269

A cross-cultural study of young children’smapping abilities

Mark Bladesa, J M Blautb, Zhra Darvizehc, SilviaElguead, Steve Sowdene, Dhiru Sonif, ChristopherSpencerg, David Steah, Roy Surajpauli and David Uttalj

The mapping abilities of four-year-old children in York, England, Durban, SouthAfrica, Tehran, Iran, Mexico City, Mexico and Evanston, Illinois, USA wereinvestigated, using a methodology involving air-photo identification and simulatednavigation on an air photo. The results show that essential mapping abilities(perspective and scale transformations) are well developed by the age of four inthese cultures, and provide some evidence in support of the hypothesis thatmapping abilities emerge without training in very young children of all cultures.

key words mapping abilities of preschoolers cognitive developmentcross-cultural environmental psychology

aDepartment of Psychology, University of Sheffield, Sheffield S10 2TN, UKbDepartment of Geography, University of Illinois at Chicago, 1007 West Harrison Street, Chicago,

IL 60607–7138, USA Corresponding authorcDepartment of Psychology, Al-Zahra University, Vanak, Tehran, IrandDepartment of Humanities, Universidad Autonoma Metropolitana, Mexico City, DF, MexicoeDepartment of Psychology, University of Sheffield, Sheffield S10 2TN, UKfDepartment of Geography, University of Durban-Westville, Durban 4000, South AfricagDepartment of Psychology, University of Sheffield, Sheffield S10 2TN, UKhDepartment of Geography and Planning, Southwest Texas State University, San Marcos, TX 78666, USAiDepartment of Geography, University of Durban-Westville, Durban 4000, South AfricajDepartment of Psychology, Northwestern University, Evanston, IL 60208–2710, USAe-mail: jblaut@uic.edu

revised manuscript received 21 December 1997

Introduction

We report here the results of a limited cross-cultural study of the mapping abilities of four-year-old children. The research was carried out inYork, England, Durban, South Africa, Tehran, Iran,Mexico City, Mexico and Evanston, Illinois, USA,and sought to determine whether childrencan interpret a vertical aerial photograph andcarry out a make-believe navigation task on thephotograph. If they can do so, they havedemonstrated an ability to engage in a simpleform of map-reading and simulated map-use;simple in the sense that it does not depend on anability to read text and interpret conventional

Trans Inst Br Geogr NS 23 269–277 1998ISSN 0020-2754 © Royal Geographical Society (with The Institute

signs, although it does depend on an ability toperform the basic mapping transformations ofperspective rotation and scale reduction. Ifthese children can indeed read and ‘use’ maps inthis sense, then we will have obtained goodevidence that mapping abilities of people inthese communities are well developed by the ageof four. We will also have obtained evidencethat has a strong bearing on several importanttheoretical issues, among them the question ofwhether the development of spatial (macro-environmental) cognition proceeds more rapidlythan is postulated in Piagetian theory, and thequestion of whether early development ofmapping abilities is a cross-cultural universal.

of British Geographers) 1998

Mark Blades et al.270

Before the 1960s, very little research was doneon the development of mapping abilities in youngchildren, and it was broadly assumed that childrencannot learn how to read and use maps before theage of about eight (Fairgrieve 1930; Shryock 1939)or later (Renner 1951). Piaget’s theory of the devel-opment of spatial cognition, which became widelyknown in the late 1950s (see especially Piaget andInhelder 1956), postulated that children youngerthan about seven are ‘preoperational’, hence can-not perform the cognitive ‘operations’ involved inmap-reading. This provided support for the pre-vailing view that map-learning is a late attainmentin children, and quickly gave the Piagetian pointof view the status of a hegemonic paradigm (seePrior 1959; Satterly 1964; Almy 1967; Eliot 1970;Rhys 1972; Graves 1975; Robinson and Petchenik1976; Gerber 1981; Naish 1982; Boardman 1983;Petchenik 1987; Winn 1987; Downs et al. 1988). Inthe late 1960s and early 1970s, a few studies carriedout by Piagetian geographers claimed to find thatyoung children do not possess mapping abilities(Miller 1967; Towler and Nelson 1968; Towler1970; 1971), but these studies were crudelydesigned and are unpersuasive. A more importantgroup of studies from the Piagetian perspectivewas carried out by Downs and Liben in the 1980s.They concluded from these studies that most chil-dren younger than about seven cannot perform theperspective and scale transformations to a map-like view of the world (Downs and Liben 1988;Liben and Downs 1989; Downs et al. 1988; Liben1991; Liben and Downs 1991), a conclusion consist-ent with Piaget’s postulate that children youngerthan this age cannot perform spatial operations,and do not possess spatial concepts, both of whichare needed for map comprehension. However,Blaut (1997a; 1997b) has criticized some aspects ofthe methodology used in these studies, and sug-gests that the data do not support the negativeconclusions that have been drawn from them. (SeeConning and Byrne 1984; Matthews 1992; Spenceret al. 1989; and Spencer and Darvizeh 1981 ontendencies to underestimate pre-school children’smacrospatial abilities.) Thus, there is at presentlittle empirical evidence against the view thatyoung children possess significant mappingabilities.

