ISSN: 0004-5608 (Print) 1467-8306 (Online) Journal homepage: https://www.tandfonline.

com/loi/raag20

The Nature of Geographic Knowledge

Reginald G. Golledge

To cite this article: Reginald G. Golledge (2002) The Nature of Geographic Knowledge, , 92:1,
1-14, DOI: 10.1111/1467-8306.00276

To link to this article: https://doi.org/10.1111/1467-8306.00276

Published online: 15 Mar 2010.

Submit your article to this journal

Article views: 802

View related articles

Citing articles: 36 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=raag21
PRESIDENTIAL ADDRESS

The Nature of Geographic Knowledge

Reginald G. Golledge
Department of Geography & Director of the Research Unit on Spatial Cognition and Choice
University of California, Santa Barbara

The nature of geographic knowledge today is very different from what it was fifty years ago. It has evolved from phe-
nomenal (declarative) to intellectual (primed by cognitive demands). Surges of interest in systematic specialties and
technical innovations in representation and analysis have changed the nature of geographic knowledge, advanced
geographic vocabulary, defined and examined geographic concepts, and developed spatially explicit theories relat-
ing to human and physical environments. Explorations of interactions between these domains has generated a new
interest in integrated science. This interest has produced a unique way of examining human-environment relations,
and has provided the basis for a vastly different underlying knowledge structure in the discipline. But the future still
challenges and significant problems face geography if it is to remain a viable academic discipline in the new infor-
mation technology society. Key Words: geographic knowledge, spatial thinking and reasoning, incidental and intentional
learning, spatial relations, geographic skills.

Everything is related to everything else, but near things are used to summarize data that has been constructed from
more related than distant things. information sensed by human or technical means and
—Tobler (1970, 234) then analyzed and interpreted to unpack embedded (and
often obscured) spatial existence and relational charac-
The Changing Nature of teristics.
Geographic Knowledge In the latter part of the 20th century there has been a
substantial change in the nature of geographic knowl-

G
eographic knowledge is the product of geo- edge. Throughout most of the history of the discipline,
graphic thinking and reasoning about the geographic knowledge has been declarative—i.e., it has
world’s natural and human phenomena. Eliot focused on collecting and representing the physical and
(2000, 2) suggests that “knowledge of space is phenome- human facts of existence. In the latter part of this cen-
nal, knowledge about spaces is intellectual.” In geogra- tury there has been a change from inventory dominated
phy, knowledge of space represents the accumulation of activity to the creation of knowledge generated by em-
facts about the spatial arrangement and interactions phasizing cognitive demands, such as understanding
comprising human-environment relations and recogni- “why” and “how” in addition to “what” and “where.”
tion of fundamental concepts—i.e., the declarative base This has required a change from an emphasis on form to
of geographic knowledge. Knowledge about space con- an emphasis on process. The accumulation of geographic
sists of the recognition and elaboration of the relations knowledge has consequently changed from item recogni-
among geographic primitives and the advanced concepts tion, place labeling, and place inventory or gazeteering
derived from these primitives (such as arrangement, or- to feature and distribution matching in real or image set-
ganization, distribution, pattern, shape, hierarchy, dis- tings, item manipulation, and item transformation (e.g.,
tance, direction, orientation, regionalization, categoriza- using logical reasoning, deductive and inductive infer-
tion, reference frame, geographic association, and so on) ence, analysis of complex forms, and multi-modal repre-
and their formal linking into theories and generaliza- sentation). This has facilitated solution of tasks such as
tions. Using Eliot’s terms, intellectual or created knowl- recognizing geographic associations, understanding spa-
edge extrapolates far beyond simple sensory or observa- tial colinearity in either positive or negative directions,
tional information. In geography this extension is undertaking map overlaying, understanding the results
captured in part in the various forms of representation of scale transformation and rotation of separate displays
Annals of the Association of American Geographers, 92(1), 2002, pp. 1–14
© 2002 by Association of American Geographers
Published by Blackwell Publishing, 350 Main Street, Malden, MA 02148, and 108 Cowley Road, Oxford, OX4 1JF, UK.
2 Golledge

to achieve maximal coincidence, and so on. A new way of this equation, “the environment” began to take on
of thinking and reasoning in turn required the develop- many new facets. Initially, the term was expanded to in-
ment of new data (e.g., primary data on attitudes, percep- clude the “built” environment (i.e., the tangible addi-
tions, beliefs, values, emotions, preferences, all collected tions that humanity had made to the physical world).
directly from individuals), new methods of representa- Then, in the latter part of the century, the term “envi-
tional formats and modalities (e.g., computer cartogra- ronment” was expanded to include the behavioral en-
phy and graphics, tactile and auditory maps), new modes vironment (the environment of human interactions and
of analysis or interpretation (e.g., spatial analysis, mul- movements), the social-cultural environment (i.e., the
tiple analysis of variance, metric and non-metric multi- hidden structures of customs, beliefs, and values that
dimensional scaling, spatially based cluster analysis, gap constrained human relations), the political environment
analysis, spatial autocorrelation, measures of geographic (i.e., the human-defined boundaries, legal structures, and
association), new sources and types of information (e.g., organizational structures within which human action
rankings, ratings, scaled values, sketches, model build- takes place), and the cognitive environment (i.e., the in-
ing, simulation, long term chemical analysis, tree ring ternal representation of the world in our memories). Dur-
analysis, hydrologic stations, recordings from ocean ing the last half-century, geographic knowledge of each of
buoys, satellite images), and the recognition that geogra- these environments has advanced remarkably. In differ-
phers must consider multiple environments (natural, ent decades, one or another of the “ism”-perspectives and
built, interactional, socio-cultural, and cognitive) the most relevant related environment have become
(Golledge and Stimson, 1997, 7–22). dominant and have contributed to our understanding of
Perhaps the single greatest changes in the nature of the complexities of human-environment relations.
Geographic Knowledge over the last half-century have Each of the themes that have temporarily dominated
been: geographic thinking and reasoning has defined its own
criteria, elaborated its own methods, experimented with
1. Recognition of the difference between the acts of
its own data types and representational modes, devel-
accumulating geographic facts and representing
oped its own theories (or adapted acceptable ones from
the spatial form embedded in these facts, and un-
other disciplines), and has differentially chosen qualita-
derstanding the processes involved in understand-
tive or quantitative criteria for evaluating the signifi-
ing and analyzing those facts to produce new infor-
cance of efforts to produce and accumulate geographic
mation and knowledge that is not directly observed
knowledge.
during data gathering; and
The expansion of the nature of geographic knowledge
2. The development of spatially relevant theories
has occurred for a number of purposes. These include:
about the location, arrangement, and distribution
of geographic phenomena and the spatial interac- 1. increasing our understanding of place-to-place re-
tions among both physical and human components lations and variations;
of those phenomena. 2. obtaining a more complete base for interpreting
In different decades of the 20th century, different pur- human-environment relations at scales ranging
poses have dominated thought about what was consid- from personal to global spaces;
ered important in a geographic knowledge accumula- 3. assisting us to think about the spatial arrangement
tion. These have included regionalism, behavioralism, or organization of features, interactions, and rela-
Marxism, neo-Marxism, structuralism, postmodernism, tionships; and
critical theory, feminism, environmentalism, and infor- 4. facilitating the performance of efficient and effec-
mation science. A significant part of the quest for geo- tive spatial behaviors.
graphic knowledge has been detoured by attempts to Geographers generally are aware that, to pursue these
understand the latest “ism” rather than advancing geo- goals, an integrated approach gives the maximum under-
graphic knowledge—i.e., geographers have focused on standing. While regional geography—an approach that
perspective rather than substance and in doing so have epitomized integrated thinking—dominated much of
wasted much effort in internecine conflict and criticism. geography up to the early 1960s, during the next thirty
Understanding Human-Environment Relations (HER) years, systematic specialties grew at the expense of the
has been a constant theme throughout the history of ge- integrated regional approach. There is no doubt that sys-
ography. For much of that history “environment” largely tematic specialization was a necessary and inevitable step
referred to physical space (i.e., the tangible natural in the development of geographic knowledge. It was dur-
world). With an increasing emphasis on the human side ing this period that the full impact of the “Theoretical
 The Nature of Geographic Knowledge 3

