Spatial Cognition

of the utility of continuing the trip from that des-tination. Dellaert et al. (1998) generalized Kitamura'sapproach to account for multipurpose aspects of thetrip chain.

Trip chaining is only one aspect of multidayactivity /travel patterns. Several models have recentlybeen suggested to predict more comprehensive activitypatterns. These models can be differentiated into theolder constraints-based models (e.g., Lenntorp 1976),utility-maximizing models (e.g., Recker et al. 1986),rule-based or computational process models (e.g.,Golledge et al. 1994, ALBATROSS -Arentze andTimmennans 2000).

6. Conclusions

This article has given a very brief overview of spatialchoice modeling. Due to the limited available space,we have focused on some of the key modelingapproaches, disregarding alternatives (e.g., dynamicmodels) altogether.

If we examine the accomplishments, one cannotescape the conclusion that the models developed overthe years have been increasing in complexity. Thetheoretical underpinnings of the recent activity-basedmodels are much richer than the Markov or gravitymodels of the 1960s. Moreover, the brief overviewindicates that geographers have been using alternativedata, such as experimental design data and activitydiaries, in addition to the more conventional surveydata.

See also: Spatial Analysis in Geography; SpatialDecision Support Systems; Spatial Interaction; SpatialInteraction Models; Spatial Optimization Models;Spatial Pattern, Analysis of; Spatial StatisticalMethods; Spatial Thinking in the Social Sciences,History of

Kitamura R 1984 Incorporating trip chaining into analysis ofdestination choice. Transportation Research 188: 67-81

Lenntorp B 1976 Paths in space-time environments. LundStudies in Geography, Series B, no. 44

Louviere J J 1984 Hierarchic~1 information integration: a newmethod for the design and "nalysis of complex multiattributejudgment problems. In: J(innear T C (ed.) Advances inConsumer Research 11, ACRA, Provo, UT, pp. 148-55

Louviere J J, Woodworth G 1983 Design and analysis ofsimulated consumer choice or allocation experiments: Anapproach based on aggregate data. Journal of MarketingResearch 20: 35~7

Oppewal H, Louviere J J, Timmermans H J P 1994 Modelinghierarchical information int~gration processes with integratedconjoint choice experiments. Journal of Marketing Research31: 92-105

Recker W W, McNally M G, Root G S 1986 A model ofcomplex travel behavior: part I -theoretical development.Transportation Research, A '20A: 307-318

Rushton G 1969 Analysis of spatial behavior by revealed spacepreference. Annals of the As~ociation of American Geographers59: 391-400

Rushton G 1974 Decompositil>n of space preference structures.In: Golledge R G, Rushton G(eds.) Spatial Choice and SpatialBehavior. Ohio State University Press, Columbus, OH, pp.119-33

Timmermans H J P 1979 A spatial preference model of regionalshopping behaviour. Tijdschrift voor Economische en SocialeGeografie 70: 45-8

Timmermans H J P 1984 Decl>mpositional multiattribute pre-ference models in spatial choice analysis: a review of somerecent developments. Progress in Human Geography 8: 189-221

Timmermans H J P, Golledge R G 1990 Applications of behav-ioural research on spatial choice problems II: preference andchoice. Progress in Human Geography 14: 311-54

Williams H C W L 1977 On the formation of travel demandmodels and economic evalqation measures of user benefit.Environment and Planning, A 9: 285-34

Wilson A G 1971 A family of spatial interaction models andassociated developments. Environment and Planning, A 3: 1-32

H. Timmennans

Spatial Cognition

Spatial cognition concerns the study of knowledge andbeliefs about spatial properties of objects and events inthe world. Cognition is about knowledge: its ac-quisition, storage and retrieval, manipulation, and useby humans, nonhuman animals, and intelligentmachines. Broadly construed, cognitive systems in-clude sensation and perception, thinking, imagery,memory, learning, language, reasoning, and problem-solving. In humans, cognitive structures and processesare part of the mind, which emerges from a brain andnervous system inside of a ,body that exists in a socialand physical world. Spatial properties include lo-cation, size, distance, direction, separation and con-nection, shape, pattern, and movement.