Research by a number of geographers andpsychologists has produced strong evidence thatyoung children do, indeed, possess significantmapping abilities. For example, Blaut, Stea and

their co-workers investigated the mapping abilitiesof young children in several studies in the late1960s and early 1970s, using air photos as mapsurrogates (Blaut et al. 1970; Blaut and Stea 1971;Stea and Blaut 1973). They found that the five- andsix-year-olds whom they studied in the UnitedStates, Puerto Rico and St Vincent could, in alimited sense, read vertical black-and-white airphotos with scales ranging from about 1:1000 to1:5000, and they found that this ability does notdepend on prior experience with an aerial view.Children identified landscape features and land-scape complexes (e.g. towns), and in one test theymade a tracing over the photo and then, after thephoto had been removed, drew a route on thetracing. Muir developed a successful first-grademap-skills curriculum based on air-photo interpre-tation (Muir and Blaut 1970). Spencer et al. (1980)found that children as young as three can identifyfeatures on a vertical black-and-white air photo;Walker (1980) and Matthews (1985a; 1985b)obtained comparable results with children agedfive and older.

Studies employing other methodologies havealso provided strong evidence of significant map-ping abilities in pre-schoolers. In a toy-play studywith three- to five-year-old children, about half ofthe threes and most of the fours and fives con-structed very simple landscapes in freeplay withtoys (Blaut and Stea 1974). Bluestein and Acredolo(1979) asked three- to five-year-old children to usea very simple map to find a hidden toy in a room;half the threes and most of the fours succeeded inthe task when it was aligned, map to room, andmost of the fives succeeded with an unalignedmap. Blades and Spencer (1986) showed, further,that fours can use a (simple) unaligned map afterinstruction. Rieser et al. (1982) showed that two-year-olds can use an aerial (overhead) view to plana route in a simple maze. Blades and Spencer(1986), DeLoache (1989) and others have shownthat three-year-olds can relate a scale model of aroom to the room itself in toy-finding tasks.Reviews of this literature are to be found inSpencer et al. (1989), Matthews (1992) and Bladesand Spencer (1994). Thus, there is substantial evi-dence that young children have mapping abilities.

The research reported in the present papersought to extend our knowledge of young chil-dren’s mapping abilities in several ways. Weemployed a research design that provided moredefinitive evidence than had been obtained in

Young children’s mapping abilities 271

previous studies concerning a child’s comprehen-sion of an aerial photograph as an iconic map. Weobtained this evidence for very young children:children between the ages of 48 and 55 months(4,0–4,11), and we tested children from five cul-tures, although from roughly comparable socialfractions in these cultures. The research may bringus closer to an adequate understanding of themanner in which mapping behaviour develops inhumans everywhere.

Conceptual foundations and purposes ofthe research

A very general hypothesis guided this research. Wethink it likely that mapping behaviour – the think-ing and action involved in reading, making andusing map-like models – is a cognitive and culturaluniversal, an ability that responds to universalecological needs and is acquired very early inchildhood by human beings of all cultures. Weallow for both strong and weak versions of thishypothesis. For example, it may be true that clearlyidentifiable maps (or map-like models) are madeand understood in all cultures; or it may be truethat people in all cultures have a basic mappingability, but some do not put it to much use. If thelatter is true, we should expect to find at least somerecognizable mapping everywhere – most con-spicuously in children’s representations of the geo-graphic landscape in toy-play – but specializeddevelopment of mapping only in some culturesand some ecological contexts. For another example,mapping behaviour may depend heavily on innateprecursors, or it may emerge from early learningand experience, with only minor contribution frominnate precursors (such as an awareness of verti-cality and horizontality). For still another example,mapping behaviour may or may not depend, asargued in Piagetian theory, on the prior acquisitionof spatial-cognitive operations, but Piagetianswould have to accept the evidence that such opera-tions or abilities appear in early childhood (not aslate as seven or thereabouts, as they suggest). Suchdiffering viewpoints are consistent with the generalhypothesis of universality. On the other hand, theold but still widely accepted belief that people insome non-Western cultures cannot deal with com-plex and abstract ideas about space (Werner 1961;Sack 1980; Deregowski 1989) is not consistent withthe hypothesis.