Revolution” (often loosely called “The Quantitative and more evident that the HER integrative approach
Revolution”) was made evident. General knowledge was that has a natural home in geography has resurfaced as an
replaced with detailed knowledge. As specific human- important knowledge seeking procedure.
environment relations were closely examined, increased
awareness of these relations facilitated the development
of axiomatic, law-like, theoretical and generalized state- Geography as a Unique Way
ments, and laid out a formal knowledge base that could of Thinking and Reasoning about
justify policy recommendations. Normative theories the World and Its Inhabitants
(e.g., location theory, central place theory, urban popula-
tion density gradient theory, and so on) first emerged be- In the mid-1960s, a psychologist at Clark University,
cause they simplified the world by assuming away much after working with an inspired group of geographic re-
of the geographic variability continued within it (Har- searchers including Jim Blaut, Roger Hart, and others in-
vey, 1969). Then less rational and optimal “theories” (or terested in spatial knowledge (such as psychologist David
partial theories) emerged (e.g., Migration, Mobility, and Stea and architect Gary Moore), began asking questions
Spatial Interaction Theories) (e.g., Amedeo and as to why geographers “think differently” from other
Golledge 1975); then social theory (e.g., social justice, academics. Although not fully answering this question,
political economy and Marxist theory applied to Ameri- Beck suggested that spatial thinking and spatial imaging
can urban life) (Harvey 1973) attracted supporters who and representation (i.e., the way geographers organized
were disenchanted because no immediate ethical and so- their thoughts and presented data to others) was unique
cially responsible solutions were offered to counteract so- to the discipline (Beck 1967). Finding an answer to ques-
cial ills by existing approaches. Researchers then drifted tions of how and why geographers think the way they do
on to the next “ism”—post modernism (Scott and Soja has stimulated much of my research over the last 35
1996), then to critical thinking (often regarded as the es- years, and I have constantly wondered why other geogra-
cape from Marxism) (Gregory 1994), and, more recently, phers have neglected the active (not incidental) pursuit
to spatial information science (SIS) (Longley, Good- of this question. In this section, I offer some of the rea-
child, Maguire, and Rhind 1999) and environmentalism sons that I have entertained at various times that might
(nature/society) (Turner, Kates, and Meyer 1994). All help understand the uniqueness of geographic thought.
emphasized restricted points of view, and all but the Beck argued that development of spatial meaning in
SIS and nature/society developments drifted away an environmental context requires an interpretation of
from an integrated human-environment relations theme. both the physical and social components of the world as
Regionalism—still a paramount example of integrated it is observed or perceived. He argued that neither geog-
HER—maintained a presence throughout the years, and, raphy (which at that time he saw as being concerned
allied with a new cultural geography, has resurfaced as an with tangible physical properties of the earth-system)
important component of the discipline in recent years nor psychology (which he saw as being concerned with
(Hudson 1994). In the past decade, this revival of inte- personal attributes and functional and symbolic transac-
grated approaches in geography has been stimulated in tions between humans and their physical environment)
part by the emergence of global communities and global was equipped alone to deal with the development of spa-
domains that required integrations of knowledge about tial meaning. And, although he contended that finding
place, culture, interactions, politics, economics, resources, spatial meaning might require input from many disci-
and natural environmental characteristics. Also, it has plines, he (like another psychologist, Uttal, 30 years later
been the result of the efforts of Gilbert White and his [2000]) suggested that the way geographers reasoned
many students, associates, and followers whose work on about space, and particularly their penchant for repre-
the occurrence and impacts of a variety of natural and senting complex spatial meanings in a clearly under-
technological hazards required the systematic integra- standable form (spatially based maps, graphics, and
tion of knowledge of physical events, human attitudes, images) emphasized that geographic thinking and rea-
and the concepts of risk and uncertainty to present a soning gave a perspective that was not matched by any
powerful HER integrative approach that at this time ap- other single academic discipline.
pears to have culminated in “Sustainability” and “Vul- So what comprises Geographic Thinking and Reason-
nerability” studies (Cutter 1984a, 1984b, 1985, 1996; ing? By examining published geography literature over
Kates and Burton 1986; Turner, Kates, and Meyer 1994). the past 50 years I have compiled a partial list of thinking
As the world of academe once again turned its face to and reasoning processes that should help to answer this
ever more evident global problems, it has become more question. The list includes:
4 Golledge

Figure 1. Aligned global maps. Figure 1(A) shows the results of a longitudinal continental “alignment” process that results in people believing
that the South American west coast city of Santiago, Chile is west of the North American east coast city of Miami, Florida, while in geographic
fact, the reverse is true. Figure 1(B) illustrates the latitudinal “alignment” process in which the continents are perceptually moved North or
South producing misstatements such as “the equator passes through North Africa.” Source: Compiled by Susan Baumgart, University of Cali-
fornia, Santa Barbara.

