computation
Thesis Supervisor
Accepted by:
__________________________________________________________________________
Roy Strickland
Principal Research Scientist in Architecture
Chairman, Committee for Graduate Students
2
Readers:
George Stiny
Professor of Design and Computation
William J. Mitchell
Professor of Architecture and Media Arts and Sciences
Dean, School of Architecture and Planning
3
Point, Line, Plane:

Basic Elements of Formal Composition in Bauhaus and Shape Computation Theories
By
Sotirios D. Kotsopoulos
Submitted to the Department of Architecture on May 18, 2000 in Partial Fulfillment of the
Requirements of the Degree of Master of Science in Architecture Studies
ABSTRACT
Architecture is not representational. It does not stand for something else. However, the process
of its formation is inclusively dependent upon a series of dynamic graphic calculations that
result into a series of spatial descriptions. This process can be equated to a non-linear se-
quence of computations with points, lines, planes, and solids, on the plane and in physical
space. This study examines the functional and perceptual properties of points, lines and planes.
How do basic elements behave in formal composition, and how do computations of form affect
basic elements? The context of the study is composite. Shape computation theory that involves
algebras of basic elements and shape rules provides a flexible and expressive computational
apparatus, while the systematic approach of the Bauhaus on nonrepresentational composition,
and the theories of P. Klee and W. Kandinsky in particular, provide artistic insight at a percep-
tual and interpretational level.
Thesis Supervisor: Terry Knight