This general hypothesis presupposes somemeasure of agreement on the characteristics oftypical, ordinary, traditional maps. They depict(model) a portion of the Earth’s surface as it wouldbe seen from far overhead, thus a view that hasbeen rotated in perspective about 90 degrees fromthe horizontal and greatly reduced in scale. Suchmaps generally model areas too large to be seenas a whole from any one earthbound vantagepoint. They represent features and areas withsigns that may be iconic (pictorial), abstract(in the sense of reduced information) or purelysymbolic (having no resemblance to thatwhich they signify). They may be two- or three-dimensional models; most, of course, are two-dimensional, but we also use dioramas, raised-relief maps and the like. More fundamentally,three-dimensional scale models of landscapes,viewed ordinarily from overhead, hence map-like,are very widespread in culture. Indeed, when achild constructs a landscape on the floor or groundwith toys or other symbolic objects, that child ismaking a map-like model. In this paper, we willuse ‘map’ to mean any such map-like model, whilenoting that very different and more complex andsophisticated models, from ancient world mapsto modern location diagrams, are produced bycartographers.

A vertical aerial photograph is a highly iconicmap-like model. It is a rotated, scale-reduced rep-resentation of the landscape, hence it has the basicsyntactic properties of a map. It does not employarbitrary or linguistic signs, although it presents aslightly abstracted image in the sense that anyphotograph is an abstracted (or degraded) imageof the scene being photographed. If its scale is largeenough for a child to distinguish features withwhich they are familiar (when seen from theground-level view), it serves as a useful map-surrogate for testing the ability of pre-school chil-dren to perform the map-syntactic transformationsof rotation and scaling. If a not-yet-literatechild who has never seen the world from thisperspective, and has been given no instructionon how to read an air photo, can, indeed, read anair photo – identifying landscape features andareas on it – then the child can successfullyengage in the form of map-reading that involves arecognition of the Earth’s surface when seen fromthe map perspective. If, in addition, a child canplay navigation games on the air photo, correctlyshowing a route that one would follow (by car or

Methods2

Mark Blades et al.272

on foot) between two distant points on thephoto, then that child can read the air photo as awhole, as a landscape-gestalt, and can ‘use’ themap in make-believe way-finding. This reasoningled us to employ air photos as instruments forprobing the mapping behaviour of very youngchildren.

The purpose of the study reported here was todiscover whether a four-year-old child can read anair photo and perform a simulated navigationtask on it, hence engage in mapping behaviourat an elementary level. Prior research, citedabove, has given some evidence that pre-schoolchildren can do these things, but the experimentaldesigns used to test the ability have beensomewhat imprecise, and very little research hasbeen done with children younger than five yearsof age.

In the framework of our universality hypothesis,we sought to discover whether mapping ability iswidely cross-cultural; that is, whether it occurs infour-year-olds from a number of very differentcultures. The decision to work with children inYork, Durban, Tehran, Mexico City and Evanstonmainly reflected the fact that we live and work inthese areas. The five cases do not give us cross-cultural universality. However, the children stud-ied in all five cities come from communities thatcan be described (on local criteria) as middle classto upper lower class, in which there is a high levelof adult literacy: hence there is some similarityamong the communities in social terms. The factthat these cultures exhibit great diversity, com-bined with the fact that we studied children fromroughly the same social fraction in each culture,suggests that some degree of cross-cultural gener-alization is warranted.1 Future research will takeplace in other cultures and social fractions,but evidence of universality will still remainlimited and partial for various reasons, not leastof which is the fact that truly culture-neutralresearch designs have not been devised forcross-cultural psychological research with children(see Cole 1996). This limitation is of very greatrelevance for the present research, so much so thatwe must caution readers not to infer culturaldifferences from the (rather small) quantitativedifferences in performance by the five groupsof children. Our specific goal was to find outwhether mapping can be carried out by thefour-year-old children studied in each of the fivecultures.