• Comprehending scale transformations (Montello

1993; Montello and Golledge 1999; Clarke 2001;
Montello and Golledge, in preparation).
• Being able to transform perceptions, representa-
tions and images from one dimension to another
and the reverse (Monmonier 1996).
• Comprehending superordinate and subordinate re-
lations and frames of reference (cardinal, rela-
tional, local, global) (Longley et al. 1999).
• Comprehending problems of spatial alignment
(Figure 1A, 1B).
• Comprehending distance effects (e.g., distance de-
cay, Morrill 1963) (Figure 2).
• Comprehending spatial association (positive and
negative [Robinson and Bryson 1957; Cliff and Ord
1973; Hubert and Golledge 1982; Getis, 1989])
(Figure 3).
• Comprehending orientation and direction (e.g., Figure 2. Typical empirical distance decay curves. These constitute
forward-backward; left-right; up-down; back-front; examples of empirically derived distance decay functions for dis-
tances apart of marriage partners (exponential), urban population,
horizontal-vertical; north/south/east/west) (Nyerges density (lognormal), and urban land values (Pareto). Morrill (1963)
et al. 1995). pointed out that many different mathematical functions could be fit
• Comprehending spatial classification (regionaliza- to the same empirical data and that selecting a specific function
tion) (King 1969). should be justified by reference to relevant theory.
 The Nature of Geographic Knowledge 5

Figure 3. Three maps of spatial associa-

tion. These figures (top to bottom) show
a spatially mixed moderate positive cor-
relation (r  0.67), a spatially obvious
moderate positive association (r  0.67),
and a spatially negative moderate cor-
relation (r  0.67). Source: adapted
from Hubert et al. (1985, 45–46, figures
4, 5, and 6). Reprinted by permission.
Copyright © 1985 by The Ohio State
University Press. All Rights Reserved.

• Comprehending clustering and dispersion (central- lation density gradients in different cultural set-
izing and dispersing tendencies) (Gould 1975, 1993). tings) (Berry, Simmons, and Tenant 1963; Casetti
• Comprehending spatial change and spatial spread 1973).
(spatial diffusion) (Hägerstrand 1968; Brown 1981). • Comprehending spatial shapes and patterns (geom-
• Comprehending non-spatial and spatial hierarchy etry and topology) (Sack 1972).
(Golledge, Rayner, and Rivizzigno 1982) (Figure 4). • Comprehending locations and places (Haggett
• Comprehending densities and density decay (popu- 1965; Tuan 1989).
6 Golledge

Figure 4. (A) Non-spatial hierarchy. This shows an inclusionary hierarchy that is semantically encoded. (B) Spatial hierarchy. This spatial
hierarchy is based on interpoint distances (proximities).

• Comprehending overlay and dissolve (spatial ag- tures; relating traverses or cross-sections to three-
gregation and disaggregation) (Albert and Golledge dimensional block diagrams or images) (Marsh and
1999). Dozier 1983; Longley et al. 1999).
• Comprehending integration of geographic features
represented as points, networks, and regions In short, geographic thinking and reasoning has pro-
(Golledge 1978) (Figure 5). vided a basis for understanding—or reasoning out—why
• Comprehending spatial closure (interpolation) there are spatial effects, not just finding what they are!
(Boyle and Robinson 1979; Muehrcke and Muehr- Further, it enables us to reveal patterns in spatial distri-
cke 1992). butions and spatial behaviors that may not be obvious to
• Comprehending proximity and adjacency (nearest a casual observer in the real world (e.g., the pattern of
neighbor) and their effects (distance decay) (King shopping centers in a city) and consequently helps us un-
1969; Getis and Boots 1978; Boots and Getis 1988; derstand the reasons for occurrences of episodic behav-
Golledge 1995). iors (e.g., obligatory and discretionary activities) in
• Recognizing spatial forms (such as city spatial struc- terms of spatial processes.
 The Nature of Geographic Knowledge 7

5. Geographers use place-based reasoning in their sci-

entific endeavors, because, whether identified in
absolute (e.g., global co-ordinate) or relative (prox-
imal or local relational) ways, the place-specific na-
ture of all things in real, imagined, or virtual worlds
is paramount. Today’s geographers—more so than
ever before—relish this factor and make a search
for understanding place-specificity their guiding
objective.
6. In a digital world where scale transformation
(zooming) comes at the click of a mouse button,
understanding the implications of scale change
(scale-dependent and scale-independent processes)
for theory, practice, and policy has reached critical
levels. Geographic reasoning has made people
Figure 5. Anchor point theory. This figure shows a spatial knowl- aware of scale, and has graphically illustrated how
edge structure consisting of best-known points (landmarks), paths
changing scale can subtly (and not so subtly)
(transport routes), and areas (communities and neighborhoods).
change the world as we represent and interpret it
(Monmonier 1996; Montello and Golledge in
In a paper presented at the Pittsburgh meetings of the preparation). New scale-dependent relationships are
Association of American Geographers, Goodchild and brought to view by spatially explicit reasoning.
Anselin (2000) outlined some unique contributions And finding scale-independent relations is the first
made to qualitative and quantitative science by the step to developing robust geographic theory.
thinking and reasoning processes of geography practi-
tioners. Loosely interpreted and modified somewhat,
these include: Spatial Concepts and Relations:
The Bases of Geographic Knowledge
1. The integrative nature of geographic science, link-
ing human and physical dimensions by seeking to The development of geographic knowledge has been
understand human-environment relations. hampered because the discipline has not developed a
2. The development over four decades of Spatial Anal- widely accepted vocabulary. We do have dictionaries
ysis—the set of spatially based analytical tools that that provide defined lists of terms (e.g., Johnson et al.
explicitly focus on comprehending the spatial 2000). But little attention has been focused on defining
component of geo-referenced data. primitives, combining them to derive more complex
3. The use and comprehension of spatial representa- spatial/geographic concepts, and exploring what “error”
tions—whether prepared in paper and ink and gratuitously accumulates as we build primitives and low
hardcopy format (e.g., cartographic maps), as on- order concepts into more highly ordered (complex)
screen visualizations (e.g., images and graphics), or terms. Exceptions include the works of Nystuen (1963),
in digitized remotely sensed imagery (e.g., satellite Papageourgiou (1969), Golledge (1995), and Nyerges
data and air photos), these unique ways of repre- (1995). Place-specific identity, location, magnitude,
senting data have changed the way geographers time, boundary, and distance have been suggested as
reason and infer the existence of patterns, distribu- primitives upon which geographic knowledge is built.
tions, and relationships in spatial data. Examples of “first order” derivable concepts include: dis-
4. Geography now has spatially explicit theory to com- tribution or arrangement (from multiple locations), re-
plement the economic, social, political, cultural, gions (from aggregations of place-specific identifiable
biotic, meteorological, hydrological, geomorpho- phenomena), frames of reference (providing structures
logical, and other symbiotic theories that it “bor- that allow absolute or relative locations to be identified),
rowed” from adjacent disciplines during the first orientation and directions (from location, identity, time
sixty years of the 20th Century. Whether in human and reference frames), spatial hierarchies and dominance
science or physical science domains, spatially ex- (from magnitude and location). Higher order (complex)
plicit theory has added much to general under- concepts include such things as pattern, clustering and
standing of the world around us. dispersion (from internal arrangement of distributions),
8 Golledge