Title: Associate Professor of Design and Computation
4
5
to Maria
6
ACKNOWLEDGEMENTS
My thanks go to Professor Terry Knight who has been a source of ideas and support to me,
and to my readers Professor George Stiny and Professor William J.Mitchell. This work has been
made possible, in part, with the help of Greek State Scholarship Foundation [S.S.F.].
7
CONTENTS
INTRODUCTION
I. POINT
A. Mathematical concept
B. Graphical concept
C. Dictionary
1. Appearance
1.1 Form
1.2 Texture
1.3 Size and relation to other forms
D. Composition
1. Centric structure
2. Eccentric structure
3. Quantitative increase
E. Points in architecture
II. LINE
C. Dictionary
1. Straight line
1.1 Typical straight lines
1.2 Atypical straight line
2. Angular and curved lines
2.1 Angular line
2.1.1 Angles and lengths
2.1.2 Compositional parameters of angular lines
2.2 Curved line
D. Composition
1. Complexes of lines
2. Lines angles and shapes
3. Linear, medial, and planar shapes
4. Divisions of shapes
E. Graphic properties
1. Boundaries and texture
2. Emphasis
3. Transition to plane
8
III. PLANE
C. Dictionary
1. Planar forms
1.1 Interaction of planes
1.2 Actual and perspective planes
1.3 Activation of planes
2. Graphic context
D. Basic elements on the graphic context
1. Line on plane - Relation to the boundaries
2. Plane on plane
CONCLUSIONS
REFERENCES
9
INTRODUCTION
Architecture is not representational. It does not stand instead of something else. However, the
process of its formation is inclusively depended upon a series of dynamic graphic calculations
that result to a series of spatial descriptions. This process can be equated to a non-linear
sequence of computations with points, lines, planes, and solids, on the plane, and the physical
space. Basic elements such as points, lines, and planes, arranged on the plane, compose two-
dimensional drawings, while planes and solids arranged on the physical space, compose three-
dimensional models. Basic elements and their combinations have a profound value. They are
abstract and inclusive. Without intrinsic structure, they allow us to attribute structure and
meaning, as we want. In design, dynamic figuration proceeds from the general and abstract to
the particular and concrete. This study examines the functional and perceptual properties of
points, lines and planes. How do basic elements behave in formal composition, and how do
computations of form affect basic elements? The context of the study is composite. Shape
computation theory that involves algebras of basic elements and shape rules provides a flexible
and expressive computational apparatus, while the systematic approach of the Bauhaus on non
representational composition, and the theories of P. Klee and W. Kandinsky in particular, pro-
vide artistic insight at a perceptual and interpretational level. The reason for attempting this
linking is threefold. First, because all three approaches are systematic and focus on non-
representational composition. Second, because they are not concerned with the artefact as a
material object, but with the manner it is produced, not with form as an immutable value, but with
formation as a process. Third, because the examination of the properties of basic elements is a
step towards the essential, towards the functional as opposed to the impressional. We learn to
look down on formalism, and to avoid taking over finished products. This should prevent us from
regarding artefacts or designs as something rigid, fixed, or unchanging. After all, the world in its
present configuration is not the only world possible!
My interest in basic elements begins from the observation that verbal description is not enough
to explain architectural form and the process of formation. Objects demand visual description.
However, visual description as fixed representation does not say enough either. If we are
interested in form we need to approach it dynamically, as a process. Architecture does not build
images, and moreover it does not need to produce structures in agreement to a predefined
mode of expression. It introduces new instances into an already existent language, the architec-
tural language. This becomes evident in the experience of a significant work, which usually
resembles a discovery, it is sudden and unforeseen, and it is not immediately available for
description. In both architectural and computational terms it is not interesting to have to deal
with fixed descriptions. Therefore we need to develop a flexible and dynamic computational
10
process. The process is computationally ambiguous since there is no golden rule to predefine
what is better, but only the empirical flux of things. Basic graphic elements in their abstraction
constitute the elementary visual qualities that we use in this abstract process. Points, lines, and
planes manipulated on the plane become immaterial expressions of our spatial ideas. All basic
elements and their arrangements may be approached in unanticipated ways. What becomes
essential is that graphic calculations follow our perception, independently from what this may
cause to description.
T. Ando
The idea of an abstract process of composition in the arts is old. Aristotle (Poetic 1459 A 19)
referring to the process of synthesis, notes that “the composition does not have to be similar
with something existing in the real world, but it should be approached as an individual singular
action that completes itself”. In the first half of 15 th century Alberti introduced a fundamental, as
well as useful, abstraction. He distinguished the conceptual process of design of an artefact
from its material manifestation. Alberti warns that the architect must know exactly what to do
before construction can begin. The building must already be fully complete in his mind. Accord-
ingly he defines design as “a firm and graceful preordering of the lines and angles, conceived in
the mind, and contrived by an ingenious artist” (1955, p. 1-2).
Paul Valery, (1924, p. 58) observed a similarity between architecture and music: both are
abstract and non-representational, because they are making us see things other than them-
selves. They do not imitate but they force matter to take conceptual forms. They give form to
rules and they take form by rules. The explanation to this mystery can be found in their affinity
to those half –abstract, half-concrete beings we call shapes. According to Valery “shapes have
impact in vision, in touch, in hearing but also in thought, and in number”. He defines shapes as
combinations of lines, where lines are the image of an “unreasonable oddity” without beginning,
without end, and without any other meaning except that of freedom. The architect organizes
lines in shapes geometrical or not, by using rules. Finally the rules are revealed to the viewer
through form. For the sense of harmony that architecture causes to the viewer, the viewer
responds actively with an infinite perceptual multiplicity that is constructed by the eye, without
the slightest effort. This supreme delight is forced by the awareness of the calculated shapes,
and the organized dimensions that force perception to build infinite imaginary relations.
11
Klee (1956, p. 24) declared that “art does not reproduce the visible but makes visible”. For Klee
“the very nature of graphic art lures us to abstraction”. Combinations of graphic elements
produce forms that express the schematic quality of the imaginary with great precision. The
more emphasis the graphic work puts on the basic formal elements, the less well suited is to
representation. He sets a distinction between basic visual elements, without which a work in
any art cannot come into existence, and secondary elements. Formal elements of graphic art
are “points, and linear, planar, and spatial energies”.
third
white (light grey) white
second
(dark grey)
first
white (black) white
Klee “Architecture” Compositional rule for the contrast-relation (Klee)
The composition of elements proceeds according to rules. Klee rejects “prefiguration” and sees
the process of formation (Gestalltung) dynamically. Formation rather than the form, becomes
the primary aim of art. Formation is based on a world of diversity where elements are moving
self evidently, in a state of primal motion and rule application causes momentarily development
and fixation. For Klee “to be abstract means to distil pure formal relations”. The creative pow-
ers, which are named “perception” and “formation”, can also be expressed as “perception rules”
and “formation rules”. Accordingly, the “Theory of Formation” proceeds to the study of percep-
tion and formation rules and the study of basic elements. Klee’s approach to graphic art
struggles for systematic precision, but there is an equal effort to avoid the limitations of unifor-
mity. Logic and calculation, which determine the mechanism of generative and productive
techniques, must be developed into more subtle and penetrating procedures harnessing action
and knowledge, manual and mental activity. The formalism that Klee proposes is based on
rationality without numerical formulae, but rooted in experience.
Kandinsky (1926, p. 20) suggested that his primary concern is “the role of the artist in con-
scious creation”. The artist should not wish to initiate “appearances”. Abstraction is opposed to
empathy and scientific approach is proposed as a method of artistic inquiry. The underlying
theme is the discussion on basic elements and the need for a theory for creation and interpreta-
tion of art. “The progress achieved by systematic research will give birth to a dictionary of
12
elements that, developed further, will lead to a “grammar” and finally to a theory of composition.
The investigation proceeds to the examination of points, lines and planes, while emphasis is
given on methods of progressing from analysis to synthesis. Kandinsky employs the parallel
concept of “language”, the “graphic language” has its own evolving laws and grammar. There is
also a profound interest in mathematical expression, which aims to establish an analytical basis
for art theory. The interest in mathematical expression tends in theoretical and practical direc-
tions. In the theoretical approach logic plays the most important role, while practice is purely
purposive. Here, logic is subordinated to purpose, so that the work attains the highest quality.
Kandinsky: Geometric composition
Kandinsky worked almost exclusively with designers and architects, attempting to construct and
justify methods of teaching that would relate the ‘immaterial’ and the ‘material’ in the creation of
a new artistic realm. The study of form includes study of basic elements and study of rules that
Kandinsky names “developing attitudes”. The notion of “basic design” and “basic course”
(Formlehre) are explicitly introduced as educational methods for designers. Kandinsky states
that “the form of the work and its elements are the material for analysis, and not the psychology
of the creation, nor the psychology of aesthetic perception, nor the historical, cultural, sociologi-
cal or other problems of art”.
When Paul Valery said: ‘I did not want to talk, but to “construct”’, he was expressing the prefer-
ence of a modern poet. Before him, poets like Malarme saw the existence of a constructive rule
as precedent to the poetic sentence. Cubists, poets, painters, and artists confronted this di-
lemma: how to ordain the content to the provisos of a “tectonic” composition of form and color.
The concept that makes these ideas similar with the systematic teaching of Klee and Kandinsky
is construction according to rules. Construction, in its architectural sense, underlies the ap-
proach of the sculptural and pictorial composition at the Bauhaus. Klee and Kandinsky isolate
basic formal elements that underlie all visual expression. Their theoretical writings are con-
cerned with the rules that govern the distribution and interaction of those elements.
13
Kandinsky Klee
It is not an oxymoron that as symbolic analytic processes support a growing fraction of human
activities, the creative and visual ones look more and more obscure and incomprehensible. In
the past decade new powerful descriptive and generative approaches in design have been
developed in parallel with the introduction of computer aids in the design process. Today as the
computer influences deeply the way we think and practice architecture we need to address the
questions of creative processes in a new way. An additional reason is that a rigorous computa-
tional approach may open new directions to design. But neither the perspective of radical
transformation of architectural practice nor the change of architectural education is a reason to
disconnect us from the original ideas that the knowledge and experience of the discipline of
architecture offers.
Shape grammar theory aims, amongst others, to preserve the autonomy, the ambiguity and the
creative potentiality of the design process, and to incorporate the qualitative character of formal
composition in computational processes. A parallel goal of the grammatical theory is the rigor-
ous analytic and synthetic approach of “style” through the development of flexible formalistic
apparatuses such as that of “grammars” and “design languages”. In analysis “grammars” inves-
tigate the compositional methods of an existent style. The analytic approach does not examine
the work as something fixed. Exercises of this type have as main goal to introduce sets of rules
that can describe expressively a specific style. In synthesis, the artist just like the scientist
sets out to find a solution to a spatial problem. These innovations can be applied again towards
the solution of other spatial problems, much like variations to a theme. In these variations the
artist recognizes a personal style. Similarity does not mean superficial copying but only prefer-
ence to a set of rules.
In shape grammars basic elements such as points, lines, planes, and solids and combinations
of those are organized according to their dimension in algebras and are manipulated according
to rules. As has been defined in Stiny (1991), an algebra U ij , contains shapes. Each shape is a
14
finite or possibly empty set of basic elements that are maximal with respect to one another.
Basic elements are defined in dimension i=0,1,2 or 3 and manipulated in dimension j > i. In
these algebras shapes can interact in computations that may occur in any dimension. The
shape algebras provide the mathematical machinery for shape computations. Shapes can be
compared, manipulated, and combined in all possible ways. Operations such as sum, differ-
ence, and similarity transformations, and relations such as the part of, which allows on shapes
parts to be embedded in one another, facilitate calculations with shapes in all dimensions. A
shape rule A B applies on a shape C in two steps. In the first step the rule “recognizes” any
part of C that much to any transformation of A. In the second step subtracts this part from C
and adds the same transformation of B in its place.
U00 U01 U02 U03
U11 U12 U13
U22 U23
U33
The U 02 , U 12 , U 22 algebras combined with symbols, that we call labels, and attributes of texture
and color, that we call weights, can express computationally all the activity that takes place on
the drafting table. The elements here are points, lines and planes that are manipulated on the
plane. The operations give us the computational tools to erase a shape, or a line, to add one
shape to another but also to implement computationally notions such as rhythm, repetition, etc.
The transformations extend our perception on the spatial properties of shapes and the different
ways that they can be arranged on the plane. Rules provide the expressive tool to relate spa-
tially one shape with the other with the immediacy that we think, from the very early stage when
the first lines are executed to the very end of the composition.
This study has three goals. First, to look at a historic precedent computational design theory
and expose the similarities and the differences with shape grammars, from technical and per-
ceptual points of view. Second, to show that our design intuitions are inhabited to a large
degree by the ambiguous visual properties of basic elements, and to exhibit how these can be
effectively handled computationally by the algebras of design. Third to look with analytic inter-
est, and from an educational viewpoint at the specific expressive basic visual vocabulary of the
Bauhaus and its corresponding design language as an attempt to deal with design, and design
education, computationally.
15
I. POINT
The first known definition of the point is that given by the Pythagoreans. Proclus (Proclus, p.
95, 21) defines the point as a “monad having position”, the only indivisible element in the sub-
ject matter of geometry. Aristotle (Metaphysics 1016 b 24) uses an equivalent definition: some-
thing, which is indivisible in respect of magnitude but has no position, is a monad, while that
which is similarly indivisible and has position is a point. For Plato the existence of points as a
discrete genus, is a geometrical fiction. He defines the point as the beginning or end of a line,
and he frequently refers to ‘indivisible lines’. Euclid (Book I. 1.) defines the point as that which
is indivisible into parts, omitting the necessity that the point must have a position. Aristotle
illustrates further his conception: a point is not a body, and has no weight. He also attempts to
explain the transition from the indivisible, or infinitely small, to the finite or divisible magnitude.
It is clear that no accumulation of points however far it may be carried can produce anything
divisible, whereas a line is a divisible magnitude. Then, it holds that points cannot make up
anything continuous like a line, a point cannot be continuous with another point, and a line is
not made up of points. A point, Aristotle explains, is like now in time. Now is indivisible and is
not a part of time, it is only the beginning or end, or a division, of time. Similarly, a point may be
an extremity, beginning or division of line but is not part of it or of magnitude. It is only by
motion that a point can generate a line (De anima I. 4, 409 a 4) and thus be the origin of magni-
tude. In modern geometry, the effort to define the point a priori is not made. The mathematician
Max Simon observes (Euclid, p. 25) that “the notion ‘point’ belongs to the limit notions, the
necessary conclusions of continued, and in themselves unlimited, series of presentations. The
point is the limit of localization. If this is more and more energetically continued, it leads to the
limit-notion ‘point’, or better ‘position’. Content of space vanishes, relative position remains”.
Then, ‘point’ is the extremest limit of that which we can still think of, but not observe as a
spatial presentation, where not only does extension cease but even relative place, and in the
Euclidean sense any ‘part-notion’ becomes nothing.
In their approach to basic elements, Kandinsky and Klee adopt many of the Pythagorean con-
ceptions, but they broaden mathematical formalization to treat points, lines, and planes, as
needed in their graphic computations. The point preserves its fundamental geometrical proper-
ties of indivisibility and position, but its incorporeal conception is opposite to the essentiality of
pictorial expression. For a painter like Kandinsky (1924, p. 25) the geometric point considered
in terms of substance equals zero. However, the point in the graphic arts takes material form
and reason of existence: it is the result of the initial collision of the tool with the material plane.
16
Kandinsky departs from the geometric conception of the bodiless and weightless point, to define
the corporeal “graphic point”. He identifies it as the “briefest” or “smallest” form, the more
sharply defined unit of pictorial expression, or the proto-element of painting. The invisible and
indivisible geometric point still mirrors its symbolic significance to the graphic point but in its
actual graphic use the point becomes a sharply defined, indivisible shape. Its boundaries
enclose the innermost concise form of graphic expression.
The beacon guiding Klee through his adventures in seeing, is the line. He adopts the exegesis
of Proclus that the point in itself is “negative”, or “dead”, because it denies all dimensions, and
he makes use of the Aristotelian example to associate the “concept-point” with that of an instant
(1956, p. 494). However, always authentic to his dualistic convictions, Klee never accepts a
concept in itself: a concept is not thinkable without its opposite. Concepts appear in pairs.
Then, when he determines the alternate forces of the design procedure as “rest and unrest”, the
point plays the role of an agent at rest, while mobility is the condition of change (1956, p.19).
The weighted point is not invisible, but an infinitely small planar element. The self-moving point
introduces the first dimension by generating a line, just as a line comes into being after the
application of a pencil on the graphic surface.The abstract geometrical conception of the point
demands that is a 0-dimensional element, a definite indication of position. The graphic point is a
visible but non-divisible small shape. The point is conceptually predefined as a symbol although
graphically acquires form, texture, and size. This follows the necessity that each point corre-
sponds to exactly one position. As a graphic form the point appears equally disjoined from all
sides, its fusion with the surrounding forms is minimal, and seems non-existent in the case of
perfected roundness.
The expressive minimalism of the point and its strong concentric qualities can still find use in
the arts and in architecture. Shape grammars accommodate all the different abstract properties
and possible practical utilities of points. A shape algebra U 0 contains 0-dimensional geometric
points and shapes that are arrangements of points. We can manipulate these points in 0,1,2
and 3 dimensions. In U 00 a single point is the only shape of the algebra, while in U 01 , U 02, and
U 03 shapes are finite arrangements of points manipulated in the line, the plane, or in three-
dimensional space. Correspondingly, an algebra V 0 (V 00 , V 01, V 02 , V 03) contains 0-dimensional
labeled points that can take any form and responds to the graphic demand for symbols of
various forms, while a W 0 (W 00, W 01 , W 02 , W 03 ) algebra supports all the above additionally with
weights, colors, textures and other attributes. Moreover, the algebras can always be combined
with others to accommodate more complex graphic demands. But how do points behave?
17
In the general case shapes are finite arrangements of maximal elements and can be identified
uniquely by sets of basic elements without definite parts. In shape computations, the parts of a
shape and the possibilities for rearrangement are indefinite, depending on rules and transforma-
tions. The parts of any shape made out of lines, or planes may be approached in infinite many
ways each time a rule is applied. In the case of indivisible points (U 0 ) the part relation that
allows lines and planes to be embedded in others, to fuse and to separate naturally, holds only
as identity. Thus, the finite number of the power set of the subsets of any set of points exhausts
all the parts of the shape that is made out of points, and limits the possibilities in finite number
of combinations. This idea is exhibited in Stiny 1991. The following example with three shapes:
the first contains 7 points (U 02 ), the second 8 maximal lines (U 12 ), and the third three planar
shapes a circle, a triangle, and a square. The shape in U 02 can produce a finite number of
combinations between points, that may be seen as arrangements, while the shape in U 12 and
U 22 contains infinitly many parts, segments of lines or planes. The visual power of the example
depends also on the number of points included in the initial shape in U 02 .
. . .
. .. .. . .. . .. .. . .
. .
. .
C. Dictionary
Kandinsky (1924, p. 83) and Klee (1924, p.21) organized an elementary visual dictionary. Their
ultimate goal was the creation of a basic “grammar” that would lead to a theory of composition.
In this effort they established classes of abstract elements that constitute important formal
factors. This basic dictionary extends to points, lines and some of their elementary combina-
tions on the plane, and develops the basis of a “visual language”, which cannot be attained with
words. Kandinsky sees this basic dictionary as one of a living language, being immutable as it
undergoes perpetual changes: “ Words become submerged, or die. Words are created, come
new into the world; foreign words are brought home from across the borders”. Kandinsky and
Klee give two kinds of definitions of basic elements and their relations: relative or absolute.
18
In the first case definitions are relative if they refer to a specific graphic context or to a coordi-
nate system, while in the second they are independent of context. In shape grammar theory,
elements are defined only relatively, in relation to a coordinate system. However, both Bauhaus
artists suggest rules that do not restrict visual ambiguity, but incorporate perceptual transitions
in computations. This can be expressed by transfers of graphic elements from one algebra to
another i.e., points, or lines can be seen and manipulated as planes, etc.
1. Appearance
An essential characteristic of form creation is fusion and mutation of parts that come from
diverse elements. The point cannot participate dynamically in computations by changing parts,
but only through sum, however, its appearance still can acquire an unlimited variety of forms.
For Kandinsky (1924, pp. 30-50) the graphic point depends on:
1. form
2. texture
3. size, and relation to the other forms of the plane
1.1 Form
All graphic elements have some visible form. The abstract point is ideally small and round. But
even the minimal point can get other forms, depending on its graphic use. The need for points
with various forms comes partialy from its symbolic nature, and has no further implications in
the way the point behaves in computations. The point can assume an unlimited number of
forms: jagged, geometrical, or develop into entirely free shapes. Different tools in their contact
with the graphic plane can produce an unlimited diversity of points. No boundaries can be fixed
and the realm of points is unlimited.
Kandinsky
The above concept can be supported computationally in a shape algebra W 0 that contains
weighted points. Points of various forms that retain the properties of indivisibility and definite
position, and can also be manipulated in all dimensions as symbols in V 0 algebra. The change
from a typical round point to one with different form could be expressed in the following rule.
U 02 W 02
19
1.2 Texture
The texture of an element affects the manner it is visually combined with other elements and
with its graphic context. It can be expressed by a variety of different weights, and tones.
Weights are used extensively in architecture, and the visual arts, where different thicknesses of
graphic elements become a fine medium of expression. Kandinsky (1924, p. 50) observes that
texture possibilities should be given even in the limited field of the point. Weights and textures
can be handled computationally in a W algebra that can be extended to points lines and planes.
In the case of points a W 0 algebra contains weighted points that are manipulated in a line, a
plane or in 3-dimansional space.
1.3 Size and relation to other forms

