Chapter 4.

3
Cognitive Maps and Spatial Behaviour:
Process and Products

Roger M. Downs and David Stea

Editors’ overview something newly discovered [. . .]. Instead, we are con-

cerned with phenomena so much part of our everyday
At the start of the 1970s, it was intellectually fashionable amongst lives and normal behaviour that we naturally overlook
behavioural geographers to investigate the significance of cog- them and take them for granted.
nitive maps, and their impacts on people’s spatial behaviour. [. . .]
Downs and Stea’s book was probably the most influential overview We find that planners try to alter cognitive maps,
of the field and brought together papers from almost all of the astronauts need them, the news media use them, and
leading exponents of this kind of research. We have excerpted
advertisers tempt us with them; they are part of our
Chapter 1, which explores the dimensions of cognitive mapping,
distinguishing between cartographic images and the cognitive
everyday lives. [. . .] We offer a formal definition: cognitive
constructs that are the focus of their attention. This conceptual mapping is a process composed of a series of psychological
piece is informed by a communications model of information transformations by which an individual acquires, codes,
transmission and explores processes and defines concepts under- stores, recalls and decodes information about the relative
pinning research. The authors define the concepts of perception, locations and attributes of phenomena in his everyday
cognition, attitude and preference, before explaining the differ- spatial environment.
ences between what people need to know and what they actually In this paper we will expand this definition and examine
know. Amongst other concepts they focus on differences between the conceptual frameworks which are subsumed within it.
locational and attribute information, the role of incomplete,
distorted, schematised, and augmented cognitive maps, and
some of the behavioural reasons for the mismatch between theory An analysis of cognitive mapping
and practice. They conclude by urging further experimental
investigation of behavioural evidence of cognitive mapping. processes
Originally published in 1973: Chapter 1 in Roger M. Downs and Cognitive maps and adaptive behaviour
David Stea (eds) Image and Environment: Cognitive Mapping and
Spatial Behavior, Aldine Press, Chicago, 8–26. Underlying our definition is a view of behaviour which,
[. . .] can be reduced to the statement that human spatial
behaviour is dependent on the individual’s cognitive map of
Introduction the spatial environment. That this formulation is necessary
is indicated by a comparison of the characteristics of the
A surprising fact is associated with studies of cognitive individual with those of the spatial environment.
mapping: although the emergence of this vigorously devel- The environment is a large-scale surface, complex in
oping research area has been recent, we are not discussing both the categories of information present and in the

The Map Reader: Theories of Mapping Practice and Cartographic Representation, First Edition. Edited by Martin Dodge, Rob Kitchin and Chris Perkins.
© 2011 John Wiley & Sons, Ltd. Published 2011 by John Wiley & Sons, Ltd. ISBN: 978-0-470-74283-9
 ROGER DOWNS AND DAVID STEA 313

number of instances of each category. Things are nether which has the functions of the familiar cartographic map
uniformly distributed, nor ubiquitous; they have a but not necessarily the physical properties of such a pic-
‘whereness’ quality. In contrast the individual is a relatively torial graphic model (Blaut et al. 1970). [. . .] The carto-
small organism with limited mobility, stimulus seeking graphic map has a profound effect on our concept of a
capabilities, information processing ability, storage capac- cognitive map.
ity and available time. The individual receives information Spatial information can be represented in a variety of
from a complex, uncertain, changing and unpredictable ways. [. . .] All media share the same function not structure,
source via a series of imperfect sensory modalities, and thus cognitive maps are derived from analogies of
operating over varying time spans and intervals between process, not product.