spatial association (from location, magnitude and distri- Table 1. Tangible Spatial Concepts and Geographic
bution) and density and distance decay (from boundary, Accommodations
distance, magnitude and distribution). The structure of ⇒ Geographic units are almost always irregular in shape and area.
the language of geography is ill-defined, under-researched, Our cognitive behavioral tendency is to make irregular shapes
and has been only casually taught and learned. An im- and areas more regular (smoothing and generalizing).
portant step to mitigate our ignorance of the structure of ⇒ Geographic units invariably do not maintain a common
our disciplinary vocabulary has been undertaken by the uniform orientation relevant to the cardinal points of the
joint efforts of the AAG, NCGE, and NGS (Geography compass. We may mentally rotate them to “fit” a reference
Standards—Geography for Life Curriculum, 1994). This frame.
curriculum defines important geographic concepts and ⇒ Hierarchical ordering is common to both physical systems
and to human organization (e.g., stream networks and
builds a K–12 curriculum around those concepts. Yet they geopolitical entities). Geographic units are cognitively and
are presented as independent units—their derivative empirically organized into a nested hierarchical form (e.g.,
nature and use in developing more complex concepts is school districts).
under-emphasized. ⇒ Our knowledge about locations, places, regions, and other
Some basic premises embedded in the spatial concepts geographic units is not perfect. Even with imperfect geographic
and relations that are the essence of geographic knowl- knowledge, effective geographic decision-making can take
edge and the accommodations made by cognitive pro- place, partly because we realize that geographic phenomena
occur in proximal spatially distributed forms (Tobler’s Law).
cessing are summarized in Table 1. This offers examples
⇒ Geographic phenomena may be irregularly distributed over
of findings from geographic observation over time as well space. There are underlying cognitive behavioral forces
as statements of how geographers accommodate tangible working to facilitate the meaningful clustering and
observations. categorization of geospatial information.
Partly as a result of cognitive filtering and processing, ⇒ Objectively defined geographic data is theoretically designed to
and partly because of inevitable technical errors in data have less bias than subjective data. Perceptual and cognitive
capture and representation, biases occur in geographic sensory filters and behavioral restraints invariably produce
knowledge. These biases include: biased personal knowledge.
⇒ Information on geographic phenomena can cover widely
• Conceptual biases based on improper thinking different scales. Geographic thinking uses plausible reasoning
and reasoning. processes to operate on imperfect and incomplete beliefs about
geography from local to global scales.
• Perceptually based biases resulting from inade-
⇒ Geographic data compiled by and for machine use has to be
quate use of gestalt principles for grouping, sym- more accurate and complete than that compiled by and for
metry, figure-ground clarity, closure, and so on. human use. Accuracy of geographic data depends on levels of
• Biases occurring during the processes of encoding, human familiarity with the nature and source of the
internal manipulation, or decoding geographic information.
information. ⇒ Geographic knowledge can be spatially fragmented.
• Biases that occur when geographic features are Generalizations are made from knowledge-rich to knowledge-
poor domains.
cognitively misaligned with respect to their actual
⇒ The nature of geographic knowledge is often made evident by
positions in objective physical space, as repre-
the way it is represented. Bias in geographic judgments is often
sented in conventional reference frames such as evident in the representation of the results of those judgments
latitude and longitude. (Harley 2001).
• Biases that occur when geographic information ⇒ The processes underlying geographic knowledge are different
summarized in regional form is simplified to a sub- from the processes underlying other complex knowledge
stantial degree, and inclusion/exclusion errors domains. Geographers are still ignorant of many of the
abound. processes (spatial and non-spatial) underlying geographic
knowledge.
• Biases resulting when non-symmetric geographic
⇒ Not all geographic knowledge is spatial in nature (e.g., some is
information is made symmetric, while symmetric
hierarchical and inclusionary in a semantic rather than in a
relations are perceived as non-symmetric. Percep- spatial sense). Non-spatial factors influence memory for
tual and cognitive factors combine to allow spatial locations (e.g., where function is more important than
shorter distances to be overestimated, and longer place in identifying phenomena). Biases often result when non-
distances to be underestimated in comparison to spatial knowledge frames are used to reason about spatial
physical distances (regression toward the mean). knowledge.
Perceived distances to and from a particular place ⇒ As we progress from local to global scale, geographic knowledge
are often regarded as being asymmetric. They may becomes more categorical than spatial.
 The Nature of Geographic Knowledge 9