BibliographyArentze T A, Timmermans H J P 2000 ALBATROSS: a

Learning-based Transportation-oriented Simulation System.ElRASS, Eindhoven University of Technology, The Nether-lands

Dellaert B, Arentze T A, Borgers A W J, Timmermans H J P,Bierlaire M 1998 A conjoint based multi-purpose multi-stopmodel of consumer shopping centre choice. Journal ofMarketing Research 35: 177-88

EttemaD F, TimmermansH J P 1997 A ctivity-based Approachesto Travel Analysis. Elsevier Science, Oxford, UK

Fotheringham A S 1983 A new set of spatial interaction models:The theory of competing destinations. Environment andPlanning, A 15: 1121-32

Girt J L 1976 Some extensions to Rushton's spatial preferencescaling model. Geographical Analysis 8: 137-56

Golledge R G, Kwan M P, Garling T 1994 Computationalprocess modeling of travel decisions using a geographicalinformation system. Papers in Regional Science 73: 99-117

14771

Spatial Cognition

and decision-making involved in spatial behavior e.g.,migration, vacationing, and trip scheduling. Thisapproach gained insight from the work of KevinLynch (1960), a planner who argued that 'images' ofcities guide people's behavior and experiences of thosecities. All of these historical developments constituteda new focus on disaggregate or microscale approacheswithin human geography, approaches that focused onthe geographic behavior of individuals as anaccompaniment to tradi~ional approaches that fo-cused on human settlements and larger aggregationsof people.

1. The Geographic Study of Spatial Cog"ition

Geographers study spatial cognition for three reasons.First and foremost, spatial cognition provides -ques-tions of geographic interest in their own right, giventhat cognition about space and place is an expressionof human-environment or human-earth relations.Second, geographers have hoped that spatial cognitionwould improve their understanding of traditionalphenomena of interest; for example, where peoplechoose to shop should depend in part on their beliefsabout distances and road connections. Finally, therehas been an interest in improving the use and design ofmaps and other geographic information productswhich depend in part on human understanding ofdepicted spatial relations.

In addition to geographers and cartographers,researchers from many different disciplines studyspatial cognition, including psychologists, architectsand planners, anthropologists, linguists, biologists,philosophers, and computer scientists. While over-lapping and interacting with these disciplines, thegeographer's interest in spatial cognition is distinct insome ways. Geographers are interested in the earth asthe home of humanity. Their concern is with thecognition of humans, not the cognition of barnswallows or underseas robots. Geographers focus onspatial scales relevant to human activities on theearth's surface, such as temporary travel, migration,settlement, and economic activities; a geographerwants to understand human conceptualizations of thelayout of a city or a region, not of an atom or the solarsystem. Finally, the geographer ultimately desires todescribe, predict, and explain external human-environment relationships, and focuses on internalmental and neuropsychological structures and pro-cesses only insofar as it helps in the understanding ofexternal relationships.

2. Topics in the Geographic Study of Spatial

Cognition

The study of spatial cognition in geography sheds lighton questions such as hlj>w spatial knowledge andbeliefs are acquired and develop over time; the natureof spatial knowledge structures and processes; howpeople navigate and stay oriented in space; how peopleuse language to communicate with each other aboutspace; and how aspects of spatial knowledge andreasoning are similar or different among individuals orgroups.

1.1 History of Spatial Cognition in Geography

The notion that a subjective understanding of spacepartially mediates human relationships with an ob-jective physical world is undoubtedly quite old. Withinthe modern discipline of geography, the study ofspatial cognition first appeared with work on geo-graphic education early in the twentieth century.However, interest in spatial cognition flowered duringthe 1950s and 1960s. The subdiscipline of experimentalcartography emerged with the insight that maps couldbe improved by understanding how humans perceiveand think about information depicted in map symbols.In an attempt to explain human responses to floodsand other natural hazards, the subdiscipline of en-vironmental perception began to study people's sub-jective beliefs about the occurrence and consequencesof hazard events. Most importantly for the study ofspatial cognition, behavioral geographers began devel-oping theories and models of the human reasoning

14772

2.1 Acquisition and Deveiopment

Humans acquire spatial knowledge and beliefs directlyvia sensorimotor systems that operate as they moveabout the world. People also acquire spatial knowl-edge indirectly via static al)d dynamic symbolic mediasuch as maps and images, 3-D models, and language.Geographers are interested in how different mediahave consequences for the nature of acquired knowl-edge.

Spatial knowledge ch~nges over time, throughprocesses of learning and development. Both changesin the child's spatial knowledge and reasoning, andthat of an adult visiting a new place for the first time,are of interest to geographers. Both physical mat-uration and experience influence development. A per-son's activity space; the set of paths, places, andregions traveled on a regular basis, is an importantexample of spatial experience that influences people'sknowledge of space and place, no matter what theirage. Most people know the areas around their homesor work places most thoroughly, for example (as in theanchor-point theory of Cohclelis et al. 1987).