time spans. From such diversity the individual must aggre-
gate information to form a comprehensive representation
of the environment. This process of acquisition, amalgam- Cognitive mapping signatures
ation and storage is cognitive mapping, and the product of
this process at any point of time can be considered a
and cognitive representations
cognitive map. [. . .] All of the media rely upon the same sort of spatial
Given a cognitive map, the individual can formulate the information, and all are employed in the same sorts of
basis for a strategy of environmental behaviour. We view spatial behaviour: thus the inputs and outputs are
cognitive mapping as a basic component in human adap- specified, while the intervening storage system (the
tation, and the cognitive map as a requisite both for human black box) is not. The way in which spatial information
survival and for everyday environmental behaviour. It is a is encoded (mapmaking) and decoded (map reading or
coping mechanism through which the individual answers interpreting) gives rise to a set of operations called the
two basic questions quickly and efficiently: where certain signature of a given mapping code. Thus a cartographic
valued things are, and how to get to where they are from map signature is dependent upon three operations: rota-
where he is. tion of point of view to a vertical perspective, change in
scale, and abstraction to a set of symbols [. . .]. Many other
signatures are feasible; we have no reason to anticipate
Cognitive maps and spatial behaviour that cognitive maps should necessarily have the same
form of signature as cartographic maps. Above all, we
[. . .] We believe that a cognitive map exists if an individual should avoid getting locked into a form of thinking
behaves as if a cognitive map exists (Stea and Downs 1970). through which we as investigators force a subject to
[. . .] Normal everyday behaviour such as a journey to work produce a cartographic cognitive map and which we
[. . .] would be impossible without some form of cognitive then verify against an ‘objective’ cartographic map. [. . .]
map. [. . .] Admittedly, much spatial behaviour is repeti- The issue of mapping signatures involves some funda-
tious and habitual [. . .] but even this apparent stimulus mental theoretical and methodological issues. [. . .] Under-
response sequence is not so simple: you must be ready for lying the whole approach is the basic question: How is
the cue that tells you to ‘turn here’ [. . .]. You are thinking information derived from the absolute space of the envi-
ahead and using your cognitive map. In human spatial ronment in which we live, transformed into the relative
behaviour we consider even a series of stimulus–response spaces that determine our behaviour? The transformation
connections as a ‘simple’ (or impoverished) form of a can be viewed [. . .] as involving any or all of three
cognitive map, in which the general aspects of spatial fundamental operations: change in scale, rotation of per-
relationship implicit in cognitive mapping play a minimal spective and a two-stage operation of abstraction and
role. [. . .] The person knows that an object is valued and symbolisation, all of which result in a representation of
one way of getting to it, but knowledge of the whereness in relative space.
relation to the location of other objects is absent. [. . .] Thus [. . .]
someone who knows only one route knows more about Thus, we should be interested in developing theoretical
that route than just the appropriate responses at certain statements about the cognitive signatures that are
choice points and, because he thinks ahead, is also engaging employed in dealing with information from the spatial
in cognitive mapping. We are postulating the cognitive environment. [. . .] The only differences between
map as the basis for deciding upon and implementing any Lynch’s (1960) ‘images’ and city maps of cartographers
strategy of spatial behaviour. lie in the degree of abstraction employed and the types of
However, we must make it perfectly clear that a symbols chosen to depict information. [. . .] We should be
cognitive map is not necessarily a ‘map’. [. . .] We are concerned with the nature or signature of relative space as it
using the term ‘map’ to designate a functional analogue. is construed and constructed by the individual. Only if we
The focus of attention is on a cognitive representation, do this can we ask how relative and absolute spaces
 314 CHAPTER 4.3 COGNITIVE MAPS AND SPATIAL BEHAVIOUR: PROCESS AND PRODUCTS

compare and differ. [. . .] Some aspects of our composite offered by Stea (1969). He suggests that cognition occurs
cognitive maps may resemble a cartographic map; others in a spatial context when the spaces of interest are so
will depend upon linguistic signatures (in which scale and extensive that they cannot be perceived of apprehended at
rotation operations are irrelevant), and still others upon once [. . .]. This scale dependent distinction [. . .] also
visual imagery signatures derived from eye-level viewpoints suggests that we are concerned with the nature and
(in which the scale transformation may be disjointed or formation of environmental cognitions rather than with
convoluted). [. . .] briefer spatial perceptions.