further be distorted, depending on the direction of

viewing.
• Biases can arise because of improper renderings of
superordinate geographic structures.
• Biases in general geographic knowledge often re-
sult from a combination of an alignment heuristic
and a rotation heuristic used at the time of encod-
ing information. The alignment effect simply tends
to align features relative to one another, regardless
of true global setting, and the rotation heuristic
includes the tendency for a feature to be rotated to
fit a reference frame (e.g., “You are here” maps are
often incorrectly oriented with respect to objec-
tive reality and thus require the difficult process of
mental rotation to match with the real world).
• Biases occur in subjective estimates of locational
precision and constancy (e.g., people don’t always
perceive the same object to be in the same place at
different points in time); this justifies a need for
accurate objective records and representations
(e.g., maps).
• Biases when perspective changes influence evalu-
ation of spatial relations (an object to the left of
another becomes to the front or behind it follow-
ing perspective change).
• Biases that result when internal representations
are distorted or fragmented and produce error rid-
den spatial products (Figure 6), but note that a
geographically correct “map” is not necessarily
stored in long-term memory or created in working Figure 6. Three cognitive maps of newcomers, mid-term residents,
memory and is not needed to successfully solve and long-term residents. Examples of grids recovered using non-
geographic tasks. metric multidimensional scaling of cognitive interpoint distances
• Biases that occur because geographic language for long-term residents (top), 3-year residents (middle), and new-
lacks the metric information needed to build a comers (bottom).
correct spatial configuration: for example, when
we say, “A is behind B and C is to the left of B” we
can generate an infinite number of spatial config- Casually observing environments without a repertoire
urations that comply with the logical rules embed- of spatial concepts, theories, and generalizations, pro-
ded in the statement. duces this informal or incidental “knowledge.” Inadequa-
cies in this knowledge base result from (1) spatial biases
in observing and internally representing information;
Informally and Formally Acquired (2) improper manipulation of stored information re-
Geographic Knowledge called into working memory for spatial decision-making;
(3) lack of spatial skills that help comprehend perceived
Geographic knowledge consists of “informally ac- information; (4) insensitivity to sample size (generaliz-
quired,” “incidental,” or “naïve” knowledge (Egenhofer ing from n  1); (5) misconceptions of the nature of
and Mark 1995) and “formally acquired” or “intentional” chance events (e.g., “we’ve just had a 20-year storm and
(“taught,” “learned,” or “expert”) knowledge. Informally won’t have another for 19 years”); (6) illusions of valid-
acquired knowledge dominates in most of our everyday ity (e.g., “my experience is typical”); and (7) personal
decision-making and thought processes. And a good deal aggrandizement (“I wasn’t included in the sample so it’s
of that knowledge is acquired using general guidelines not representative!”). Unfortunately for many people
that produce vague or error-prone knowledge. (particularly in countries like the USA where geography
10 Golledge

is not an essential and integrated part of general educa- and knowledge base that is not casually accessible or eas-
tion in all K–16 environments), informal or incidentally ily (naively) accumulated. Rather, it is a concept rich
acquired geographic knowledge is the main source for and structured body of knowledge that is based on spe-
understanding the world. Fortunately, during the last two cific modes of thinking and reasoning that usually have
decades, major assaults on geographic ignorance have to be taught. This is especially true of geographic knowl-
been launched by cooperative action among the Associ- edge today.
ation of American Geographers (AAG), the National To more completely understand the nature of geo-
Council for Geographic Education (NCGE), the Na- graphic knowledge, we must be aware of differences in
tional Geographic Society (NGS), the American Geo- the levels of spatial abilities among people. But, even
graphic Society (AGS), and the Committee on Geogra- more fundamentally, we must be aware of the very nature
phy (COG). Important results include the Geography of the spatial abilities and skills that accommodate the
Standards: Geography for Life Curriculum (1994) and acquisition of geographic knowledge. An ongoing Na-
the ARGUS and ARGWORLD projects (Gersmehl tional Research Council initiative on “Spatial Think-
1999). A National Research Council initiative brought ing” (headed by geographer Roger Downs) is focusing on
forth a book summarizing the mission and content of this precise problem.
geography today (NRC 1997). Other significant efforts
have included projects on “Human Dimensions of The Usefulness of Geographic Knowledge
Global Change” (Hanson 1997), Mission Geography
(Bednarz and Butler 1999), and the institution of an Ad- Geographic knowledge is useful for two fundamental
vanced Placement (AP) test for high school geographers reasons: (1) to establish where things are and (2) to re-
(Boehm and Petersen 1999). Each of these emphasizes member where things are to help us in the process of mak-
geographic concepts and geographic relations as well as ing decisions and solving problems. Establishing where
offering insights into distributions, arrangements, pat- things are has produced the need for exact forms of loca-
terns, and interactions. And all reflect the changing na- tional determination and place and feature representa-
ture of geographic knowledge. tion as epitomized in cartographic developments. Re-
There is a remarkable difference between the quality membering where things are is part of everyday life and
and accuracy of the informal or incidental geographic everyday decision-making. Inadequacies in remembering
knowledge that we acquire by personally experiencing force us either to make mistakes or to consult representa-
places during activities dominated by other purposes tions of where things are before we can effectively use
(e.g., “experiencing” an urban environment during a knowledge of geographic patterns and relations.
work or shopping trip) and the deliberately structured But establishing where things are is but one way that
formal or intentional geographic knowledge that we ac- geographic knowledge is found useful. Other aspects con-
quire via teaching and learning processes (e.g., by using sist of making us aware of the spatial relations amongst
the Geography for Life Curriculum, 1994). As geogra- things (e.g., soils and vegetation), the regional or categor-
phers, we are amazed at the appalling geographic igno- ical classes to which things belong (e.g., groupings of
rance of those persons whose knowledge repertoire is urban functions; culture regions), the extent to which
dominated by informal or incidentally acquired informa- things interact (e.g., urban land values and population
tion. Such people often cannot name the major conti- densities), and the extent to which things are co-related
nents, may not be able to identify the USA on a map or in terms of their spatial occurrence and distribution (e.g.,
globe (Earhardt 1998), or may have completely spurious professional sports teams and large cities). In other words,
understanding of place (e.g., confusing the Mediterra- it helps us to know why things are where they are, and how
nean with the Caribbean). and why they are spatially related to other things.
Geographic knowledge levels change dramatically Acquiring geographic knowledge helps to develop a
when formal or intentional knowledge is gained—partic- capacity for recognizing occurrences of similar phenom-
ularly when people are taught to observe fundamental ena in different environments (knowledge transfer) and
geographic principles like location, place, connectivity, recognizing one, two, or three-dimensional transforma-
interaction, distribution, pattern, hierarchy, distance, di- tions of phenomena (this knowledge includes a capacity
rection, orientation, reference frame, geographic associa- for transforming three-dimensional objects into two-
tion, scale, region, and geographic representation (many dimensional representations as when making a map or
of these have been incorporated into the Geography creating a geologic profile from a block diagram). Other
Standards-Geography for Life Curriculum). This implies transformations include rotation and alignment (e.g., of
that geography—like other disciplines—has a language maps and the scenes they represent or resulting from per-
 The Nature of Geographic Knowledge 11