A widely accepted model, of spatial learning suggeststhat spatial knowledge of places develops in a sequenceof three stages or elements. First is landmark knowl-edge, unique features or views that identify a place.Second is route knowledge, based on travel routinesthat connect ordered sequences of landmarks. Thefinal element is survey knowledge, knowledge of two-dimensional layouts that includes the simultaneous

Spatial Cognition

interrelations of locations; survey knowledge wouldsupport detouring and shortcutting. Not everyone hasdetailed or comprehensive survey knowledge of theirsurrounds, and exposure to maps of places clearlyadvances the development of such knowledge.

tures, and interpreting verbal route directions. Way-finding tasks generally do involve a cognitive map ofthe environment. Various locomotion and wayfindingtasks vary greatly in tl1eir demand on attentionalcapacity. Following a familiar route is not verydemanding on attention; finding one's way in a newcity may be quite demanding.

Orientation is knowing 'where you are,' althoughthe precision and compr~hensiveness of this knowingvaries greatly in different situations and for differentpeople. Two fundamental processes are involved inorientation during navigation. The first process is therecognition of external, features or landmarks-pilotage. In some cases, th~ recognized landmark is thegoal, but more commonly the landmark aids orien-tation by serving as a key to knowledge of spatialrelations stored in an internal cognitive or externalcartographic map. The second process is dead reck-oning, updating a sense of orientation by integratinginformation about move~ent speed, direction, and/oracceleration without reference to recognized features.

People also use symbolic media such as maps tonavigate and stay oriented. Maps used for navigationshow robust alignment effects, a common confusion inusing maps when the tqp of the map is not thedirection in the surrounds that one is facing whenlooking at the map. It results from the fact that spatialperception is 'orientation-dependent'; objects andpictures have a top and a bottom in the perceptualfield. In the case of road or trail maps, many peopleturn the map to bring about alignment. Unfortunately,this option is not available with ubiquitous 'You-Are-Here' maps, which are surprisingly often placed in amisaligned way with respect to the viewer's perspec-tive. The resulting alignment effect is reflected by thefact that misaligned maps !ire interpreted more slowlyand with greater error by most people (Levine et al.1984). While an annoyanCe in many situations, mis-aligned emergency maps placed in public buildings arepotentially disastrous.

2.2 Spatial Knowledge Structures and Processes

The metaphor of a cognitive map was coined byTolman in a 1948 paper to refer to internallyrepresented spatial models of the environment. Thecognitive (or mental) map includes knowledge oflandmarks, route connections, and distance and dire-ction relations; nonspatial attributes and emotionalassociations are stored as well. However, in manyways, the cognitive map is not like a cartographic 'mapin the head.' It is not a unitary integrated represent-ation, but consists of stored discrete piecesincluding landmarks, route segments, and regions.The separate pieces are partially linked or associatedfrequently so as to represent hierarchies such as thelocation of a place inside of a larger region. Furt-hermore, spatial knowledge is not well modeled bymetric geometries such as the Euclidean geometry ofhigh school math. The ways in which spatial knowl-edge is internally represented in human minds lead tocertain patterns of distortions in the way peopleanswer spatial questions. For example, people oftenbelieve the distance from place A to B isdifferent than from B to A. Turns are frequentlyadjusted in memory to be more like straight lines orright angles. At larger scales, most people havesomewhat distorted ideas about the sizes and locationsof continental land masses on the earth; for example,South America is thought to be due south of NorthAmerica when it is actually rather far to the east aswell.

2.3 Navigation and Orientation

Humans travel over the earth's surface to reachdestinations. This typically requires planning and theability to stay oriented while moving. Navigation isthis coordinated and goal-directed travel throughspace. It consists of two components, locomotion andwayfinding. Locomotion refers to the guidance ofoneself through space in response to local sensori-motor information in the immediate surrounds, andincludes such tasks as identifying surfaces of support,avoiding obstacles, and moving toward visible land-marks. Locomotion generally occurs without the needfor an internal model or cognitive map of the en-vironment. Wayfinding refers to the planning anddecision-making that allows one to reach a destinationthat is not in the immediate sensory field, and includessuch tasks as choosing efficient routes, schedulingdestination sequences, orientating to nonlocal fea-

2.4 Spatial Communication via Language

Spatial infonnation is oft<ln communicated verbally,in natural language such as English. People give andreceive verbal route directions, read spatial descrip-tions contained in stories; and increasingly interactwith computer systems via verbal queries. There aretwo notable characterist~s of the way languageexpresses spatial infonnation. One is that languageexpresses mostly nonquantitative or imprecise('fuzzy') quantitative infortnation about space. State-ments about connections and approximate locationare more important thaQ precise statements. Forexample, people say 'turn l<lft at the gas station' ratherthan 'turn 800 after you go 0.6 miles.' Great precisionis typically unnecessary or' even confusing. A secondcharacteristic is that various aspects of the context of

14773

Spatial Cognition

communication are critical in interpreting spatiallanguage. Context is provided by knowledge of who isspeaking, where they are, physical features in thesituation, the previous topic of conversation, and soon. 'The bicycle is near the church' is generallyunderstood differently than 'Wisconsin is nearMichigan.'

information about routes and destinations. Severalresearch issues concern how locational information isbest displayed or communicated, and how the avail-ability of such information might change people'sexperiences and behavio~ in space and place.