Before considering the nature and functions of cognitive
maps in more detail, we must discuss some basic defini-
tions and attempt to clarify a few misconceptions which Attitudes, predictions, preferences
currently prevail. and cognitive maps
[. . .] The parallels between the concepts of cognitive map
and attitude are marked. [. . .] Fishbein (1967) replaces the
The concepts of perception, holistic concept of an attitude with a formation containing
cognition, attitude and preference three components: cognitions or beliefs, affect or attitude
and conations or behavioural intentions. Fishbein claims
Perception and cognition: distinctions that the fact that affect, cognition and action are not always
highly correlated necessitates this more complex typology.
Unfortunately, perception and cognition have been [. . .] Fishbein points out that attitudes, beliefs and
employed in a confusing variety of contexts by psycholo- expressed behavioural intentions are frequently brought
gists and other social scientists. [. . .] It is difficult to into line with actual behaviour. [. . .] In other words, if the
determine whether the process of perceiving is being dis- behaviour can be specified an attitude can usually
cussed or [. . .] the product of the perception process. [. . .] be postdicted.
Perception has been used in a variety of ways: to experi- Finally, we must distinguish among attitudes, prefer-
mental psychologists it involves the awareness of ences and traits. [. . .] Preferences are usually considered to
stimuli through the physiological excitation of sensory be: less global [. . .]; and less enduring over time. [. . ..]
receptors; to some social psychologists it implies both When a given attitude pervades a wide variety of objects
the recognition of social objects present in ones immediate over a considerable period, it becomes a personality trait.
sensory field and the impressions formed of persons or [. . .]
groups experienced at an earlier time. To many geogra- Hypothetically, one could construct a scale from pref-
phers perception is an all-encompassing term for the erence through attitude to trait, increasing in both inclu-
sum total of perceptions, memories, attitudes, preferences siveness and duration of the cognitive, connotative and
and other psychological factors which contribute to the effective components. [. . .]
formation of what might better be called environmental
cognition. [. . .]
Given the varied uses of the terms it is difficult to The nature and function of cognitive
distinguish between perception and cognition. [. . .] We
reserve the term perception for the process that occurs
maps
because of the presence of an object, and that results in the What do people need to know?
immediate apprehension of that object by one or more of
the senses. Temporally it is closely connected with events in [. . .] There are two basic and complementary types of
the immediate surroundings and (in general) linked with information that we must have for survival and everyday
immediate behaviour. [. . .] Cognition need not be linked spatial behaviour: the locations and attributes of phenom-
with immediate behaviour and, therefore, need not be ena. Cognitive maps consist of a mixture of both. [. . .] We
directly related to anything occurring in the proximate must also know what an object is.
environment. [. . .] Locational information is designed to answer the ques-
However this distinction falls short of establishing a tion, ‘where are these phenomena?’ and leads to a subjective
clear dichotomy [. . .]. Both refer to inferred processes geometry of space. There are two major components of this
responsible for the organisation and interpretation of geometry, distance and direction. Distance can be mea-
information [. . .]. Cognition is the more general term sured in a variety of ways and we are surprisingly sensitive
and includes perception as well as thinking, problem to distance in our everyday behaviour. [. . .] Knowledge of
solving and the organisation of information and ideas. distance [. . .] is essential for planning any strategy of spatial
A more useful definition from a spatial point of view is behaviour. [. . .] Direction is no less important in the
 ROGER DOWNS AND DAVID STEA 315

geometry of space, although we are less conscious of existence is defined by the presence of phenomena in the
directional information. [. . .] subject’s cognitive representation. [. . .] [However] we
By combining distance and direction we can arrive at must be careful in interpreting the absence of phenom-
locational information about phenomena, but not nec- ena from cognitive maps as reflecting discontinuity
essarily the same as that provided by Cartesian co of space.