spective changes). In movement, spatial knowledge is 1. Will there be a Geography Discipline in, say, 50
necessary to undertake route reversals, take shortcuts, years? That is, will emerging trends that de-
and navigate through unfamiliar territory. It also helps to emphasize disciplinary boundaries, reorganize col-
create integrated images from separate independent bits legiate structures, and facilitate the develop-
of evidence (e.g., compiling an integrated understanding ment of research communities in universities,
of an environment from piecemeal explorations of it). In eliminate the need for departments, particularly
many circumstances, geographic knowledge is required those—like geography—that appear to lend them-
to complete problem-solving tasks in imagined as well as selves to diffusion of sub-areas into other aca-
real space (e.g., spatial simulations), understanding spa- demic units?
tial correlations among geographically dispersed phe- 2. What will distinguish Geography’s contribution
nomena, and imagining integrative representational modes to knowledge from that of other disciplines with
for communicating spatial information about phenom- Spatial Information or Integrated Science inter-
ena in visual, cartographic (maps), auditory (auditory ests? What aspects of Geographic Knowledge will
maps), kinesthetic (mental records of effort), and haptic help us compete equally for students and research
(tactile map) domains. funds with other information processing disci-
Geographic knowledge is useful not only for effectively plines and agencies?
participating in everyday life, such as helping to remem- 3. How will Geographic Knowledge increase if the
ber the location of ATM’s, shopping centers, and routes world we live in becomes more reliant on digital
to local schools, but at all other scales of living. Federal information technology? What will be the re-
and state policies are implemented at regional and local search and instructional role of technical innova-
levels, and the process of regionalization is in essence, tions such as virtual worlds and wearable wireless
spatial classification. A significant geographic problem computers?
that emerges after every population census is to redistrict 4. What aspects of geographic thinking and reason-
electoral regions. This is necessitated by spatial variation ing help us to create an accepted identity as the
in natural population growth and migration (the process pre-eminent link between human and natural
of redistributing population over space). Geographic sciences? Geographers for decades have claimed
knowledge is useful in every political decision ranging to best represent that link, but in fact for over 30
from the determination of national boundaries to the al- years much of both human and physical parts of
location of funds to maintain local transportation sys- the discipline have virtually ignored each other.
tems. There is little point in attempting to list all the What is needed to reintegrate the discipline? Is it
areas in which geographic knowledge is pertinent. It is necessary to do this?
used universally, in all cultures, in all regions, but it is im- 5. What geographically specific knowledge can be
portant to understand why it is used universally. Which accessed to help solve some of the world’s “Great
brings me to the crux of this section. The natural world Challenges” such as effects of Climatic Change,
and the human activities embedded in it are incredibly the task of creating or maintaining Sustainable
diverse. The human brain does not deal well with ex- Natural and Urban Environments, or deciding
treme diversity (chaos), but can handle variability. We do how many people the Earth can support and
this in a spatial sense by searching for locational regu- where they will be located?
larities that can be cognitively categorized and spatially 6. What unique output or consequences occur due
associated so that information can be ordered and com- to spatial and geographic thinking? The role of
municated. In this way, variability and diversity can be geographic knowledge in the history of the
interpreted, and what might otherwise be considered growth and spread of human populations has
chaos can be comprehended at some scale of organiza- not been articulated, so we do not at this time
tion. Geographic knowledge thus helps us make sense out know why geographic knowledge has been im-
of chaotic or apparently highly diversified environments. portant in the emergence and growth of human
civilization.
7. What will be the ideal format for geographic rep-
Problems Involved in Pursuing Geographic resentations in the future? Will they be electronic
Knowledge in the 21st Century or hard copy? Single or multiple media?
8. How can Geographic Knowledge contribute to
Some questions and tasks that the discipline must face the comprehension and solution of problems in-
as we move further into the 21st century include: volved in society-space relations? Can geogra-
12 Golledge