The point grows in size and occupies an area of the plane. The crossing of a size-boundary can
make the point appear as a plane. This perceptual transition is emphasized by Kandinsky as an
important feature of abstract basic elements. In the case of the point it “reveals the expressive
multiplicity of the smallest form attained by slight changes in its size”. Moreover, it is unlikely
that any numerical expression could possibly inhibit our sensory perception and determine this
size-boundary. Kandinsky and Klee steadily avoid the constraints of computational uniformity. A
form or element can change dimension at any stage of the synthetic process. The relation to the
other forms seems here the only parameter that we can refer. As new elements inform the
process it can become necessary that a point should be treated as plane or as linear shape.
This remains entirely a matter of personal decision, since no fixed boundary between points and
other elements can be found. A weighted dot that has been placed on the empty plane as a
point,
can be seen as a plane, when a very thin line appears next to it.
Kandinsky
20
Kandinsky (1924, p. 29) introduces the notion of a rule that doesn’t seem constructive. In fact it
changes the field: the point becomes a planar shape and can be manipulated like any planar
element. This visual-computational transition can be expressed in the following shape rule.
W 02 W 22
The rule substitutes the weighted dot (W 02 ) on the left side, with a weighted plane (W 22 ) on the
right side, depending on how we choose to handle the dot. A perceptual shift can be expressed
computationally as a transfer from a particular algebra to another. The combination of all three
parameters of form, texture, and size, and the perceptual transition of the point to a small plane
can be expressed in a W 22 algebra .
C. Composition
The examination of elementary arrangements of basic elements with the purpose to reveal their
major expressive attributes in relation to some basic context, becomes a standard experimental
procedure for Kandinsky. In the case of the point, position and quantity are explored in the
predefined context of a square plane.
1. Centric structure
The positioning of a point at the center of a shape illustrates one of the most common uses of
this basic element in all different areas of design. The point acquires the character of a center,
and also, reveals the nature of the center itself. In architectural design the manipulation of the
center of any given context or spatial arrangement plays a consequential role in the process of
composition.
Kandinsky
The concept of the center can be approached more constructively, to produce composite forms.
Kandinsky and Klee introduce examples in which the point evolves in concentric, radial, or
spiral structures. In the simplest case, computational rules that arrange shapes or planes in
relation to an initial center, at a distance d or a radius r, can be used to generate designs in a
21
U 12 or U 22 algebra. The same concept can produce concentric structures with colors or weights
in a W 22 algebra. The productive growth of the point in concentric waves of geometrical or free
forms generates a two-dimensional structure emanating from nuclear strata.
radius r : → ⇒ ⇒ ⇒
distance d : → ⇒ ⇒ ⇒
Klee
The more characteristic centric structure is the circle. The spiral progress of aline from a center
point produces spiral structures which hold different properties from the circle. Rhythmic repeti-
tion of any type of line, about a common center creates the radial structure. Any combination of
the above centric structures is possible by superimposition of their centers.
Kandinsky
As an extension of the axial concept we can generate through rotation of any linear shape or
planar element, about a common center point, the two point groups of symmetry. Rotations
about a single point through 2 p /n generates the cyclic group C n while, an additional reflection
through an axis that passes from the given center generates the dihedral group D n . Arrange-
ments of this type are widely used in the arrangement of architectural plans (example, F. L.
Wright Suntop Homes).
March-Steadman
22
The above arrangements evolve though computations that are based on the concept of the
centric structure. That is, form-production about a center, where the concentric qualities of the
point are accentuated and projected in two, and three dimensions. The rules apply in reference
to an initial point to generate forms that mirror its abstract concentric qualities. Apparently, each
of the centric structures (concentric, radial, spiral) that is generated from an initial center-point,
and a distinct set of rules, possess distinct properties. According to the functionalistic artistic
approach of Kandinsky and Klee this is not a thing that should be overlooked.
2. Eccentric Structure
An eccentric structure consists of a point moved from the center of the plane. The arrangement
can be seen as the counter concept of the centric arrangement. It transforms the visual impact
of the graphic point from the absolute position of the center to a relative position that depends
on the relationship with the boundary. More generally, the eccentric point starts to interact with
its context. Klee proposes a constructive use of the eccentric point. The displacement of the
point can be used in the construction of pictorial depth.The eccentric position of the point
effects the nuclear strata of a centric structure, and the density of the boundaries of the
cocentric linear or planar shapes creates the illusionistic three-dimensional space. This compu-
tation can be executed with lines on the plane in U 12 algebra, with planes on the plane in a U 22
algebra, or in a compound U 12 , U 22 algebra that includes planes and their linear boundaries.
Klee uses the computation often in his paintings by superimposing several eccentric structures
that create the visual effect of multiple perspective views on the same canvas.
Klee
23
3. Quantitative Increase
The repetition of an element produces more complex visual results. Accumulation of points
especially in the case that they are not identical can develop a multifaceted graphic effect.
Multiplicity and diversity of form, texture, and size, which establishes the graphic point in count-
less entities with distinct values, is the fundamental expressive medium of the typical graphic
techniques like etching, woodcut, and lithography. Properties of this type could be accommo-
dated computationally in a W 22 algebra, where points can be treated as elementary planes
diverse in shape and tone.
Kandinsky
However, the use of sequence of points with the intention to construct a line is not justifiable for
Kandisnky (1924, p. 53) : ”This line is condemned to a poverty-stricken half-life”. True to his
functionalistic approach to basic elements he emphasizes the value of differences as originally
purposeful and well-founded. Elements in computations should be treated for what they are.
E. Points in architecture
Points can also be the medium of subordination since repetition in an ordered fashion is a
source of rhythm. In architecture, the organization and interaction of multiple overlapping
networks of elements and activities makes the use of overlapping grids a common medium of
subordination. Here the point signifies a noteworthy position of an “event” that takes place in
the abstraction of the plane or the three-dimensional space. Changes in the graphic size, the
form, or the weight of the points may signify corresponding changes in the interpretation.
However, the concentric qualities of points remain unaltered since graphic points grow out of
their centers.
T. Ando
24
The implications of repetition depend upon the type of points, and the canonic or free mode of
repetition. A repetitive series of points of the same weight, arranged linearly, in an architectural
ground plan can be seen as an indication of rhythm of form, structural order, and direction of
movement. In space the point is the result of cross-section of several planes, or the termination
of an angle. The peak of any architectural form can be seen as a point in an elevation.
T. Ando
Architectural design occurs in algebras that contain points, lines, planes, and solids. Combina-
tions of algebras like U 12 , V 02, W 12 and W 22 , which manipulate lines, labeled points, weighted
lines, and weighted planes, are employed to support computations that occur on the drafting
table. The action of points can be restrictive or creative. The restrictive action of points in
calculations with lines and planes is twofold. First, labeled points (V 02 ) are used as symbols to
to assign graphic and symbolic attributes like weights and colors, functions, materials, etc. to
lines and planes. Second, weighted round points (W 02), or points of various forms, arranged on
the plane and manipulated as weighted elementary planes (W 22) are used to visualize bound-
aries of elements, or grids. In this way, points support the articulation of “meaning” or “rhythm”,
by restraining the free interaction of lines, and shapes thus establishing an underlying structure.
25
However, we should not disambiguate the action of points, which also has deep creative impli-
cations. The accumulation of points of diverse form that can be suddenly seen and manipulated
as planes, the slight convergence or divergence of ambiguous stains, spots, and dots, their
density or looseness, their weights, the continuous or discontinuous character of their rhythm
etc. Analogous repetitive marks in drawings and models should not be overlooked as symbolic
indications, or side effects of a capricious technique since often they can be traces of the
deepest intentions for the composition.
The geometric constraints for points in this case should be overpassed. The graphical character
of the point allows all kinds of points to occur naturally. Points that begin as simple indications
of position demand to express form. They become planes or linear shapes that are manipulated
dynamically and change size, texture, and dimension in ananticipated ways. This ambiguous
interaction of scale, dimension, and parts, between graphic elements becomes a usefull and
expressive tool in composition where the solutions of form emerge graphically and the dimen-
sion of graphic elements cannot be fixed.
26
Klee
The restrictive and creative action of points has its foundations in graphic representation. Klee
notices that points restrict lines by determining, and visualizing their boundaries. In another
illustration, the point discharges its tension towards another point, to generate a line. Points of
equal or different intensity that lie on the plane, or others that belong to restricted lines visual-
ize free or restricted rhythms.
Linear points
Planar points
Klee
Klee (1956, p. 152) rejects any notion of predetermined “structure” or “prefiguration” and sees the process
of formation (Gestalltung) rather than the form, as the primary aim of art. Dynamic figuration, which can be
expressed as dynamic rule application, is based on a world of diversity, where things move freely for the
sake of going, without aim or will, and without obedience. In this phase, elements are moving self evi-
dently, in a state of primal motion. The opposite (counter concept) is what happens afterwards as the result
of figuration: change, development, fixation, measurement and determination. The end, what Klee names
“appearance”, is only a part of what is essential. The true essential form is a synthesis of figuration and
appearance. But form is momentarily and all rules are made to be broken. In art it is essential to create
movement according to rules and to create deviations while bearing the rules in mind.
27
II. LINE
Proclus (p. 96, 21-97, 3) defines the line as “magnitude in one dimension” or “magnitude ex-
tended in one way”. While the definition of the point is negative, the line introduces the first
“dimension”, and to this extent its definition becomes positive, while it also sustains a negation
which denies the other dimensions. Aristotle (Metaph. 1016 b 25-27) defines the line as a
magnitude “divisible in one way only” in contrast to a magnitude “divisible in two ways” which is
a surface, and a magnitude “divisible in three ways” which is a solid. Aristotle also connects the
line with the notion of “length”, while “breadth” is connected with the surface, and “depth” with
the solid. Euclid states that a line is breadthless length. Proclus gives an alternative definition
of the line as “ flux of a point”, or the path of a point when moved. The same idea appears in
Aristotle (De anima I. 4, 409): “ They say that a line by its motion produces a surface, and a
point by its motion a line”. Proclus (p. 97, 8-13) comments that “this definition is a perfect one
as showing the essence of a line. He who called it the flux of a point seems to define it from its
genetic cause, and it is not every line that he sets before us but only the immaterial line. For it
is this that is produced by the point which, though itself indivisible, is the cause of the existence
of things divisible”. Aristotle (Metaph. 986 a 95) notes that the Pythagoreans distinguished
straight and curved lines. This distinction appears in Plato (Republic X. 602 c) and in Aristotle
(Anal. post. I. 4, 73 b 19). From curved lines Plato and Aristotle separate off the “circular” as a
distinct species. Aristotle recognizes broken lines forming an angle as one line. Thus, “ a line if
it be bend, but yet continuous, is called one” while for Heron “a broken line is a line which,
when produced, does not meet itself”. Aristotle also (De caelo I. 2, 268 b 17) states that “all
motion in space, which we call translation is (in) straight line, a circle, or a combination of the
two; for the first two are the only simple (motions)”.
Following the Pythagorian view Kandinsky (1924, p. 57) sees the line as a derivative element.
Line is as the track of the moving point that signifies the visual transition from the static to the
dynamic. Movement that alters the stillness of the point requires a natural cause, which
Kandinsky identifies, in the alternate or simultaneous action of one or two external forces. Here
an external reason causes the transformation of an existing basic element into a new one with
entirely different properties. The characteristic visual property of the line becomes the ability to
evoke motion. Our eyes follow the line in its full length. The line is also divisible in infinitely
many ways in opposition to the indivisible point. The line remains always a line, even in its
smallest subdivision, since the notion “line” is qualitative. Thus, the line is the antithesis to the
static point, as “motion” is the antithesis to “position”.
Klee
28
For Klee (1956, p. 105) the line is the most “authentic” basic element and the most interesting
to explore because it is the most active. The point in dynamic terms is only an agent, while line
is the condition of change, and the generator of endless phenomena in the domain of vision. In
opposition to Kandinsky, Klee negates the idea of an external force that moves the point. The
point sets itself in motion and a line comes into being. However, a restricted straight line does
not produce a dynamic form. While the line in its most original, primitive condition is structure-
less and develops freely the line between two points is static. The question as to whether a line
is dividual, which means that can be divided freely, or individual, which means that it cannot be
divided without loosing its qualitative significance, is decided by what Klee names the criterion
of indefinite extension, or definite measure. The criterion that refers to the context of the com-
putation and as a concept has deep implications in the way we compute with unstructured lines
versus restricted lines. When there is indefinite extension, arbitrary divisions can be made
without changing the structural style (dividual). But where an individual has definite measure,
nothing can be added or subtracted without changing it into another individual (Klee 1956, p.
245)
Klee
The concept corresponds to two distinct computational approaches that can be applied in
different cases. The idea of definite measure corresponds to the deterministic or combinatorial
mode where the use of predefined units, restricted elements, or modules that refer to a scale of
distinct quantitative values, prohibits arbitrary divisions that may cause confusion in the evalua-
tion of the components and in the way rules apply to combine them whithin the system. In that
sense Klee defines these elements as “individuals” and the quantitative evaluation,or the
number, determines the qualitative properties of the elements. The notion of indefinite exten-
sion corresponds to non-deterministic or unrestricted elements of arbitrary measures, where
arbitrariness is guided by criteria that emerge in place, and parts are picked out by rules that
act on shapes, parts of shapes, and divisions of lines. Elements and all their possible divisions
or extensions are free to interact as desired, and both qualitative, and quntitative “values”
emerge on the way. Klee defines these elements as “dividuals” because they can be divided or
extended in infinetely many ways . Here the qulitative evaluation determines the quantitative
properties of the elements. Each of two modes produce corresponding structures that Klee
names individual and dividual structures. However, arrangements of elements can be ap-
proached as dividual or individual, depending on how we choose to treat them. Arrangements
that include determined number of parts can be seen as individuals, while if the number of
parts is undetermined can be seen as dividuals.
29
Examining the concept of dividual and individual in relation to shape algebras, in U 0 algebras
points and shapes with points are individuals , whereas algebras that include lines and planes
can support both dividual and individual structures. The combinatorial approach can be suitable
when standardization of components and forms is preferable or where preanalysis is possible,
whereas the non-deterministic approach avoids the need for foreknowledge, leaves more
freedom to innovation, and appears more open-ended. As an extension of the above ideas the
two approaches can be combined to produce composite forms in which individual structures
produced in the deterministic way combine with dividual structures produced in the non-deter-
ministically. An example of composite structure, is the combination of the structural grid of a
building (parti), which is individual, with the actual plan of the building which is usually dividual.
For Kandinsky and Klee the formal elements of graphic art are points, and linear, planar, and
spatial energies. In the general case basic elements are used in formal composition as distinct
qualities. Linear, and planar forms can be constructed from shapes that are arrangements of
lines and planes. Planes, and lines, weighted or not, are combined through application of
compositional rules to produce forms. Also operations like addition, subtraction, and multiplica-
tion and the set of transformations are explicitly introduced by Klee (1956, p. 159) as an essen-
tial part of a computational machinery of dynamic figuration.
+ =
Addition
Klee
The process of composition becomes equally important with form. In the general case when a
linear form is combined with another linear form according to a rule, the linear parts are consid-
ered to have “active” character. In a linear shape that lies on a finite plane the linear part takes
on a decidedly active character and the plane a passive character. Klee suggests that the
active part, the line, can accomplish two things: it may divide a planar form into two parts or it
may give rise to a displacement of the parts (Klee 1956, p. 7).
Klee
30
In shape grammars, a line is a distinct quality and is considered always a line. The line is
bounded by definite and distinct endpoints, and has finite nonzero length. Linear forms made up
of linear shapes can be manipulated on the plane, or in other dimensions according to rules. In
the case of two-dimensional space, a shape is a finite arrangement of lines on the plane. A
shape can contain occurrences of straight or curved lines that are connected or disconnected.
Rules that refer to relations between shapes apply to generate new arrangements. The linear
parts can be embedded in one another can be separated, can be added, subtracted, or trans-
formed. (rotated, reflected, translated or scaled).
Klee
C. Dictionary
Klee (1924, p. 23) defines the three formal means that apply to all basic elements as length,
weight, and color. Length is quantitative and can be equated with measure, weight is medial can
be equated with texture, and color is qualitative. Quantitative properties like length, angles, length
of radius, and focal distance, are subject to measurement. Weight is a “relation” and includes tone
value, which can be expressed as degree of shading between black and white, and as texture.
Finally the color can be neither weighed nor measured. According to Klee neither with scales nor
with ruler can any difference be detected between two surfaces, one pure yellow and the other
pure red, of similar area and similar brilliance. Basic lines are considered the straight and the
curved. Definitions of elements like the vertical, the horizontal, and the diagonal lines, are relative
and refer to a specific graphic context determined by a coordinate system, while definitions like
that of the angled or the curved lines, are independent of context. In shape grammars elements
are defined only in relation to a coordinate system, while the graphic context of the dimension in
which we manipulate the element is considered infinite.
1. Straight Line
Kandinsky (1924, p. 57) illustrates the generative cause of a line as action of one external force
to a point. In their abstract definition of straight lines Kandinsky and Klee see the line having
infinite length, therefore consideration of endpoints seems unnecessary. However, this need
soon appears reasonable to Klee who observes that any line can be the result of tension
between two distinct points. In shape grammars a line is always bounded by definite and dis-
tinct endpoints, and has finite nonzero length. Moreover, a line can be transformed to any other
line when is embedded on it and its endpoints are moved. This happens naturally if we draw
one line on another and for Kandinsky and Klee there is no need to define this relation. In
shape grammars lines can have common linear parts or can be discrete, whereas for Kandinsky
and Klee linear forms can also touch in points. In both theories the position of lines on the plane
and their relation with other lines is determined by the position of their endpoints. Kandinsky
(1924, p. 57) distinguishes lines in typical and atypical.
31
1.1 Typical straight lines