ordinates of cartographic map. [. . .] Thus, locational
information is not as simple as it might appear. We must
store many bits of distance and direction data to operate
efficiently in a spatial environment, a process involving Distortion and schematisation
relatively accurate encoding, storage and decoding. Use By the distortion of cognitive maps we mean the cognitive
of locational information [. . .], however, requires a transformations of both distance and direction, such that
second type of information: that concerning the attri- an individual’s subjective geometry deviates from the
butes of phenomena. Euclidian view of the real world. Such deviations can
Attributive information tells us what kinds of phenom- have major effects on the patterns of spatial use of the
ena are out there and is complementary to locational environment. [. . .] If people are sensitive to distance,
information. [. . .] An attribute is derived from a charac- consequent spatial behaviour patterns will be dependent
teristic pattern of stimulation regularly associated with a on such distance distortions.
particular phenomenon, which, in combination with other Far more significant and as yet little understood are the
attributes, signals the presence of the phenomenon. [. . .] results of schematisation (the use of cognitive categories
We can divide attributes of phenomena into two major into which we code environmental information). We are,
classes: descriptive, quasi-objective or denotative; and as Carr (1970: 518) suggests, victims of conventionality.
evaluative or connotative. [. . .] Here we are separating This conventionality can be expressed in two ways. The first
attributes which are affectively neutral (descriptive) from involves the use of those spatial symbols to which we all
those that are affectively charged (evaluative). subscribe and which we use both as denotative and con-
[. . .] An object is identified and defined by a set of notative shorthand ways of coping with the spatial envi-
attributes and bits of locational information. However, ronment. [. . .] However, there are other symbols dealing
what is an object at one spatial scale can become an with geographic entities [. . .] which owe their cogency
attribute at another [. . .]. The scale of analysis of the and importance to their mere existence. [. . .] Such entities
problem at hand defines what is an object and what is have been termed the invisible landscape. As images these
attributive and locational information. elements are perhaps the most purely symbolic. [. . .]
A second aspect of schematisation or conventionality
involves the very limited set of cognitive categories or
What do people know? concepts that we have developed in order to cope with
information derived from the spatial environment. [. . .]
If we compare a cognitive map with a base map of the real Our understanding of the semantics of cognitive maps is
world [. . .] we find that cognitive mapping does not lead to remarkably limited.
a duplicative photographic process [. . .], nor does it give an The controversy over linguistic relativity suggests that
elaborately filed series of conventional cartographic maps there are cross-cultural differences in the ways in which
at varying spatial scale. Instead cognitive maps are com- spatial information is coded. Such differences are not only
plex, highly selective, abstract, generalised representations cross-cultural. [. . .] Downs (1970) assumed that a neigh-
in various forms. [. . .] We can characterise cognitive maps bourhood shopping centre would be clearly defined and
as incomplete, distorted, schematised and augmented, and commonly agreed upon spatial unit, with the edge of the
we find that both group similarities and idiosyncratic commercial area defining the shopping centre boundary.
individual differences exist. However, residents of the area recognised four distinct
subcentres.
The incompleteness of cognitive maps
The physical space of the real world is a continuous Augmentation
surface which we have come to understand through a
classic geometrical framework: that of Euclid. [. . .] [. . .] There is some indication that cognitive maps have
There are no gaps or bottomless voids. [. . .] Yet all non-existent phenomena added as embroidery. [. . .] Such
cognitive maps are discontinuous surfaces. Seemingly distortions may be highly significant, but we know little
some areas of the earth’s surface do not exist when their about their causes, and nothing about their eradication.