phers help solve problems of inequality, inequity, unique modes of representation, and practical usefulness
and social justice as well as just identifying and for decision-making and policy formulation.
cataloging their existence? Geographic Knowledge is universal. But it can not be
9. What future role can Geographic Knowledge acquired only informally or incidentally by casual obser-
play in establishing global international, na- vation. We must define and accept a comprehensive set
tional, regional, and local policy? of concepts on which knowledge can be based. Without
10. What geographic knowledge can we create to en- such a base, our knowledge structure is speculative and
hance understanding of global societies, cultures, hard to justify or defend. With such a base, we can erect
economies, and political and information struc- a formidable and rich array of concepts, generalizations,
tures? Will global mapping provide the ultimate laws and theories which are equivalent to acceptable
knowledge structure to achieve such goals? bodies of knowledge in other disciplines. It is important
that we realize that we need to deal with well-defined
concepts as opposed to vague ones. Understanding this
Summary will in part provide a basis for assessing the validity, reli-
ability, and justifiability of our arguments and conclusions—
Johnson (1997, 35) stated that: “To most of us, there is much debated questions by critical thinkers throughout
no such thing as geography, other than as a vaguely de- the discipline. Geographic knowledge represents a body
fined discipline to which we are attached as much for po- of science that has much to offer humanity, and, as its
litical and economic (that is, job security) reasons as for professors, we must do our best to continue adding to all
intellectual ones.” He further states “And does it matter? facets of that knowledge set. Geography is a healthy disci-
I believe not. There is no such thing as geography, only a pline, and maintaining it as such in the future does matter.
lot of separate geographies all of which share character-
istics with the others, but are quite considerably self-
sufficient.” While such an impression may be gained be- References
cause of the specialization and fragmentation of geography
Albert, W. S., and R. G. Golledge. 1999. The use of spatial cog-
over the last 40 years, I cannot agree with this statement
nitive abilities in geographical information systems: The
because (1) it ignores the fact that topical specialization map overlay operation. Transactions in GIS 3 (1): 7–21.
has been a necessary step in the development of detailed Amedeo, D., and R. G. Golledge. 1975. An introduction to scien-
geographic knowledge by providing the deep under- tific reasoning in geography. New York: John Wiley and Sons.
standing of knowledge segments prior to their integra- Beck, R. (1967). Spatial meaning and the properties of the envi-
ronment. In Environmental Perception and Behavior. Re-
tion into relevant theories and models; (2) it fails to ac- search Paper no. 109, ed. D. Lowenthal, 18–29. Chicago:
knowledge the existence of a common set of primitives Department of Geography, University of Chicago.
and concepts that span both physical and human sides of Bednarz, S. W., and D. R. Butler. 1999. “Mission geography” and
the discipline; and (3) such a position is defeatist and the use of satellite imagery in K–12 geographic education:
only provides fuel for others to continue their assault on A NASA-GENIP partnership. Geocarto International 14
(4): 85–90.
the discipline in terms of departmental closures and ex- Berry, B. J. L., J. W. Simmons, and R. J. Tennant. 1963. Urban
clusion of geographers from multidisciplinary challenges population densities: Structure and change. The Geographi-
to which they have much to offer. Rather, I believe that cal Review 53:389–405.
geographic knowledge is concept-rich, has a substantial Boehm, R. G., and J. F. Petersen, eds. 1997. The first assessment:
Research in geographic education. San Marcos, TX: The Gil-
theoretical base, is replete with distinct analytic forms
bert M. Grosvenor Center for Geographic Education.
(both qualitative and quantitative), lends itself to graphic, Boots, B. N., and A. Getis. 1988. Point pattern analysis. Beverly
cartographic, and other forms of geo-representation, and Hills: Sage Publications.
is imbued with knowledge derived from using place- Boyle, M. J., and M. E. Robinson. 1979. Cognitive mapping and
specific reasoning to integrate components of its physical understanding. In Geography in the urban environment, vol.
2, ed. D. Herbert and R. Johnson, 59–82. New York: John
and human domains. Wiley and Sons.
The changes that have taken place over the last 50 Brown, L. A. 1981. Innovation diffusion: A new perspective. New
years have given structure and substance to a discipline York: Methuen.
that was concerned largely with inventory. Today we Casetti, E. 1973. Urban land value functions: Equilibrium versus
have all the components deemed necessary to define and optimality. Economic Geography 49 (4): 357–65.
Clarke, K. C. 2001. Getting started with geographic information sys-
justify the existence of a scientific discipline—a huge tems. 3rd ed. Upper Saddle River, NJ: Prentice Hall.
array of empirically verified factual data, spatial theories Cliff, A. D., and J. K. Ord. 1973. Spatial autocorrelation. London:
and models, innovative methods of spatial analysis, Pion.
 The Nature of Geographic Knowledge 13