Typical straight lines are the horizontal, the vertical, and the diagonal because all three can be
perceived, and executed without measurements. Kandinsky sees the horizontal as a flat sup-
portive base while in complete contrast the vertical, has height. For Klee this pair of lines refers
directly to the symbolic scale of the coordinate system, where the horizontal is dependent on
the vertical. The diagonal diverges equally from both.
Kandinsky
1.2 Atypical straight lines

Kandinsky names atypical all straight lines that are not horizontal, vertical, or diagonal (1924, p.
63), because their relation to the coordinate system cannot be determined visually. However, in
a given finite context, which is determined by a finite plane, Kandinsky implies that we can refer
to the center and characterize these lines as centric or accentric.
Kandinsky
In shape algebras elements are considered finite but the space in which we manipulate these
elements is infinite. Still the above organization of lines in relation to the center of a finite
graphic context, or in relation to multiple “centers”, could be constructively useful in painting
and in urban planing or architecture, where arrangements of blocks and roads need to be
organized in relation to particular centers. The above organization of lines could be attainable
by defining the limits of the dimension in which we manipulate the element, in this case the
plane. Then, the intuitively suitable notion of finite graphic context needs possibly to be exam-
ined as an optional feature of a U ij algebra in order to accommodate the standard practical
demand for rule application in the finite surface of a canvas or a specific site in the case of
architecture. One example of constructive application of the above ideas in painting is illus-
trated by Klee (1956, p. 190). In these arrangements of straight, vertical, horizontal, diagonal,
centric and acentric lines, a predetermined scale of convergence to the vertical and the hori-
zontal and multiple centers are applied. The visual effect is the creation of pictorial depth.
32
Klee
2. Angular and curved lines

Kandinsky and Klee categorize angular lines, with curved lines as the result of a more complex
movement of a point by two external forces. Moreover the angular line is considered as one
line. The action of two forces in the realm of the line can take place in two ways:
1. Point k Alternate action of two force k Angular line
2. Point k Simultaneous action of two forces k Curved line
Kandinsky
2.1 Angular line

The notion of an external force that acts on one element, the moving point, and changes its
straight course, is a mechanical way that Kandinsky employs to describe how a given context
interacts with the straight line and changes its form. Angular lines are also described in Klee’s
illustration of a sequence of straight restricted lines. The angular lines in computational terms
can be seen as shapes that contain a finite number of straight maximal linear elements and are
manipulated in a U 11 , U 12, and U 13 algebras, where the part relation holds if the elements have
common linear parts. The U 11 algebra contains one straight line that is the special case of an
angle of 180 . For the rest of the algebras maximal linear parts combine on the plane and the
three-dimensional space in a way that every two share one common endpoint and in the general
case do not coincide. An extension of the concept of angular lines is also the notion of angular
planes, or solids that are manipulated in space in U 23 and U 33 algebras. While straight lines can
33
be described as length, angular lines depend on two characteristics: their angles and the
lengths of their straight linear segments. Many-angled lines can also derive by addition of pairs
of simple angular lines.
→ ϕ
2.1.1 Angles and lengths

The angle is the simplest case of angular line.For Klee and Kandinsky the notion “angle” is
qualitative and belongs entirely to the domain of vision, without further numerical extensions.
In shape computational terms the simplest angular line contains only two maximal linear seg-
ments that share one endpoint, and in the general case do not coincide. Angles are character-
ized by Kandinsky (1924, p. 69) according to visual criteria as typical or atypical.Typical are the
angles of 45 , 90 , and 135 because their characteristically exact angle-sizes can be perceived
and executed intuitively without measurements, while all other angles are characterized simply
as atypical.
Angle Quality
α. Acute – 45 ⋅ sharp
β. Right – 90 ⋅ controlled (Typical)
γ. Obtuse – 135 ⋅ passive
δ. Free (Atypical)
In shape computation a parameter ϕ can be introduced in a shape rule to express the value of
the angle and to construct different angles. The shape rule takes the initial line on the left side
and returns an angle with value ϕ on the right side. The further difference between the simple
angular lines consists in the lengths of their individual linear sections, which greatly modifies
the basic character of these forms. In the sense that Kandinsky describes thegenerative cause
of angles, as action of one external force to a staight line, we may parameterize further limita-
tions in the values of angles and lengths according to external conditions and translate the
interaction of an external cause with straight line into specific angular forms.
Klee
34
Typical angular forms have distinct ways to tile the plane. Accordingly, four right angles tile the
plane and can form a cross, or a square, eight acute angles tile the plane, while, the obtuse is
not capable of tiling the plane exactly but two obtuse angles added together leave an exact
portion of 90 unconquered.
+ =
+ + +
2.1.2 Compositional parameters of angular lines