 316 CHAPTER 4.3 COGNITIVE MAPS AND SPATIAL BEHAVIOUR: PROCESS AND PRODUCTS

Inter-group and individual differences in cognitive to. [. . .] Reinforcement and checking are continuous:
maps and mapping erroneous beliefs about locational and attribute informa-
tion are rapidly corrected by feedback from spatial
[. . .] Underlying group perspectives are the result of three behaviour.
factors. Firstly, the spatial environment contains many Vicarious information about the city is by definition
regular and recurrent features. Secondly, people share second-hand. [. . .] This is true of a verbal description, a
common information processing capabilities and strate- cartographic street map, a TV film, a written description, a
gies [. . .]. Thirdly, spatial behaviour patterns display colour photograph or a painting. In the mapping context
similar origins, destinations and frequencies. These fac- these modes of representation, though similar in function,
tors in combination yield inter-group differences in are different in form because they display different signa-
cognitive maps. tures. [. . .] The result of this filtering is a complete repre-
The individual differences among cognitive maps sentation, which varies with both the individual and his
emerge primarily from subtle variations in spatial activity group membership [. . .]. We may not be able to translate
patterns. [. . .] Such idiosyncrasies are particularly notable from the signature of the street map back to the spatial
in verbal descriptions of cognitive maps – the choice environment. [. . .] Both active and passive information
of visual details shows tremendous variation from subject processing are tied to the spatial environment, and result
to subject. from symbolic elaboration, embroidery and augmentation.
In answer, therefore, to the question, ‘What do we [. . .]
know?’, we can conclude that we see the world in the Thus we have three sorts of information available to us at
way that we do because it pays us to see it in that way. [. . .] any point in time. Each has distinct characteristics, validity
People behave in a world ‘as they see it’ – whatever the flaws and utility.
and imperfections of cognitive maps, they are the basis for
spatial behaviour.
A terminology for change
How do people get their knowledge?
To this point our whole discussion of cognitive mapping
We have postulated a set of basic characteristics that our has been static – concepts of learning time and change have
knowledge of the spatial environment should possess, and been omitted. [. . .] We acquire the ability to know things
we have indicated the characteristics that our knowledge about our environment through a process of development.
(or cognitive map) actually possesses. Some of the [. . .] Development clearly includes change; taking place
differences [. . .] can be attributed to the ways we acquire over a considerable period, such change is assumed to be
spatial information. irreversible and [. . .] is also regarded as progressive. Devel-
opment encompasses both growth [. . .] and maturation.
Sensory modalities [. . .] What effects or learned changes can spatial informa-
tion induce? Boulding (1956) suggest three possibilities: no
In our studies of cognitive maps we have overlooked the effect, simple accretion and complete reorganisation. The
range and number of sensory modalities through which no effect case is the most frequent in the normal adult: the
spatial information is acquired, and have ignored the information simply confirms what he already knows [. . .].
imaginative nature of cognitive processes [. . .]. The visual, Most of the spatial information that we receive, although
tactile, olfactory and kinaesthetic sense modalities combine essential for the successful use of the environment of any
to give an integrated representation of any spatial envi- point in time, has no effect on the stored knowledge or
ronment. The modalities are complementary despite our cognitive map.
intuitive belief [. . .] that visual information is predomi-
nant. [. . .] The quality of distinctiveness or memorableness
is not solely the result of the way the environment looks.
[. . .] A typology of change: accretion,
diminution, reorganization
Direct and vicarious sources of information
The simple accretion case relates to minor changes to the
Sources have a different degree of validity, reliability, cognitive map. [. . .] Both locational and attribute infor-
utility and flexibility. Direct sources involve face-to- mation are added to the cognitive map: a simple additive
face contact between the individual and, for example, a change has occurred through learning. [. . .] Diminution
city and information literally floods the person from all develops directly from deletion. There is no need to
his sensory modes. He must be selective in what he attends assume that cognitive maps undergo only progressive
 ROGER DOWNS AND DAVID STEA 317

change [. . .] Either through the passage of time or through Stea, D. (1969) The measurement of mental maps: an exper-
maturation we forget – the amount of information avail- imental model for studying conceptual spaces, in Behavioral
able through the cognitive mapping diminishes. [. . .] All Problems in Geography: A Symposium (eds K.R. Cox and R.