SubjectsA total of 144 four-year-old children were studiedindividually in nursery schools in York, Durban,Tehran and Mexico City, and in a university psy-chology laboratory in Evanston. Approximatelyeven numbers of males and females were tested.Information on the socio-economic background ofchildren was not sought, but it was clear that theYork and Evanston children came from middle-class communities, the Durban and Tehran chil-dren came from lower middle-class communitiesand the Mexico City children came from a commu-nity that was slightly poorer, though not a slum. InDurban, nineteen children were of Indian descentand one was of Zulu descent. Testing was done inEnglish in York, Durban and Evanston, in Farsi inTehran, and in Spanish in Mexico City, by nativespeakers of each language.

MaterialsChildren were tested with vertical aerial photo-graphs. Different photographs were used in thedifferent areas. We had rejected the idea of testingall children with a single photograph, since thecharacteristics of the depicted landscape would bemore familiar to children of some cultures than tothose of others. Initially, our intention was to testwith photographs of the areas surrounding thenursery schools, at a common scale, but these wereunavailable in some areas and unobtainable withour limited resources in others; in the end we useda variety of images. In one sense, this is a limitationon the cross-cultural generality of the study, but theuse of images of an unfamiliar landscape shouldhave the effect of reducing, not enhancing, thechildren’s performance. York children and Tehranchildren were tested with a black-and-white photoof a portion of Sheffield, England, at a scale ofapproximately 1:1300.3 Durban children weretested with a black-and-white photo of the portionof the city of Durban in which their nurseryschool is located, at a scale of approximately1:4000. Mexico City children were tested with a1:1100 colour air photo of another (very dissimilar)part of Sheffield. Evanston children were testedwith a 1:720 black-and-white photo of theirneighbourhood. In the portion of the test thatinvolved make-believe navigation on the photo,children in York, Durban and Tehran used a felt-tipped marker, drawing on a transparent overlay

Young children’s mapping abilities 273

placed rigidly on the photo and retained as arecord of the child’s performance. Mexico Citychildren also drew on a transparent overlay; mostdrew their routes on it with a felt-tipped marker,but a few children, who had not yet acquiredcompetence with writing instruments, navigatedby moving a tiny toy car across the photo, theexperimenter carefully marking the routes used bythese children. In the Evanston study, childrennavigated with a tiny toy car or a tiny toy bird.Testing was recorded with camcorder and taperecorder.

ProceduresThe children were tested individually. A singlebasic methodology was used in the York, Durban,Tehran and Mexico City studies. The Evanstonstudy was the last to be carried out, and themethodology was varied slightly in the light of ourexperience in the other localities. The procedurethat was used in York, Durban, Tehran and MexicoCity can be described as follows.

The experimenters had visited the schools pre-viously and were well-known to the children.Before testing began, experimenters gave the chil-dren a brief practice session with the writingimplement to confirm their competence in its use.The same instructions (as phrased in the differentlanguages) were used for all children. The experi-menter first showed the air photo to the child, andasked, ‘What is this called?’. If the child repliedeither ‘a photograph’ or ‘a picture’, or used otherlanguage conveying comprehension, the experi-menter then asked, ‘What is this a photograph(picture) of?’, followed by, ‘Where do you think theperson was who made this photograph (picture)?’.Next, the experimenter asked, ‘What can you see inthe picture?’. If the child named a feature andpointed to it, the experimenter asked, ‘What elsecan you see?’. This continued until the child nolonger spontaneously mentioned and pointed tonew features on the photo. At this point, theexperimenter began to point to features on thephoto and asked, ‘Can you tell me what this is?’ (orwords to that effect). Next, the experimenter said,‘Let’s play a game. Let’s pretend that this is whereyou live’. The experimenter pointed to a predeter-mined starting point (a house) and then asked,‘What’s the name of your best friend?’. The childreplied with a name. The experimenter thenpointed to a predetermined finishing point

(another house) and said, ‘Let’s pretend that this iswhere [the person named] lives’. The same twohouses were used for all children in a given com-munity; the houses were on opposite sides of thephoto, and would not have been intervisible onthe ground. The experimenter next evaluated thechild’s understanding of the instructions by point-ing to the starting point and asking, ‘Now, in ourgame, who did we say lives here?’. If the childreplied correctly, the experimenter pointed to thefinishing point and asked, ‘And who did we saylives here?’. The experimenter then gave the pen ortoy to the child and said, ‘Suppose you wanted tovisit [the person named as best friend], can youshow me by drawing with your pen how youwould go from where you live?’. After the childhad performed the navigation task, correctly orincorrectly, the experimenter asked, ‘Can youtell me what you have just done?’. Very minorvariations in wording were allowed.