Other computer technologies will lead to newresearch and thinking in spatial cognition. There are ahost of spatial cognition questions inspired by effortsto improve the effectivehess and efficiency of geo-graphic information syst~ms (GISs). How can com-plex geographical information be depicted to pro-mote comprehension and effective decision-making,whether through maps, graphs, verbal descriptions, oranimations? How does exposure to new geographicinformation technologies \lIter human ways of perceiv-ing and thinking about the world? There are alsoresearch issues concerning the way that people do ordo not spatialize their understanding of other types ofinformation networks sudh as the World Wide Web;one often speaks metaphrrically of 'navigating' theWeb. However, undoubtedly the most dramatic tech-nological development in $patial cognition will be theadvent of computer-simulated worlds known as virtualenvironments or virtual rdality.

2.5 Individual and Group Similarities andDifferencesNo two individuals know exactly the same things orreason in exactly the same way about space and place.Some people are better at tasks such as wayfinding,learning spatial layouts, or reading maps. In somecases, there may be different ways to think about suchproblems, all of which may be effective. In such cases,one might speak of 'stylistic differences' in spatialcognition rather than 'ability differences.' Geograph-ers are interested in measuring and explainingsuch individual differences. Traditional psychometricpaper-and-pencil tests of spatial abilities tend not tomeasure these differences very well (Allen et al. 1996).

There are many factors that may contribute tovariations in spatial cognition: body size, age, edu-cation, expertise, sex, social status, language, culture,and more. A first goal for geographic research is tomeasure and document ways these factors mightcovary with spatial cognition. Females and malesappear to differ in some ways in their spatial abilitiesand styles, for example, but not in others (Montello etal. 1999). In addition to describing such covariations,geographers have the research goal of distinguishingtheir underlying causes. For example, the two sexeslikely differ in their spatial cognition in part because ofdifferent travel experiences afforded them by theirsocialized roles. Explanations for such covariationsare generally quite difficult to determine, however, asone cannot readily do randomized experiments onperson characteristics such as age, sex, culture, andactivity preferences.

See a/so: Cartography; Cognitive Maps; GeographicEducation; Geographic Learning in Children; Knowl-edge (Explicit and Impliqit): Philosophical Aspects;Space: Linguistic Expression; Spatial Memory Loss ofNormal Aging: Animal Models and Neural Mech-anisms; Spatial Thinking in the Social Sciences,History of

3. Technologies and the Future of SpatialCognition Research

A variety of technologies are poised to have a largeimpact on the questions and methods of spatialcognition research in geography. Global positioningsystem (GPS) receivers have recently become smalland inexpensive, and are increasingly used by sailors,hikers, and other outdoor enthusiasts. A GPS receiverprovides information about one's location by geo-metrically combining distance signals it picks up fromseveral geosynchronous satellites constantly orbitingthe earth. This system is also part of the technology ofautomated navigation systems built into automobilesto provide drivers up-to-date and locally relevant

14774

BibliographyAllen G L, Kirasic K C, Dobson S H, Long R G, Beck S 1996

Predicting environmental le~ning from spatial abilities: Anindirect route. Intelligence 22: 327-55

Couclelis H, Golledge R G, G!\le N, Tobler W 1987 Exploringthe anchor-point hypothesis of spatial cognition. Journal ofEnvironmental Psychology 7: f)9-122

Downs R M, Stea D (eds.) 1973 Image and Environment.Aldine, Chicago

Golledge R G 1987 Environmental cognition. In: Stokols D,Altman I (eds.) Handbook of Environmental Psychology,Wiley, New York

Levine M, Marchon I, Hanley G 1984 The placement andmisplacement of you-are-here maps. Environment & Behavior16: 139-57

Liben L S, Patterson A H, Nqwcombe N 1981 Spatial Rep-resentation and Behavior ACross the Life Span. AcademicPress, New York

Lloyd R 1997 Spatial Cognition: Geographic Environments.Kluwer Academic, Dordrecht, The Netherlands