Cutter, S. L. 1984a. Emergency preparedness and planning for Gregory, D. 1994. Geographical imaginations. Cambridge, MA:
nuclear power plant accidents. Applied Geography 4:235– Blackwell Publishers.
45. Hägerstrand, T. 1968. A Monte Carlo approach to diffusion. In
———. 1984b. Risk cognition and the public: The case of Spatial analysis: A reader in statistical geography, ed. B. J. L.
Three Mile Island. Environmental Management 8:15–20. Berry and D. F. Marble, 368–84. Englewood Cliffs, NJ:
———. 1985. Rating places: A geographer’s view on quality of life. Prentice-Hall.
Washington, DC: Resource Publications in Geography, As- Haggett, P. 1965. Locational analysis in human geography. New
sociation of American Geographers. York: St. Martin’s Press.
———. 1996. Societal responses to environmental hazards. In- Hanson, S. 1997. Human dimensions of global change. Washing-
ternational Social Science Journal 48 (4): 525–36. ton, DC: Association of American Geographers.
Earhardt, D. K. 1998. The geography pretest: An indicator of Harley, J. B. 2001. The new nature of maps : Essays in the history of
knowledge. Paper presented at the Annual Meeting of the cartography. Baltimore: Johns Hopkins University Press.
Southwest Division of the Association of American Geog- Harvey, D. 1969. Explanation in geography. London: Edward Ar-
raphers, 30 October, Baton Rouge, LA. nold; New York: St. Martin’s Press.
Egenhofer, M. J., and D. M. Mark. 1995. Naive geography. Tech- ———. 1973. Social justice and the city. Baltimore: Johns Hop-
nical Report 95–8: National Center for Geographic Infor- kins University Press.
mation and Analysis (NCGIA), Orono, ME, University of Hubert, L. J., and R. G. Golledge. 1982. Measuring association
Maine, and Buffalo, NY, State University of New York between spatially defined variables: Tjøsteheim’s Index and
(SUNY). some extensions. Geographical Analysis 14 (3): 273–78.
Eliot, J. 2000. The nature and measurement of spatial intelligence. Hubert, L. J., R. G. Golledge, C. M. Costanzo, and N. Gale.
College Park, MD: Institute for Child Study, College of Ed- 1985. Measuring association between spatially defined vari-
ucation, University of Maryland. ables: An alternative procedure. Geographical Analysis
Gersmehl, P. 1999. Activities and readings in the geography of the 17:36–46.
United States (ARGUS) Minneapolis: University of Minne- Hudson, J. 1994. Making the Corn Belt: A geographical history of
sota. Midwestern agriculture. Bloomington: Indiana University
Getis, A. 1989. A spatial association model approach to identi- Press.
fication of spatial dependence. Geographical Analysis 21 (3): Johnston, R. J. 1997. Australian geography seen through a glass
251–59. darkly. Australian Geographer 28:29–37.
Getis, A., and B. Boots. 1978. Models of spatial processes. Cam- Johnston, R. J., D. Gregory, G. Pratt, and M. Watts, eds. 2000.
bridge, U.K.: Cambridge University Press. The dictionary of human geography. 4th ed. Oxford:
Golledge, R. G. 1978. Learning about urban environments. In Blackwell.
Timing space and spacing time, vol. 1, Making sense of time, ed. Kates, R. W., and I. Burton. 1986. Geography, resources, and en-
T. Carlstein, D. Parkes, and N. Thrift, 76–98. London: Ed- vironment: Themes in the work of G. F. White. Vols. 1 and 2.
ward Arnold. Chicago: University of Chicago Press.
———. 1992. Do people understand spatial concepts: The case King, L. J. 1969. Statistical analysis in geography. Englewood
of first-order primitives. In Theories and methods of spa- Cliffs, NJ: Prentice-Hall.
tiotemporal reasoning in geographic space: International Con- Longley, P. A., M. F. Goodchild, D. J. Maguire, and D. W. Rhind,
ference GIS—From space to territory: Theories and methods of eds. 1999. Geographical information systems: Principles, tech-
spatiotemporal reasoning. Pisa, Italy, September 21–23, Pro- niques, applications, and management. New York: Wiley.
ceedings, ed. A. U. Frank, I. Campari, and U. Formentini, Mark, D., and S. Freundshuh. 1995. Spatial concepts and cog-
1–21. New York: Springer-Verlag. nitive models for geographic information. In Cognitive as-
———. 1995. Primitives of spatial knowledge. In Cognitive as- pects of human-computer interaction for geographic informa-
pects of human-computer interaction for Geographic Informa- tion systems, ed. T. L. Nyerges, D. M. Mark, R. Laurini, and
tion Systems, ed. T. L. Nyerges, D. M. Mark, R. Laurini, and M. J. Egenhofer, 21–28. Dordrecht: Kluwer Academic
M. J. Egenhofer, 29–44. Dordrecht: Kluwer Academic Pub- Publishers.
lishers. Marsh, W. M., and J. Dozier. 1983. Landscape planning: Environ-
Golledge, R. G., J. N. Rayner, and V. L. Rivizzigno. 1982. Com- mental applications. New York: John Wiley and Sons.
paring objective and cognitive representations of environ- Monmonier, M. 1996. How to lie with maps. 2nd ed. (first ed.
mental cues. In Proximity and preference: Problems in the 1991). Chicago: University of Chicago Press.
multidimensional analysis of large data sets, ed. R. G. Golledge Montello, D. R. 1993. Scale and multiple psychologies of space.
and J. N. Rayner, 233–66. Minneapolis: University of Min- In Spatial information theory: A theoretical basis for GIS. Lec-
nesota Press. ture Notes in Computer Science 716. Proceedings, European
Golledge, R. G., and R. J. Stimson. 1997. Spatial behavior: A geo- Conference, COSIT ’93. Marciana Marina, Elba Island, It-
graphic perspective. New York: Guilford Press. aly, September, ed. A. U. Frank and I. Campari, 312–21.
Goodchild, M. F., and L. Anselin. 2000. Spatially integrated so- New York: Springer-Verlag.
cial science: Building the research infrastructure. Paper Montello, D. R., and R. G. Golledge. 1999. Scale and detail in
presented in honor of Reginald Golledge at the Associa- the cognition of geographic information. Report of Spe-
tion of American Geographers, Pittsburgh, PA. cialist Meeting of Project Varenius, 14–16 May 1998, Uni-
Gould, P. 1975. Spatial diffusion: The spread of ideas and innova- versity of California at Santa Barbara, Santa Barbara, CA.
tions in geographic space. New York: Learning Resources in http://www.ncgia.org (last accessed 16 November 2001).
International Studies. Montello, D. R., and R. G. Golledge. In preparation. Scale and
———. 1993. The slow plague: A geography of the AIDS pan- the cognition of geographic information. University of Califor-
demic. Cambridge, MA: Blackwell Publishers. nia, Santa Barbara.
14 Golledge

Morrill, R. L. 1963. The distribution of migration distances. Papers Papageorgiou, G. J. 1969. Description of a basis necessary to the
and Proceedings of The Regional Science Association 11:75–84. analysis of spatial systems. Geographical Analysis 1 (2): 213–15.
Muehrcke, P. C., and J. O. Muehrcke. 1992. Map use: Reading, anal- Robinson, A. H., and Bryson, R. 1957. A method for describing
ysis, and interpretation. 3rd ed. Madison, WI: JP Publications. quantitatively the correspondence of geographical distribu-
National Geographic Research and Exploration. 1994. Geogra- tions. Annals of the Association of American Geographers
phy for Life: National Geography Standards—1994. Washing- 47:379–91.
ton, DC: National Geographic Research and Exploration. Sack, R. D. 1972. Geography, geometry, and explanation. Annals
National Research Council (NRC). 1997. Rediscovering geogra- of the Association of American Geographers 62 (1): 61–78.
phy: New relevance for science and society. Washington, DC: Scott, A. J., and E. W. Soja, eds. 1996. The city : Los Angeles and
National Academy Press. urban theory at the end of the twentieth century. Berkeley:
Nyerges, T. 1995. Cognitive issues in the evolution of GIS user University of California Press.
knowledge. In Cognitive aspects of human-computer interac- Tobler, W. 1970. A computer movie simulating urban growth in
tion for geographic information systems, ed. T. L. Nyerges, the Detroit regions. Economic Geography 46 (2): 234–40.
D. M. Mark, R. Laurini, and M. J. Egenhofer, 61–74. Dor- Tuan, Y.-f. 1989. Morality and imagination: Paradoxes of progress.
drecht: Kluwer Academic Publishers. Madison: University of Wisconsin Press.
Nyerges, T. L., D. M. Mark, R. Laurini, and M. J. Egenhofer, eds. Turner, B. L., II, R. W. Kates, and W. B. Meyer. 1994. The earth
1995. Cognitive aspects of human-computer interaction for geo- as transformed by human action in retrospect. Annals of the
graphic information systems. Dordrecht: Kluwer Academic Association of American Geographers 84 (4): 711–15.
Publishers. Uttal, D. H. 2000. Seeing the big picture: Map use and the de-
Nystuen, J. D. 1963. Identification of some fundamental spatial velopment of spatial cognition. Developmental Science
concepts. Papers of the Michigan Academy of Science, Arts, 3:247–86.
Letters XLVIII:373–84.

Correspondence: Department of Geography, University of California, Santa Barbara, CA 93106-4060, e-mail: golledge@geog.ucsb.edu.