The verbal responses were scored as correct orincorrect, according to the words used by the child.If the child did not say that the image was a map ora photograph or picture taken from the air, thisresponse was scored as incorrect, unless the childconveyed the same basic meaning with differentwords. Identification of a feature was scored ascorrect if the child named the feature with anappropriate word (e.g. house, building) or used aword or words that would be a reasonable alter-native (e.g. a house being called ‘where peoplelive’). In this paper, we aggregate the free identifi-cations (‘What can you see?’) and prompted iden-tifications (‘Can you tell me what this is?’) into asingle score. The route navigated by the child wasscored as correct if the line drawn by the childfollowed roads and did not cross over rooftops(except where this clearly represented a slipof the child’s hand, as confirmed on the videotape). Given the instructions, many differentroutes were equally valid, so the child’s scoring onthe task did not depend on the route chosen, aslong as it would be an appropriate one in the reallandscape.

In the Evanston study, the navigation testemployed a somewhat more elaborate design.Children were randomly assigned to two testinggroups. Children in one group navigated with atiny toy bird; those in the other group, with a tinytoy car. After a child had finished the feature-identification task, the experimenter placedsmall sticky markers on two houses on the photo.

Mark Blades et al.274

Children in the ‘bird’ group were asked to showhow the bird would go from one house to theother. Children in the ‘car’ group were asked howthe car would go from one house to the other. Foreach group, there were four trials, with the desti-nation house placed successively farther from theorigin house. A child in the ‘bird’ group was scoredas having performed the task correctly if he or she‘flew’ the toy bird in a more or less straight linefrom one house to the other. A child in the ‘car’group was scored as correct if he or she ‘drove’ thecar realistically along streets. The difference inmethodology between the Evanston study and theothers was rather moderate, and we have com-bined the data from all five areas in the followingdiscussion of results.

navigation problem.

Results

Results are summarized in Table I. In the followingparagraphs, we report first the results obtained ateach study site, then the summary data for all

4

study sites.

YorkResponses to the first naming task (‘What is thiscalled?’) were mixed: six children used appropriatelanguage, ten used inappropriate language andfour gave ambiguous responses. None of the chil-dren correctly answered the second naming ques-tion (‘Where do you think the photo was takenfrom?’): three (of sixteen scored responses) gaveincorrect answers and thirteen gave ‘don’t know’responses. Nineteen of the twenty children cor-rectly identified at least one feature. The meannumber of different features identified was 5·2.Thirteen children made no erroneous identifi-

cations; three made one error and four made twoor three errors. Only two children made significantsyntax errors (i.e., identifications that were notconsistent with the map perspective or scale). Inthe navigation task, fourteen of the twenty children(70 per cent) drew an appropriate route on thephotograph.

DurbanResponses to the first naming question weremixed. Seven of the twenty children gave appro-priate answers (of whom four said ‘map’), ten said‘don’t know’ and three gave ambiguous answers.Responses to the second naming question werepoor: two children answered ‘aeroplane’ andeighteen answered ‘don’t know‘. On the identifica-tion task, seven different landscape features werecorrectly identified; all children made at least onecorrect identification; the mean number of correctidentifications was 6·6; the mean number of erro-neous identifications was 0·8; the mean number of‘don’t knows’ was 2·1. Twelve of the twenty chil-dren (60 per cent) solved the navigation problem,drawing a realistic route on the air photo.

TehranSixty children were tested. Few (15 per cent) gaveappropriate answers to the naming questions; themean number of correct feature identifications was2·1.5 Fifty-eight per cent of the children solved the

6

Mexico CityForty-four per cent of the children named the airphoto with an appropriate word or phrase. A meanof 3·0 features were correctly identified. Significantsyntax errors occurred in 16 per cent of theresponses. Sixteen children (80 per cent) solved thenavigation problem.