stored knowledge is subject to this time decay: we need to G. Golledge), Northwestern University Press, Evanston, IL,
repeat a spatial experience in order to remember the route pp. 228–253.
in the future. [. . .] Diminution may also be an adaptive Stea, D. and Downs, R.M. (1970) From the outside looking
process. [. . .] Given our limited capacity to store and in at the inside looking out, Environment and Behavior, 2,
handle information, diminution maybe [. . .] ensuring 3–12.
that excess information is lost but important information
retained. [. . .]
The most dramatic changes in cognitive maps are the Further reading
result of total reorganisation. Boulding (1956) suggests that
Blaut, J.M., Stea, D., Spencer, C. and Blades, M. (2003) Mapping
images are relatively resistant to change in their overall as a cultural and cognitive universal. Annals of the Association
nature. It requires an accumulation of contrary evidence of American Geographers, 93 (1), 165–185. [A comprehensive
before complete reorganisation can occur. [. . .] The most meta study reviewing the cross cultural evidence for cognitive
frequent spatial example of such a complete reorganisation mapping as a common human trait.]
is to be found in long-distance human migration and Fabrikant, S.I. and Lobben, A. (2009) Cognitive issues in
subsequent residential site selection. [. . .] geographic information visualization. Cartographica,
We have examined some aspects of our cognitive maps 44 (3). [This themed issue includes a number of useful
and how they came to be. We know they are modes of articles focusing upon the application of cognitive
structuring the physical environment [. . .]. Much of the approaches to geovizualisation some forty years after the
support in contentions concerning their existence is beha- Downs and Stea book.]
vioural, stemming from introspection and anecdotal evi- Kitchin, R. (1994) Cognitive maps: what they are and why study
dence, but the harder experimental data are beginning to them. Journal of Environmental Psychology, 14, 1–19. [A
tightly focused and critical introduction to the nature of
emerge. [. . .] Thus, the face of cognitive mapping is growing
cognitive mapping.]
clearer – only the features have yet to be fully filled in.
Kitchin, R. and Freundschuh, S. (2000) Cognitive Mapping: Past,
Present and Future, Routledge, London. [An edited research
monograph reviewing research progress thirty years after the
References heyday of behavioural geography.]
Liben, L.S. (2009) The road to understanding maps. Current
Blaut, J.M., McCleary, G.F. and Blaut, A.S. (1970) Environ- Directions in Psychological Science, 18, 310–315. [Reviews
mental mapping in young children. Environment and different psychological approaches to map understanding,
Behavior, 2, 335–349. with a rich emphasis upon contextual differences and their
Boulding, K. (1956) The Image, University of Michigan Press, impact on environmental cognition.]
Ann Arbor, MI.
Carr, S. (1970) The city of the mind, in Environmental
Psychology: Man and His Physical Setting (eds H.M.
Proshansky, W.H. Ittelson and L.G. Rivlin), Holt Rinehart See also
and Winston, New York, pp. 518–533. . Chapter 1.3: On Maps and Mapping
Downs, R.M. (1970) The cognitive structure of an urban . Chapter 1.6: Cartographic Communication
shopping centre. Environment and Behavior, 2, 13–39. . Chapter 1.11: Exploratory Cartographic Visualisation: Advanc-
Fishbein, M. (1967) Attitude and the prediction of behaviour, ing the Agenda
in Readings in Attitude Theory and Measurement (ed. M. . Chapter 3.3: Cartography as a Visual Technique
Fishbein), John Wiley & Sons, Inc., New York, pp. 477–492. . Chapter 3.6: The Roles of Maps
Lynch, K. (1960) The Image of the City, MIT Press, Cambridge, . Chapter 4.9: Understanding and Learning Maps
MA. . Chapter 4.11: Usability Evaluation of Web Mapping Sites