Kandinsky (1924, p.70) introduces a concept that is related to architecture. Although an angu-
lar line is described as lengths and angles, its potential in formal composition depends on three
parameters. These parameters can be seen as one composite parameter, they can be used
singly, or in pairs.
α. The lengths of their straight parts.
β. The inclination of their angles (more or less acute).
Kandinsky
γ. The conquest of the plane.
Besides the lengths of the straight segments and the angle, the third parameter relates the
angular line to its context. While the two first can be expressed arithmetically, the third belongs
entirely to the domain of perception and it is relative to the context of the composition, so that
numerical values do not apply. Here Kandinsky introduces a significant concept for the qualita-
tive difference between straight linear and angled elements. While straight linear elements,
simply “divide” in a more or less neutral way, angles have the tension to “conquest” in distinct
ways, something that we experience stronger in architectural space. An acute angle made out
of planes conquests the space in a distinctively different way from an obtuse, or a right angle
etc. The same concept is implied by Alberti (1955, p. 1-2) who defines “design a firm and
graceful preordering of the lines and angles, conceived in the mind, and contrived by an inge-
nious artist” thus establishing angles as a distinct quality of architectural form, although angles
are made of linear parts.
35
2.2 Curved line

Kandinsky (1924, p. 79) by applying the concept of external force generates the curved line
from simultaneous action of two forces on the point where the one force acts steadily and
continually and exceeds the other. This type of line exhibits the tendency to close itself. The
simple curved line and the straight line have different character. If the straight line expresses
the possibility for endless extension, the curved line has “elastic” character. Moreover,
Kandinsky according to the Pythagorean approach notes that the straight line and the curved
line constitute the primary contrasting pair of lines, because while the angled line includes
straight segments, in the curved line we notice complete absence of the quality “straight line”.
In a computational environment forces that are not elements of design, but that act on basic
elements, could be introduced as another means of shape transformation or as away to inter-
pret shapes.

 Kandinsky

Klee (1956, p. 123) sees the curved line as the antithesis of the straight line, a form that has
something restful and harmonious. The non-geometric curve is an expressive suggestion of a
restful walk without purpose, and in musical terms suggests a folk-song rather than a more
elaborate form. However, the curve can also appear in restricted form, and can be generated in
relation to two given points.
Klee
Curves are lines that can be manipulated in a U 11 , U 12, U 13 algebras like other lines. Geometrical
curves are easier in description with numbers, where free non-geometrical curves are more
complicated forms. The embedding relation and the transformations need to be extended to
calculate with free curves. However, complex forms of curves are usually combinations of
simpler curved segments. Kandinsky (1924, p. 85) offers computational rules for the construc-
tion of complex wave-like lines that cover all the forms of curves. Accordingly a complex curved
or wave-like line consists of geometric parts of a circle, free parts, and various combinations of
these (equal or unequal radius, horizontal or vertical course, and alternating tensions).
Shape rules can be introduced in U 12 algebra, to illustrate this idea,
Curve geometric wave-like

k
36
D. Composition
Kandinsky (1924, p. 74) observes that in formal composition the components produce a sum,
but there is not always direct correspondence of the character of the sum with the characteris-
tics of the components. The same happens in chemistry, where the sum divided into its compo-
nents in many cases fails to be restored in the combination of the components. Therefore, the
equality sign is not used but rather, an arrow  which points to relationships. Kandinsky
suggests that the quest of his research is “the relationships between the elements”. The sum, or
the form, simply supplies the missing factor necessary to balance. The components would be
derived from the sum and vice versa. The same idea is expressed in shape grammars with the
concept of rule application, or spatial relation, with a “left” and a “right” part. This corresponds
precisely to the notion of a relationship, where components and solid descriptions are not
preserved since may restrict elements to interact freely. Multiple descriptions can emerge from
rule applications on elements and the components may vary. The role of rule application in this
case is also highly emphasized by Kandisnky (1924, p. 92) who states that only the action of a
rule on an element brings life to that element, and allows to its basic properties to be ex-
pressed. Accordingly, he defines composition as an exact law-abiding organization. The compo-
sition with lines is the clearest and simplest example of this creative process which always
takes place according to rules, and therefore allows and requires rule application.
1. Lines, angles and shapes

Simple straight or angular lines produce shapes. We need at least three lines or angles to
compose the simplest shape, and with it the primary plane made out of straight lines, whereas,
one curved line can produce the primary plane of a circle. Thus, the triangle and the circle
become for Kandinsky the contrasting pair of shapes. A further correspondence between lines,
planes, can be established. The qualitative characteristics of angles are used to point to planar
shapes. The basic planar forms are introduced as products of angular lines. The square, the
triangle and the circle become the typical forms of the plane. Right angles, with balanced
qualities construct the square. Sharp acute angles construct the equilateral triangle. The pas-
sive obtuse angles are distantly related to a form without angles. They constitute the third
primary shape of the plane, the circle.
Angles Forms
k => => =>

Right angle 
k => =>
Acute angle 
r k
Obtuse angle 
37
2. Complexes of lines
A further extension of the concept of subordination of elements according to rules is possible
for lines. Kandinsky (1924, p. 93) attempts to expose the constructive power of two principles,
the principle of the parallel and the principle of the contrast. The visual result is more complex
forms and articulation of different rhythms, depending on the quantitative or qualitative nature of
the repetition and of the linear elements. The two compositional rules of parallel and contrast
can also be expressed as rules in a U 12 or W 12 algebra
Repetition of straight line with alteration of weights or intervals:
Repetition, or opposed repetition of angular lines:
k k
Repetition, or opposed repetition of curved line:
k k
Contrasting repetition of curved line:
Kandinsky
38
3. Linear, medial, and planar shapes

In U ij shape algebras boundaries of elements are defined in an algebra U i-1j . This expresses the
fact that points bound lines, lines bound planes, and planes bound solids. Examining the inter-
action of lines on the plane, Klee and Kandinsky refer to two distinct cases. First, finite arrange-
ments of lines that have at least one of their endpoints not common or lying in between the
endpoints of another line are considered linear active elements. Free straight lines, or angles
that share one common end point or arrangements of lines about a common center are included
here.
Klee
Second finite arrangements of lines that compose closed linear shapes are considered medial.
This follows the Euclidean definition of shapes where a figure is considered shape if and only if
is contained by closed boundary or boundaries. Klee (1956, p. 111) observes that lines by
circumscribing a shape bound an area of the plane and cause planar effect. If we see a shape
as an action of the hand the line remains always a line. But when it is completed the linear
impression inevitably changes to a planar impression and the linear character disappears. For
the same reason Klee characterizes closed shapes hybrids. For Klee and Kandinsky shapes
can be seen as combinations of lines or as boundaries of planes depending on how we want to
look at them, and therefore they are considered medial. All free and geometrical shapes are
included in this category,
Klee
In the case of weighted or colored planes when the boundary-lines are not visible as distinct
elements lines are considered to be “inactive” and the planar character of shapes prevails.
Klee
Then, arrangements of straight lines and closed curved lines can be visually associated to
planes. Kandinsky observes that geometry does not make this distinction which is exceedingly
important for graphic representation and mentions the qualitative difference between the circle
and the spiral. Although both are products of the curved line, the spiral can be perceived only
as a line, where the circle may also be seen as a plane.
Klee
39
This ambiguity of shapes may have implications in the way they combine. Klee distinguishes
three modes of interaction: shapes may be discrete, may lie side by side, or overlap. The
relation between their parts may be, no contact, or contact in point, in line, in plane, or in
space. Accordingly, parts of shapes may be apart, touching (one-dimensional contact) or inter-
penetrating (two-dimensional contact).
Klee
A shape in a U 12 algebra is an arrangement of lines that lie on the plane. Shapes may be dis-
crete, or may have common linear parts. Shapes with lines that share points are considered
discrete since they don’t have common linear parts. More generally, in shape grammars shapes
have common parts only if the parts belong in the same algebra with the shapes. They do not
have common parts otherwise. Thus, in the above example in a U 12 algebra only the two rect-
angles are considered to have common parts, whereas Klee would characterize, the first case
discrete, the second touching in a point and the third, interpenetrating. In the fourth arrange-
ment the rectangular shapes both in a U 12 algebra and in Klee’s approach have a common
linear part, in the case we see them as an arrangement of lines. If we see them as an arrange-
ment of planes, in a U 22 algebra, the shapes do not have common planar parts. The above
relations are illustrated in the following table,
Klee 
U 12   
U 22   
U 22 U 12  
Klee’s characterization of shapes as medial gives him enough flexibility to see them and ma-
nipulate them as arrangements of lines or as boundaries of planes. The difference with U 12
algebra appears stronger in the second example, where Klee sees a common linear part be-
tween the two shapes. This intution to pick this part as common between the two shapes, is
irrelevant from our knoweledge of the dimension in which they belong and expresses the
continuous flux between dimensions. A compound U 22 U 12 algebra, which contains planes and
40
lines manipulated on the plane would produce almost the same graphic environment that Klee
characterizes “medial”, since it contains both the planar shapes and their linear boundaries.
However, the notion of a more inclusive medial graphic environment, where shapes made out of
lines can optionally be approached as planes, or they may have common parts without neces-
sarily having common linear parts could be more appropriate. Still the introduction of a specific
shape rule could possibly capture more expressively Klee’s approach, just for the case. The
rule in a U 12 algebra would pick any linear part we need from two shapes without altering the
algebra.
U 12 U 12
To incorporate better the option to perceive shapes in different dimension and to be able to
calculate with them in a more systematic way, complex cases of similar visual transitions would
probably demand more thoughtful action. An alternative would be to retain both descriptions of
the element or the elements that “change” or transfer to another dimension, and to proceed in
parallel computation, in different algebras or combinations of algebras. In this case one can
always refer to the step that the transition occurred or to any of the indermidiate steps and
proceed from there. This resembles the parallel exploration of ideas in architectural composi-
tion, where shifts in the way one sees and manipulates shapes occur naturally. In each of these
stepsof transition the use of tracing paper leaves intact or informs the previous stages of the
composition. In the following example a rule places squares into squares, and provides a basic
form that can be explored, or decomposed in various ways. In the case that one is interested to
explore potential solutions for the plan of a house,one would go through series of parallel
computations. The one that is included here transfers linear parts of the arrangement, to
planar parts that correspond to specific spaces.
41
rule 1:
k
rule 2:
k
rule 3:
k
Derivation A: using only rule 1
(import shape (import shape

from Derivation A) from Derivation B)
Derivation B: using only rule 2 Derivation C: using only rule 3