Table I The children and their performance

Study siteNumber of

childrenIdentifications

(mean number correct)Navigation(% correct)

York 20 5·2 70Durban 20 6·6 60Tehran 60 2·1 58Mexico City 20 3·0 80Evanston 24 4·1 88

Overall 144 4·2 71

Note: children’s ages range from 4,0 to <5,0

Evanston7

Fifty-eight per cent of the children correctly iden-tified the air photo as a map-like image (‘map’,‘town’, etc), 21 per cent responded that they didn’tknow what it was, and 21 per cent gave incorrectresponses.8 The directed identification task askedchildren about eight different landscape features:51 per cent of the responses were correct, 7 per centwere ambiguous, 21 per cent were incorrect andthe remaining 21 per cent were ‘don’t knows’ andnon-responses.9 Relatively few significant syntax

Young children’s mapping abilities 275

errors were noted. Twenty-two of 24 childrenidentified at least one landscape feature; the meannumber of correct identifications was 4·1. On thenavigation task, 88 per cent of the children solvedthe task; eleven out of twelve children in the ‘bird’group were successful, while ten out of twelvechildren in the ‘car’ group were successful.

Summary resultsVerbal responses to the naming questions variedwidely among subjects and from site to site, andwe do not believe that these questions yielded veryuseful results. Nearly all children, in all sites,identified at least one landscape feature; the meansfor correct identifications at the different sitesranged from 2·3 to 6·6; the mean of these site meanswas 4·2. Very few significant syntax errors werenoted. On the navigation problem, the means ofcorrect identification for the different sites rangedfrom 58 per cent to 88 per cent, and 71 per cent ofthe children overall (mean of the site means)solved the problem.

Discussion

The crucial finding is that four-year-olds in allfive cultures demonstrated an ability to interpreta map-like model, an aerial photograph, in thesense of being able to perceive that a downward-looking, scale-reduced image of a landscape is arepresentation of a landscape. Hence, thesechildren can perform the essential syntactictransformations in map-reading: projection orperspective rotation and scale reduction. Sincefour-year-old children would not be able (withoutinstruction) to interpret either conventional map-signs or written notation on a map, the research,by necessity, utilized an iconic map-like modeland did not probe for the children’s ability tocope with the semantics of maps (a topic that isnow being investigated). However, it is clear thatthe crucial problem in map-reading for pre-schoolers is the interpretation of map perspectiveand scale. Therefore, we argue that thisresearch demonstrates that four-year-olds in thecommunities studied can read a map.

Although a common protocol was used in all ofthe sites (with some variation in Evanston), therewere several differences in the methodologies usedat the various sites. As noted previously, the airphotos were not comparable, although all were

vertical images and all had relatively large scales.There were minor differences in test administrationprocedures and testing conditions. There were, wemust assume, Eurocentric biases in the researchdesign. For these reasons, comparative analysis ofthe performance of children in the five cultures isnot warranted. (Indeed, in all cross-cultural behav-ioural research on pre-school children’s cognitiveabilities, the real methodological limitations are soserious that quantitative comparisons are generallyunreliable.) We can, however, note the following:first, most of the methodological limitations wouldtend toward errors in the direction of underesti-mating the cognitive capacities of the non-Westernchildren. Yet the non-Western children whom westudied (in Mexico, Iran and South Africa) per-formed very well indeed. Results from these sitesclearly indicate that map-reading ability is presentin these children. Second, one specific reason whythese data do not warrant comparison of cognitiveabilities across the five cultures is that methodo-logical factors alone provide compelling expla-nations for the difference in scores. For instance,the Tehran and Mexico City feature-identificationscores were undoubtedly influenced by the factthat the aerial photo displayed an unfamiliararea. By the same token, the Evanston scores musthave been affected by the fact that the air photodisplayed a familiar landscape at a very largescale.

These results give some support to the generaliz-ation that map-reading abilities are present, in theabsence of training, in four-year-old children of allcultures. However, the results give strong supportonly for a more limited hypothesis: map-readingabilities are present cross-culturally in four-year-olds in urban communities with relatively highadult-literacy rates.

The findings suggest that mapping abilities,and macrospatial learning as a whole, developmuch more rapidly than is predicted in classicalPiagetian theory, and that if, indeed, there is adiscrete developmental stage in which Piaget’sconcrete spatial operations emerge, they mustemerge in children at or before the age of four.The findings also contradict the thesis that cogni-tive development in non-Western children isslower than that in Western children (Dasen1973), and provide new evidence against thetraditional eurocentric belief that the spatial-cognitive abilities of non-Western peoples aresomehow inferior to those of Westerners.