Lynch K 1960 The Image of the City. MIT, Cambridge, MAMark D M, Frank AU (eds.) 1991 Cognitive and Linguistic

Aspects of Geographic Space. Kluwer Academic, Dordrecht,The Netherlands

Montello DR, Lovelace K L, Golledge R G, Self C M 1999Sex-related differences and ~imilarities in geographic and

Spatial Data

Table 1Terminology for describi~ geographic features

environmental spatial abilities. Annals of the Association ofAmerican Geographers 89(3): 515-34

Tolman E C 1948 Cognitive maps in rats and men. Psy-chological Review 55: 189-208

Tversky B 1992 Distortions in cognitive maps. Geoforum 23:131-8

Digital database term(and synonyms)Feature type Example

Point feature Benchmark Node (vertex, O-cell)Linear feature Road centerline Arc (edge, l-cell)Areal feature Lake/pond Polygon (face, 2-ce1l)

D. R. Montello

Spatial Data

Spatial data are the sum of our interpretations ofgeographic phenomena. Taken literally, spatial datacould refer to any piece of information that associatesan object with a location-from stars in the heavens totumors in the human body. Typically, however, whenused in a geographical context, spatial data is restrictedto describing phenomena on or near the earth'ssurface. In digital form, the data are the primaryinformation needed by geographic information sys-tems (GIS), the software tools used for spatial-dataanalysis.

ena of interest. For users to make sense of the datathey are receiving, they must understand the con-ceptual model underlying the data.

The locational aspect of a spatial data set is usuallydescribed by one of two ty~s of models, either grid-cell(sometimes called raster) data models, or vector datamodels. Grid-cell data models use collections of fixedunits of space, for example, cells of land SOm on a side,to describe the location jlnd extent of a geographicphenomenon. Digital remotely sensed images arecollected in this form. Vector data models use graphicelements (points, lines, and polygons) to describe thelocation and extent of the phenomenon under study.Most digital map data arJ in this form.

The spatial data also must include all of thenonlocational information used to describe the fea-tures in the database. In the case of digital-image data,the amount of nonlocational data is relatively small,usually a set of values measuring characteristics suchas the greenness of vegetation. With vector data theamount of nonlocational, or attribute, data can bequite large, even exceeding' the amount of spatial data.This includes information' about the various featuresincluded in the database (for example, the number oflanes of a highway) and their relationships to eachother (for example, that the railroad tracks lie betweenthe stream and the highway). In the context of GIS,features are the sum of our interpretations of geo-graphic objects or phen~mena. Buildings, bridges,roads, streams, grassland, and counties are examplesof features.

1. What Are Spatial Data?

'Spatial data' is used as an all-encompassing term thatincludes general-purpose maps, remotely sensed im-ages, and census-tract descriptions, as well as morespecialized data sets such as seismic profiles, dis-tribution of relics in an archeological site, or migrationstatistics. Information about buildings, bridges, roads,streams, grassland, and counties are other examples oftypes of spatial data. Broadly, spatial data refers toany piece of information that has an absolute orrelative location (see Location: Absolute/Relative).More commonly, spatial data are associated withinformation shown on maps and images of naturaland man-made features that range in size from 1 m to10 km. Such data are often referred to as geospatialdata (see Geographic Information Systems; SpatialData Infrastructure; Remote Sensing; Landsat Imageryin Geography; Remote Sensing and Geographic Infor-mation Systems Analysis; Global Positioning Systemfor examples of the varying types and usages of spatialdata). Spatial data sets are collections of locationaland nonlocational information about selected geo-graphic phenomena. This collection of informationforms the heart of a spatial database. Once theinformation is in the database, computers provide thecapabilities to handle geographic data in ways thatwere previously impossible.

2.1 Feature Data StructuresThe real world of geographical variation is infinitelycomplex and often uncertain, but must be representeddigitally in a discrete, deterministic manner. In onemethod of describing geogrilphic reality, characterizedby the information shown on maps, we think of theworld as a space populated by features of variouskinds-points, lines, and areas. This is commonlyreferred to as the vector or feature data model.Features have attributes that serve to distinguish themfrom each other. Any point'in the space may be empty,or occupied by one or more feature. These featuresmay exist in one combil"jed database, or may beseparated according to a theme or variable into anumber of layers or classes. The terminology used todescribe these features is given in Table 1.

2. Data Models

A digital spatial data set represents a certain model ofgeographic reality. These models describe both thelocational and nonlocational aspects of the phenom-

14775

Copyright @ 2001 Elsevier Science Ltd. All rights reserved.International Encyclopedia of the Social & Behavioral Sciences ISBN: 0-08-043076-7