42
4. Divisions of shapes
Can we define “elementary” divisions in arrangements of straight lines? In the following ex-
ample a horizontal and a vertical line are placed centrally on the square plane. This arrange-
ment is considered by Kandinsky (1924, p. 66) an extension of the concept of the centrally
placed point in the domain of lines.
Kandinsky
The issue here is to define the “elementary parts”. Kandisky manipulates the plane of the finite
graphic context as part of the composition and sees the arrangement in various ways. Visual
transitions from from plane to linear shapes occur spontaneously.The initial arrangement is
defined here in a combination of U 22 and U 12 algebras, which contains a square plane and two
lines, manipulated on the plane. A parallel computation takes place only with lines on the plane
in a U 12 algebra . First we see this shape as the result of interaction between 1 plane and 2 lines.
This gives a description that contains three elements.
+ =
However, Kandinsky notes it is equally permissible to see the arrangement as four equal
squares made out of lines, in a U 12 algebra. The three previous elements can be seen as four
new elements of entirely different nature.
+ + + =
Kandinsky observes that further possible elementary divisions of this basic schema exist: six
horizontal and six vertical lines could be seen as the elementary parts of this arrangement,
again in a U 12 algebra. The sum consists of 6 + 6 =12 lines.
+ =
43
Or, we can alternatively see the arrangement as 4 planar elements + 2 linear elements = 6.
+ =
The results of Kandisky’s attempt to define “elementary” divisions of the arrangement is illus-
trated in the following table, in parallel with a reference to the shape algebras that the parts
belong.
= = =
U 22 1 plane - - 4 planes
U 12 2 lines 4 linear squares 12 lines 2 lines
Kandinsky (1924, p. 66) ends with the conclusion that the notion of “elementary” combinations,
or divisions is relative. Relativity is part of the nature of shapes. Any choice between combina-
tions of elements remains open and can describe the same arrangement equally well. The
characterization “elementary” is clearly inadequate and imprecise, since it is not easy to fix a
limit for the complexity of shapes, or to reduce this complexity into elementary parts. However,
Kandisnky observes that the analysis of shapes through observations like the above offers the
only means of getting to the bottom of pictorial things, and serves the ends of composition.
A further conclusion that we can draw from Kandinsky’s observations is that the ambiguity of
graphic forms arrives from both ends, it is related to transitions of shape and dimension. Basic
elements may compose finite arrangements that we call shapes in a specific algebra in infinitely
many ways and shapes may change their basic components,in the shape algebra at any de-
mand. In that sense infinite compositions and decompositions of the same arrangement are
possible. Moreover, the dimension of these parts may be subject of continuous change. Parts of
shapes may turn into elements of a different algebra. The expressive power of basic elements
is based on their ability to interact in ways that are not predetermined in any sense and depend
exclusively on the eye of the beholder. This ambiguity of form and of dimension makes basic
elements abstract and inclusive, while at the same time they are concrete and visual, ready to
support any possible interpretation. The multiplicity in the way we see and treat forms in archi-
tectural design is highly dependent on these visual transitions where the dimensions and the
parts of graphic elements constantly change. In shape grammar theory the organization of basic
44
elements in distinct algebras and the possibility to transfer shapes from one algebra to another,
to calculate simultaneously in different algebras, and to combine them, serves primarily the
demand for dimensional transitions, while, the shape rules, the identities and the part relation
that allow parts to be composed and decomposed in infinite many ways satisfies the normal
expectation that forms do not have predetermined parts. However, visual transitions of dimen-
sion that occur spontaneoulsy in computations and cause phenomena of perceptual flux, need
to be treated as a significant factor of ambiguity and creativity and need to be examined and
supported computationally in further detail from shape grammar theory.
E. Graphic properties
1. Boundaries and texture

Kandinsky (1924, p. 91) observes that the outer boundaries of the line can be considered as
two independent lines that are not necessarily straight or continuously parallel. Then, the edges
of the line can be seen as having a shape. These outer edges may undertake qualitative and
expressive significance. They can be smooth, jagged, torn, rounded etc. Since line has length
the combinations are more complex than those in the case of points and each of the two sides
of the line can have a special character. However, if we remain in the domain of lines, and we
do not consider a specific line segment as a plane, the form of these boundaries can be treated
as part of the line texture in a W algebra (W 11 , W 12, W 13 ). The importance of weights as a means
of graphic expression is emphasized by Klee. The slightest change in weights of lines can alter
the perceptual value of the elements. For Klee weight is “medial” parameter is in between
length and color, where length is characterized as “measure” and color as “quality”.
Klee
2. Emphasis
In this case that the shape of the line- boundaries develops to a significant visual parameter for
the quality of the line, the weighted line can be approached as having specific atributes of
form. A special case of the above is when a specific visual effect of increase or decrease of
strength is desired. Kandinsky (1924, p. 89) names the effect emphasis. Here the transition
from a simple line to a weighted line occurs as a more radical shift and weight is used as a
computational means to enable the line to attain the shape necessary at the moment.
Kandinsky
U 12 W 12
45
3. Transition to plane
Any short weighted line often bears a relation to the plane. Kandinsky (1924, p. 90) observes
that as in the case of the point the boundaries of this transition cannot be determined analyti-
cally but are indefinite and mobile. The visual outcome depends upon proportions, relations with
the other forms of the plane, and more importantly perception. However, this visual transition
from line to plane becomes a potent source of expression and of unexpected consequences in
computations with basic elements. Here one is dependent upon visual intuition and personal
will, since a predetermined distinction between line and plane is impossible. In shape computa-
tional terms a line on the plane in a U 12 algebra, or a weigthed line in a W 12 algebra needs
often to be approached and manipulated as a small plane. Rule application that transfers the
element in a U 22 or W 2 algebra which contains planes or weighted planes, can serve this de-
mand.
U 12 U 22
In the more general case, any line, or element that belongs to a specific algebra can change
dimension. The constructive interest of this transition appears in the computations that follow,
since designers manipulate in a different way and for different reasons, a line from a plane, or a
plane from a point. As Klee (1956, p. 78) points out the work of art is first of all continuous
genesis. That means is never experienced purely as a result.The given is a whole consisting of
various unknown parts, and the problem is to find them. This implies also fluidity of dimension.
What the designer does is temporal: the liberation of the elements, their arrangement in subsid-
iary groups, simultaneous destruction and construction towards the whole. After a while some-
one stops and goes away or if something strikes him returns and may approach things in an
entirely different way. Abstract formal elements are put together to make concrete or abstract
things. In the end the form is achieved by a ceaseless action. The computational interest ap-
pears in the manipulation of elements in terms of parts and in terms of dimension. Rule applica-
tion is equal with a decision to start looking, and more importantly manipulating, the same
element in a different way with the only aim to explore the implications for the composition. The
possibilities are infinitely many, and as most designers experience decisions in this procedure
cannot be definite beforehand . Questions remain to be answered, in relation to transitions of
dimension and the possibility for a more general treatment of occurrences of dimensional flux
computationally.
46
47
III. PLANE
Proclus (Proclus, p. 117, 8) defines the plane as “a surface such that a straight line fits on all
parts of it” and Aristotle (Metaph. 1020 a 12) as a magnitude extended two ways, or “divisible in
two ways”. Aristotle (De anima I. 4, 409 a 4) mentions the common remark that “a line by its
motion produces a surface” or that is “the extremity of a solid” and Euclid (Elements, I. 7.)
states that “a plane is a surface, which lies evenly with the straight lines on itself”. In the modern
approach Bolyai and Lobachewsky (Elemente der absoluten Geometrie, 1876) evolved the
plane as “the locus of all points equidistant from two fixed points in space”. Leibniz (Uber
Euklidische Geometrie, 1888) defines plane as “the surface which divides space into two
congruent parts” with the further condition that the two congruent surfaces could be “slid along
each other without the surfaces ceasing to coincide”.
Kandinsky and Klee refer to the plane in two ways. Plane is a planar shape or a form with
closed boundaries. It is also the basic graphic surface. The material plane of the graphic sur-
face becomes the finite graphic context that has closed linear boundaries and is approached as
a shape of the composition. What is of importance, in this case, is the interaction between the
plane and the other basic elements that appear on it. The relativity of value of all forms, and the
ambiguity of their dimension can be exposed to some degree in comparison with other forms,
and in comparison with this basic finite graphic context. This context and its linear parts may
define the relative value of an isolated form as well as of the whole sum of forms.
1. Planar forms
Klee (1956, p.112) generates the plane from the moving line. The trace of a constant motion of a
straight line on the plane can be perceived as a plane with rectangular or square shape. Also,
linear rotation about a point generates a planar circle. The line becomes invisible as element, or
passive according to Klee (1956, p.112) while the plane becomes visible or active.
The above plane can be expressed computationally as a weighted or colored plane in a W22
algebra or as a simple plane in a U 22 algebra. But what Klee sees as plane form, on a plane
surface, is what emerges from the coloring of the area of a closed shape. He names it, transi-
tion from the medial shape (or inactive plane) to the active plane. Then, each active planar
element corresponds to a medial shape, which is equal with the shape of its boundaries. What
makes the plane visually active is the “activation” of its full area, and the parallel visual
“disactivation” of its boundaries as distinct elements. A compound U 12 , U 22 algebra, which
48
contains both the linear boundaries and their enclosed planar area, would support this effect. In
the case of the passive plane the planar part that is contained in the U 22 algebra, remains
inactive, while the linear part in the U 12 algebra is visible or active. Then, the activation of the
weighted or colored planar part could be equated with what Klee names visual transition form
the passive to the active plane.
U 12 U 22 U 12 U 22
In another derivation of the plane from the line Kandinsky (1924, p.60) organizes straight lines
about a common meeting point, according to the rule
After several repetitions the shape becomes denser and the intersections form a more compact
center, in which the point seems to grow. As in the case of the point, which becomes plane
Kandinsky attempts to express computationally common intuitions about basic elements, and
their transitions from one dimension to another.
... ⇒
The point can be seen as a center about which lines move and finally, flow, into one another to
construct a plane. This derivation of the plane is similar with the one we see in Klee and has
Aristotelian origins.
Kandinsky
Kandinsky introduces the notion of the plane as emergent shape. The plane can emerge from
accumulation of lines on the graphic plane. It is stated that this is a characteristic property of
the line.
49
The abstract plane in its general form is a closed line. This concept is illustrated in both Klee
and Kandinsky in accordance to what Euclid would have called plane figure or shape: “that
which is contained by any type of linear boundaries”. Accordingly, Kandinsky organizes planes
from the form of their linear boundaries. He distinguishes them as simple geometrical planes
and their derivatives, displaced geometrical planes, and free planes. Simple geometrical planes
are considered only three planar shapes, the triangle, the square and the circle. With the
derivatives form the following group of six shapes.
Triangle  Trapezium Square  Parallelogram Circle  Ellipsis
Displaced geometrical planes are all planar shapes contained by straight or geometrical curved
lines and not classified as simple geometrical planes or their derivatives. Here are included all
possible parts of simple geometrical planes, and all emergent planar shapes that have no free
lines as boundaries.
Free planes are all planar shapes contained by closed free lines.
This organazation of planar shapes is part of the analytic approach of the properties of graphic
elements and the establishment of possible relationships. Here the distinct attributes of straight
and free lines, or acute and right angles become the criteria for the classification of shapes.
However, for Klee (1956, p. 60) and Kandinsky (1924, p. 74) exactness should never be one
sided. Graphic forms cannot be conceived apart from rules that give them life. The indispens-
able imaginary part is incorporated in rule application. When Klee declares that “what we are
after is not form but function” expresses this concept. The work, he explains “is related to the
rules inherent in it. It grows in its own way, on the basis of these rules. But the rules are not a
priori”. Klee characterizes the work as projection and as phenomenon “for ever starting” and
“for ever limited”. It does match the infiniteness of the rules in this: even in its limited sphere
the reckoning does not come out even. Thus, he suggests, “consider the actual with benevo-
lence but do not define today, define backwards and forwards, spatial and many sided. A de-
fined today is over and done for”.
50
1.1 Interaction of planes