Mark Blades et al.276

Acknowledgements

The research in the United States, Great Britain,Mexico and South Africa was supported inpart by the National Science Foundation (grantSBR-9423865).

Notes

1 ‘Culture’ is used here in the traditional aggregatesense of referring to a community possessing abroadly common set of psychological, social andmaterial characteristics.

2 The methods and results from some sites arereported in more detail in the following papers: theYork study in Sowden et al. (1996) and the MexicoCity study in Stea et al. (1997).

3 This air photo is reproduced in Sowden et al. (1996).4 The study sites are discussed in a sequence corre-

sponding roughly to the dates that the studies werecompleted.

5 Incorrect answers were not categorized in thisstudy. The number of children who did not identifyany features was not recorded. The identificationscore strongly reflects the fact that landscape fea-tures on the Sheffield air photo are very unfamiliarto Tehran pre-schoolers (see the reproduction of thisair photo in Sowden et al. 1996).

6 The Tehran study, which was carried out independ-ently by Dr Zhra Darvizeh, using only localresources, also tested 60 three-year-olds and 30five-year-olds. The results for all ages on identifi-cation and navigation are shown in Table II.

7 A detailed report on the Evanston study is beingprepared by Dr David Uttal, who designed anddirected it.

8 The second naming question (‘Where do you thinkthis picture was taken from?’) was not asked in theEvanston study.

9 The free identification task was not used in theEvanston study. Most incorrect responses related tofour somewhat confusing features, and responsesfor the other four features (street, car, parking lot,house) averaged 72 per cent correct.

Table II Identification and navigation results, Tehran

Age 3,0–3,5 3,6–3,11 4,0–4,5 4,6–4,11 5,0–5,6

Identifications (mean number correct) 0·6 2·0 2·2 2·0 3·5Navigation task (% solving) 23 30 50 67 63

References

Almy M 1967 The psychologist looks at spatial conceptformation: children’s concepts of space and time in

National Council for Geographic Education ed.Research needs in geographic education: suggestions andpossibilities NCGE, Washington DC 23–40

Blades M and Spencer C 1986 Map use by youngchildren Geography 71 47–52

Blades M and Spencer C 1994 The development ofchildren’s ability to use spatial representationsAdvances in Child Development and Behavior 25 157–99

Blaut J 1997a Children can Annals of the Association ofAmerican Geographers 87 1 152–8

Blaut J 1997b Piagetian pessimism and the mappingabilities of young children Annals of the Association ofAmerican Geographers 87 1 168–77

Blaut J and Stea D 1971 Studies of geographic learningAnnals of the Association of American Geographers 61387–93

Blaut J and Stea D 1974 Mapping at the age of threeJournal of Geography 73 5–9

Blaut J McCleary G and Blaut A 1970 Environmentalmapping in young children Environment and Behavior 2335–49

Bluestein N and Acredolo L 1979 Developmentalchanges in map-reading skills Child Development 50691–7

Boardman D 1983 Spatial concept development and pri-mary school work in Boardman D ed. New directions ingeographical education Falmer Press, London 119–34

Cole M 1996 Cultural psychology Harvard UniversityPress, Cambridge MA

Conning A and Byrne R 1984 Pointing to pre-schoolchildren’s spatial competence: a study in natural set-ting Journal of Environmental Psychology 4 165–75

Dasen P 1973 Piagetian psychology: cross-cultural contribu-tions John Wiley and Sons, New York

DeLoache J 1989 The development of representation inyoung children Advances in Child Development andBehavior 22 1–39

Deregowski J 1989 Real space and represented space:cross-cultural perspectives Behavioural and Brain Sci-ences 12 51–74

Downs R and Liben L 1988 Through a map darkly:understanding maps as spatial representations TheGenetic Epistemologist 16 11–18

Downs R Liben L and Daggs D 1988 On education andgeographers the role of cognitive developmentaltheory in geographic education Annals of the Associationof American Geographers 78 680–700

Eliot J 1970 Children’s spatial visualization in Bacon P ed.Focus on geography: key concepts and teaching strategies