Colored or weighted planes that are manipulated on the plane in a W 22 algebra possibly overlap.
In this case their common planar part, which is the product of what Klee (1956, p. 117) names
two-dimensional contact on the plane, or planar interpenetration, becomes a territory that
demands treatment according to constructive and qualitative relations. Klee notes that when
planar shapes interact this way their common parts can intermesh, or one absorbs the other.
The following example illustrates schematically the basic possibilities for interaction between
two colored planes in W 22 algebra.
Klee
In the first case their common part becomes of a third distinct quality, while in the second and
third the quality of one or the other plane becomes quality of their common part. When numer-
ous planar shapes overlap the treatment of their common parts can be more consequential for
the composition. The overlapping can be approached in a variety of ways especially when
different colors and textures interact. In architectural representation, the interaction of planes
except from visual results, like the creation of graphic depth, can also signify the interaction
between architectural elements. In handling this qualitative relationship between planes, at least
two alternatives appear computationally plausible. As is examined in Knight 1989, we can
create a graphic “scale”, which is a set of qualitative rules that establishes the “ranking” of
planes and defines the qualities of their common parts beforehand, or, we can treat them by
defining the preferable result with a shape rule each time the overlapping occurs.
1.2 Actual and perspective planes

Klee (1956, p. 145) introduces the idea of perspective three-dimensional space within the two-
dimensional graphic plane as manipulation of planes and their boundaries, lines. Accordingly
the horizontal line can be seen as a surface at eye level, and two trapeziums correspond to the
view from above and from below.
Klee
51
Here, operations that applyn on the graphic plane combine elements to create a higher-dimen-
sional structure. Rules apply on planes and lines to produce this other kind of illusionistic
space. The planes can be seen as actual planes or as perspective planes, horizontal or vertical.
Also progression of length and weights of lines produces the same visual result.
Klee
This arrangement of elements on the plane creates the tendency for the viewer to orient himself
by this new dimension. Major-minor, large or small components, brilliance-darkness, behind and
front, and so the desire to bring in a third dimension gives rise to very simple rules. But what
Klee has in mind is not the conventional perspective: “the value of the whole process lies solely
in the possibility of checking. There is no merit to drawing in proper perspective “. Instead he
establishes several viewpoints, or centers and he organizes lines and planes according to a
specific graphic scale of deviation, from the horizontal or the vertical. The result is a natural
deviation from static relations.
Klee
1.3 Activation of planes

The explicit reference to medial shapes, and the use of active and inactive planes, from Klee
and Kandinsky is an intuitive manner to approach the ambiguity of shapes computationally. It
reveals the expressive attributes of lines and linear arrangements. Lines can be seen as them-
selves or as boundaries of planes.
52
What remains implicit is that often in calculations of form we use lines, and at the end we treat
their products as we prefer, in terms of dimensional properties and parts. The above becomes
apparent in computations with planes on the plane. The visually active planar form is not always
an economic element for computations, but its linear boundaries are. Thus, we can use them
instead of the total area of the planar shape, and “activate” the planar part at the end. In other
cases shapes can be activated as planes, by a figure-grounding effect that can be produced by
accentuation of contrast in their linear boundaries. Klee (1956, p. 51) offers a rule of qualitative
treatment of lines. He names it endotopic-exotopic treatment.The rule, which describes the
movement of boundary contrast, is the following: “In handling boundary contrast, always stay on
one side of the line”.
Klee
This becomes particularly useful when we want to create the illusion of depth by employing the
visual effect of overlapping, interpenetrating, and discrete planes. Here the boundaries become
spatial and progression produces the effect of depth. The spatial character of the plane is
imaginary, and depends on interpretation, but still is one of the more expressive features of two-
dimensional representation that we want to treat computationally. Thus, the treatment of the
linear boundaries of shapes becomes an important factor of visual expression.
In the example the first rule places one linear square into another and the second moves it on
one of the sides. Each time we assign a weight to one side of the line in a W 12 algebra. The
application of contrast on the linear boundaries of the square activates the plane and the linear
shape can be perceived as a plane.
→ →
53
The interaction between two squares and the interaction with the graphic plane creates the
visual effect of overlapping planes in three-dimensional space. This transition of dimension can
be handled computationally by a compound W 12 , W 22 , algebra, which contains weighted lines
and weighted planes on the plane.
after rules apply to simple and weighted lines (U 12 , W 12 ) to create shapes on the plane. In the
example five maximal lines compose a shape in W 12 .
The principle of endotopic-exotopic treatment can also apply to linear figures containing inter-
sections.
54
The endotopic or exotopic rule that applies on five lines creates an ambiguous visual effect. The
linear shape that lies on the plane becomes an arrangement of planes, or lines, in the illusionis-
tic three-dimensional space of the graphic plane. Emergent visual phenomena like the above
exhibit vividly how our perception about visual things may change unexpectedly. Similar intel-
lectual constructions require more clarification if they are to serve as constructive rules, but
nonetheless expose the expressive power of abstract basic elements and of rules, and their
intrinsic ability to generate ambiguous perceptual shifts from one dimension to another.
2. Graphic context
When basic elements are manipulated on the plane (j=2), the graphic context can be seen as a
basic plane. Kandinsky (1924, p. 114) observes that pairs of linear segments usually bound this
context which in the case of painting is considered finite. Accordingly, the context in architec-
tural composition has its own limits. In the case of a plan the shape and the boundaries of the
site define this finite context for the composition, while in sections and elevations an upper,
height- limit, and a lower ground-limit usually exist for the building. The shape of the site be-
comes in an abstract sense the graphic environment that accommodates the work and becomes
part of the composition of the plan. In sections and elevations, emphasis on the verticality or
the horizontality of the building provides a different context that individual architectural ele-
ments cannot alter. The shape of the finite context that is considered as a basic plane in the
case of plans, section and elevations can take any form: complex many-angled shapes, circular,
oval or free forms, while rectangulars appear the relatively most common in use shapes. This
finite graphic context is contained into boundary lines, which possess strong perceptual signifi-
cance, not only as the boundaries of a shape but also as limits of space.
T. Ando
Each part of the finite context is individual and has its own character. The perceptual conven-
tion of “above and below”, that one experiences when he stands, as a standard relationship
between the three dimensional space and the body, applies on the characterization of the “top
and bottom” boundary lines of the building, in section and elevation. Above and below constitute
a contrasting pair, that has implications on the perception of “weights” of elements. However,
55
since the origin of the composition remains abstract, the application of this concept is relative.
The visual convention of “right and left” amounts to the identification of the left and right limit for
the composition. By convention this correspondence applies in mirrored reflection in relation to
the human body. Right and left constitute the second pair of boundaries that has implications for
the manipulation of elements. The graphic description of the site in plan is precisely one shape.
This means that the shape and the area of the site is the same for everyone. However, the
mapping of the qualities of this finite context includes the acceptance of a number of objective
facts for the specific site together with the subjective interpretation and evaluation of a infinite
variety of existing conditions. From this “reading of the site” architects come to descriptions that
vary significantly. The process of the mapping of the site occurs in several steps, in plan and
section. The sequence of these steps is unordered and has both qualitative and quantitative
implications for the composition. New information may be added and subjective evaluation may
change as interpretation alters. The existing conditions include the condition of the site (flat,
hill, river, sea, waterfall, trees, North etc), the relation to the surrounding buildings, construc-
tions, roads, pavements etc. Their descriptions are shapes, in plan, section and elevation.
D. Basic elements on the graphic context

Any form applied on the plane interacts with a finite graphic environment. The composition
demands the placement of the graphic elements not only as graphic quantities or “measures”
but also as qualities in interaction with this finite context. Each boundary invokes different
qualities in relation to the qualities of the site. Therefore proximity of a form to the edge sub-
jects it to different influence. The edges are distinguished from one another only by the degree
of tensions in relation to the site. Each context offers two typical possibilities as regards the
elements that it bears. In the first the elements lie in relation to the boundaries and its shape. In
the second the elements are loosely associated with the shape of this context. The plane
disappears perceptually and the elements “hover” in space. In the following examples only the
square basic plane is been considered. Any other shapes of basic plane makes the visual
interactions considerably more complex than the square.
1. Line on plane - Relation to the border