Young children’s mapping abilities 277

National Council for Geographic Education,Washington, DC 263–90

Fairgrieve J 1930 Geography in school University ofLondon Press, London

Gerber R 1981 Young children’s understanding of theelements of maps Teaching Geography 6 128–33

Graves N 1975 Geography in education Heinemann,London

Liben L 1991 Environmental cognition through direct andrepresentational experience: a life-span perspective inEvans G W and Garling T eds Environmental cognitionand action: an interactive multidisciplinary approachOxford University Press, Oxford 245–76

Liben L and Downs R 1989 Understanding maps assymbols: the development of map concepts in childrenAdvances in Child Development and Behavior 22 145–201

Liben L and Downs R 1991 The role of graphic represen-tations in understanding the world in Downs R LibenL and Palermo D eds Visions of aesthetics, the environ-ment and development: the legacy of Joachim F WohlwillErlbaum, Hillsdale NY 139–80

Matthews M 1985a Environmental capabilities of thevery young: some implications for environmentaleducation in primary schools Educational Review 37228–39

Matthews M 1985b Young children’s representation ofthe environment: a comparison of techniques Journal ofEnvironmental Psychology 5 261–78

Matthews M 1992 Making sense of place: children’sunderstanding of large-scale environments HarvesterWheatsheaf, Hemel Hempstead

Miller J 1967 Measuring perspective ability Journal ofGeography 66 167–71

Muir M and Blaut J 1970 The use of aerial photographs inteaching mapping to children in the first grade: anexperimental study The Minnesota Geographer 22 4–19

Naish M 1982 Mental development and the learning ofgeography in Graves N ed. New UNESCO sourcebookfor geography teaching UNESCO, Paris 16–54

Petchenik B 1987 Fundamental considerations aboutatlases for children Cartographica 24 1 16–23

Piaget and Inhelder B 1956 The child’s conception of spaceRoutledge and Kegan Paul, London

Prior F 1959 The place of maps in the junior schoolUnpubl. dissertation, University of Birmingham

Renner G 1951 Learning readiness in elementary geogra-phy Journal of Geography 50 65–74

Rhys W 1972 The development of logical thinking inGraves N ed. New movements in the study and teaching ofgeography Temple Smith, London 93–106

Robinson A and Petchenik B 1976 The nature of mapsUniversity of Chicago Press, Chicago

Rieser J Doxsey P McCarrell N and Brooks P 1982Wayfinding and toddlers’ use of information from anaerial view of a maze Developmental Psychology 18714–20

Sack R 1980 Conceptions of space in social thought: ageographic perspective Macmillan, London

Satterly D 1964 Skills and concepts involved in mapdrawing and map interpretation New Era 45 260–63Reprinted in Bale J Graves N and Walford R eds 1973Perspectives in geographical education Oliver and Boyd,Edinburgh 162–9

Shryock C 1939 Gradations in map learning Journal ofGeography 38 181–7

Sowden S Stea D Spencer C Blades M and Blaut J 1996The mapping abilities of four-year-old children in York,England Journal of Geography 95 107–11

Spencer C and Darvizeh Z 1981 The case for developinga cognitive environmental psychology that does notunderestimate the abilities of young children Journal ofEnvironmental Psychology 1 21–31

Spencer C Blades M and Morsley K 1989 The child in thephysical environment: the development of spatial knowledgeand cognition John Wiley and Sons, Chichester

Spencer C Harrison N and Darvizeh Z 1980 The dev-elopment of iconic mapping ability in young chil-dren International Journal of Early Childhood 12 257–64

Stea D and Blaut J 1973 Some preliminary observationson spatial learning in school children in Downs R andStea D eds Image and environment Aldine, Chicago226–34

Stea D Elguea S and Blaut J 1997 El desarrollo delconocimiento del espacio a la escale macroambientalentre ninos muy jovenes Revista Interamericana dePsicología 31 141–8

Towler J 1970 The elementary school child’s concept ofreference system Journal of Geography 69 83–93

Towler J 1971 Egocentrism: a key to map-reading ability?Social Education 37 893–8

Towler J and Nelson L 1968 The elementary schoolchild’s concept of scale Journal of Geography 67 24–98

Walker R 1980 Map using abilities of five to nine year oldchildren Geographical Education 3 545–54

Werner H 1961 Comparative psychology of mental develop-ment Science Editions, New York

Winn W 1987 Communication, cognition, and children’satlases Cartographica 24 1 61–81