In U 12 , U 22 algebra, which includes lines manipulated on a bounded plane, the possibilities for
relation of a line with the linear boundaries of the plane are two. The line may be embedded on
a boundary line, or it may be discrete. In the embedding case the line practically disappears
since it fuses visually with the line of the boundary. In the case the line is discrete it still may
touch or not touch the linear boundary of the plane. The relation of the line with the shape of the
bounded plane becomes qualitatively different as the line approaches the boundaries. This
relation changes even more when the line touches a boundary.
Kandinsky
56
A form increases its visual interaction with the shape of the bounded graphic context if is placed
independently from its boundary lines. The main reason is that the shape of the boundaries and
the form appear as two distinct entities that interact visually. Positioning of a form close to a
specific boundary increases the intensity of the interaction with it. This relation becomes looser
as the distance from the boundary becomes bigger. The interaction with the shape of the
boundary changes in character when the form comes into contact with it. In the case of the line,
the element and the shape become visually one shape. However, the line and the linear shape
are still discrete. The following elementary arrangement of lines illustrates an extension of the
above
Kandinsky
In the first case the arrangement becomes a typical expression of rigidity, while in the second
the same arrangement of lines detached from the boundaries acts as an independent shape in
relation with the shape of the plane. In a relative example the diagonals of the left square
remain centered while the vertical and horizontal move of the center. The arrangement becomes
more complex. The acentric arrangement of lines of the right square intensifies the complexity
while the contact of the vertical line with the upper boundary of the plane restraints the instabil-
ity of the arrangement.
Kandinsky
2. Plane on plane
In the case that the line becomes a plane, the shape of the boundary and the shape of the
planar element do not fuse visually into one shape but their interaction alters when the shape
moves closer to the boundary. If one, or two boundary lines of the element are embedded on
the boundary lines of the plane, the shape still retains its individuality but also the shape of the
boundary is reinforced.
An additional observation is that forms “lying near” the boundary augment the dynamic interac-
tion of the composition with the context, while those lying more about the center augment
stability and give more “objective” character. The same visual results occur if we place a larger
plane on the basic plane. For the illustration of this case the general proportions of the closed
shape approach those of the basic plane.
57
Upon these bases, further principle directions of forms can be superimposed, either remaining
central or else moving away from the center in various directions. And of course, the center can
be avoided altogether- the number of constructional possibilities is infinite.
CONCLUSIONS
The idea that the process of composition in design is abstract is not new. In this process combi-
nations of immaterial graphic elements produce forms that express the schematic quality of our
imaginary spatial ideas with great precision. Basic graphic elements are points, lines, and
planes that combine in one, two, and three-dimensional space according to rules.
Computationally, and as means of expression basic elements exhibit two different kinds of
ambiguity: ambiguity of division, and ambiguity of dimension. These give rise to a series of
expressive potentials and have important constructive use in design. Consequently, it becomes
important for the process of composition to handle computationally both kinds of ambiguity.
Basic elements have been discussed exhaustively by mathematicians and artists. Points, lines,
and planes are mathematically predefined but their graphical use demands a different approach.
In their theories on non-representational composition Paul Klee and Wassily Kandinsky sug-
gested a computational model for the intuitive treatment of basic elements. Their effort was
propelled by a practical need to define equilibrium between intuition and calculation. Kandinsky
and Klee accepted the notion of dimensional flux and solved it by introducing medial elements.
A conspicuous similarity between these approaches and the shape computation theory is that
all three are expressive systems that are concerned mainly with the process of design and less
with the finished artefact. They use basic elements and rules to construct designs dynamically.
Shape grammars deal with the dimensional ambiguity of basic elements in three ways:
a) use of compound algebras and weights
b) use of shape rules
c) parallel computation
The concept of an algebra that will accommodate medial shapes as well as that of finite graphic
context are subjects of probable future research.
58
59
REFERENCES
Ando, Tadao 1990 Sketches, edited by Werner Blaser (Basel: Birkhauser)
Calas, Nicolas 1997 Art in the Age of Risk, and other Essays (New York: Dutton)
Chase S C 1989 ‘Shapes and shape grammars: from mathematical model to computer imple-
mentation’ Environment and Planning B: Planning and Design 16 pp. 215-242
Earl C F 1986 ‘Creating design worlds’ Environment and Planning B: Planning and Design 13
pp. 177-188
Earl C F 1997 ‘Shape boundaries’ Environment and Planning B: Planning and Design 24 pp.
669-687
‘Euclid’s Elements, the thirteen Books’ introduction and commentary by Sir Thomas L. Heath
1956 (New York: Dover Publications, Inc)
Flemming U 1981 ‘ The secret of the Casa Guiliani Frigerio’ Environment and Planning B 8 pp.
87-96
Flemming U 1987a ‘More than the sum of parts: the grammar of Queen Anne houses’ Environ-
ment and Planning B: Planning and Design 14 pp 323-350
Flemming U 1990 ‘Syntactic Structures in Architecture’, The Electronic Design Studio, MIT
Press, Cambridge pp. 31-47
Gips J 1975 Shape Grammars and their uses: Artificial Perception, Shape Generation and
Computer Aesthetics (Basel:Birkhauser)
Kandinsky, Wassily 1912 ‘Concerning the Spiritual in Art’ (New York: George Wittenborn, Inc.)
Kandinsky, Wassily 1924 ‘Point And Line to Plane’ (Michigan: Cranbrook Press)
Klee, Paul 1956 ‘The Thinking Eye’ (New York: George Wittenborn, Inc.)
Klee, Paul 1961 ‘Notebooks, Volume 2, The Nature of Nature’ (New York: George Wittenborn )
Klee, Paul 1924 ‘On Modern Art’ (London: Faber and Faber)
Klee, Paul 1953 ‘Pedagogical Sketchbook’ (New York: Frederick A.Praeger, Inc.)
Knight T 1980 ‘The generation of Hepplewhite-style chair back designs’ Environment and
Planning B: Planning and Design 7 pp. 227-238
Knight T 1981 ‘Languages of designs: from known to new’ Environment and Planning B 8 pp.
213-218
Knight T 1981 ‘The forty-one steps’ Environment and Planning B 8 pp. 97-114
Knight T 1986 ‘Transformation of the Meander Motif on Greek Geometric Pottery’ Design Com-
puting 1 pp. 29-67
60
Knight T 1989 ‘Color grammars: designing with lines and colors’ Environment and Planning B:
Planning and Design 16 pp. 417-449
Knight T 1989 ‘Transformations of De Stijl art: the paintings of Georges Vantongerloo and Fritz
Glarner’ Environment and Planning B: Planning and Design 16 pp. 51-98
Knight T 1994 Transformations in Design (Cambridge: Cambridge University Press)
Knight T 1998 ‘Shape grammars’ Environment and Planning B: Planning and Design Anniver-
sary Issue pp. 86-91
Knight T 1998 ‘Designing a Shape Grammar’, Artificial Intelligence in Design ’98, p. 514
Koning H and Eisenberg J 1981 ‘The language of the prairie: Frank Lloyd Wright’s prairie
houses’ Environment and Planning B 8 pp. 295-323
Krstic D 1999 ‘Constructing algebras of designs’ Environment and Planning B: Planning and
Design 26 pp. 45-57
Lupton, Ellen and Abbott Miller, J, editors 1991 ‘The ABCs of : The Bauhaus and
Design Theory’ Princeton Architectural Press, Inc.
March L 1996 ‘The smallest interesting world?’ Environment and Planning B: Planning and
Design 23
March L 1996 ‘Rulebound unruliness’ Environment and Planning B: Planning and Design 23,
p. 396
Stiny G, 1975 ‘Pictorial and Formal Aspects of Shape and Shape Grammars’, Birkhauser, Basel
March L and Steadman Philip 1971 ‘The Geometry of Environment’ (Cambridge: The MIT Press)
Michelis, P. A. 1974 ‘L’esthetique de l’architecture’ (Paris: Klincksieck)
Naylor, Gillian 1985 ‘The Bauhaus Reassassed, Sources and Design Theory’ ( London: The
Herbert Press)
Μπιρηs, Τασοs Κ. 1996 ‘Αρχιτεκτονικηs σηµαδια και διδαγµατα’ (Αθηνα: Μορφωτικο Ιδρυµα
Εθνικηs Τραπεζηs)
Stiny G, 1976 ‘Two exercises in formal composition’, Environment and Planning B 3 pp. 187-210
Stiny G 1977, ‘Ice-ray: a note on Chinese lattice designs’, Environment and Planning B 4 pp.
89-98
Stiny G 1980 ‘Kindergarten grammars: designing with Froebel’s building gifts’, Environment and
Planning B 3, pp. 409-462
Stiny G 1981 ‘A note on the description of designs’, Environment and Planning B 3 pp. 257-267
Stiny G, 1982 ‘Spatial relations and grammars’, Environment and Planning B: Planning and
Design 9 pp. 113-114
Stiny G, 1986 ‘A new line on drafting systems’ Design Computing 1 5-19

61
Stiny G, 1987 ‘Composition counts: A+E=AE’, Environment and Planning B: Planning and
Design 14 pp. 167-182
Stiny G, 1989 ‘Formal devices in design’ Design Theory 88 Eds S L Newsome, W R Spillers, S
Finger (Springer, New York) pp. 173-188
Stiny G 1990 ‘What is a design’, Environment and Planning B: Planning and Design 17 pp. 97-
103
Stiny G 1990b ‘What designers do that computers should’ The Electronic Design Studio: Archi-
tectural Knowledge and Media in the Computer Era, Eds M McCullough, W J Mitchell, P Purcel
(MIT Press, Cambridge, MA) pp. 17-30
Stiny G 1991 ‘The algebras of design’, Research in Engineering Design 2(3) pp. 171-181
Stiny G 1992 ‘Weights’ Environment and Planning B: Planning and Design 19 pp. 413-430
Stiny G 1993 ‘Boolean algebras for shapes and individuals’ Environment and Planning B:
Planning and Design 20 pp. 359-362
Stiny G 1994 ‘Shape rules: closure, continuity, emergence’ Environment and Planning B: Plan-
ning and Design 21 pp. s49-s78
Stiny G 1996 ‘Useless rules’ Environment and Planning B: Planning and Design 23
pp. 235- 237
Stiny G 1998 ‘New ways to look at things’ Environment and Planning B: Planning and Design 25
pp. 71- 80
Stiny G 1999 ‘Commentary: Shape’ Environment and Planning B: Planning and Design 19
pp. 7-14
Stiny G and Gips J 1972 ‘Shape Grammars and the Generative Specification of painting and
Sculpture’, Information Processing 71, Amsterdam, pp. 1460-1465
Stiny G and Gips J 1978 ‘Algorithmic Aesthetics ‘, University of California Press, Berkeley, CA,
Stiny G and W.J. Mitchell 1978 ‘The Palladian grammar’, Environment and Planning B 5 pp. 5-
18
Stiny G and March L 1981 ‘Design Machines’ Environment and Planning B 8 91-104
Valery P 1924 Eupalinos: ou, l’ Architecte, (Paris: Gallimard)

62
ILLUSTRATIONS
Images and drawings included in this study have been made by the author except otherwise
noted.
All images by Klee are sketches included in the publication of his artistic notebokes (1956),
under the general title ‘The Thinking Eye’
(Klee, Paul ‘The Thinking Eye’ New York 1956 : George Wittenborn, Inc.)
All images by Wassily Kandinsky are included in his book ‘Point and Line to Plane’
(Kandinsky, Wassily 1926 ‘Point and Line to Plane’ Michigan: Cranbrook Press)
The sketch in p. 10 is by Tadao Ando (VITRA, Meditation Pavillion)

The sketches in pp. 23, 24 are by Tadao Ando (Theater on the water and Rokko Housing)
(Ando, Tadao 1990 ‘Sketches’ edited by Werner Blaser Basel: Birkhauser)
The illustration of symmetry groups on the plane in p. 21 are by March and Steadman
(March L and Steadman Philip 1971 ‘The Geometry of Environment’ Cambridge: The MIT Press)
The illustrations on pp. 24, 25, on the use of points in architecture are studio projects made by
the students of the National Technical University of Athens, Greece. They are included as visual
examples of the process of architectural composition in the book Αρχιτεκτονικηs σηµαδια και
διδαγµατα. (Μπιρηs, Τασοs Κ. 1996 ‘Αρχιτεκτονικηs σηµαδια και διδαγµατα’ Αθηνα:
Μορφωτικο Ιδρυµα Εθνικηs Τραπεζηs), p. 36
View publication stats

Point Line Plane

Uploaded by

Copyright:

Available Formats

Point Line Plane

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Point Line Plane

Uploaded by

Copyright:

Available Formats

What is the thesis about?

What is the thesis about?

What are points, lines and planes examined for?

What are points, lines and planes examined for?

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Article · December 2011

The Connected Sustainable Home View project

The user has requested enhancement of the downloaded file.

Point, Line, Plane:

SUBMITTED TO THE DEPARTMENT OF ARCHITECTURE IN PARTIAL FULFILLMENT OF THE

MASTER OF SCIENCE IN ARCHITECTURE STUDIES

c 2000 Sotirios D. Kotsopoulos. All rights reserved.

Point, Line, Plane:

Thesis Supervisor: Terry Knight

white (light grey) white

white (black) white

Klee “Architecture” Compositional rule for the contrast-relation (Klee)

Kandinsky: Geometric composition

U00 U01 U02 U03

U11 U12 U13

1.3 Size and relation to other forms

1.1 Typical straight lines

1.2 Atypical straight lines

2. Angular and curved lines

2.1 Angular line

2.1.1 Angles and lengths

2.1.2 Compositional parameters of angular lines

α. The lengths of their straight parts.

β. The inclination of their angles (more or less acute).

γ. The conquest of the plane.

2.2 Curved line

Curve geometric wave-like

1. Lines, angles and shapes

k => => =>

Repetition of straight line with alteration of weights or intervals:

Repetition, or opposed repetition of angular lines:

Repetition, or opposed repetition of curved line:

Contrasting repetition of curved line:

3. Linear, medial, and planar shapes

(import shape (import shape

Derivation B: using only rule 2 Derivation C: using only rule 3

1. Boundaries and texture

Triangle  Trapezium Square  Parallelogram Circle  Ellipsis

1.1 Interaction of planes

1.2 Actual and perspective planes

1.3 Activation of planes

D. Basic elements on the graphic context

1. Line on plane - Relation to the border

Ando, Tadao 1990 Sketches, edited by Werner Blaser (Basel: Birkhauser)

Knight T 1994 Transformations in Design (Cambridge: Cambridge University Press)

Michelis, P. A. 1974 ‘L’esthetique de l’architecture’ (Paris: Klincksieck)

Stiny G, 1986 ‘A new line on drafting systems’ Design Computing 1 5-19

Valery P 1924 Eupalinos: ou, l’ Architecte, (Paris: Gallimard)

The sketch in p. 10 is by Tadao Ando (VITRA, Meditation Pavillion)

View publication stats

You might also like