Advances in Child Development and Behavior Volume 12

ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR VOLUME 12 Contributors to This Volume Marc H. Bornstein David Klahr Wil...

Author: Hayne W. Reese | L.P. Lipsitt

302 downloads 816 Views 16MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR

VOLUME 12

Contributors to This Volume Marc H. Bornstein

David Klahr

William J. Friedman

Howard V . Meredith

Howard Gadlin

Bruce M. Ross

Stephen M. Kerst

Robert S . Siegler

ADVANCES IN CHILD DEVELOPMENT AND BEHAVIOR

edited by Hayne W. Reese

Lewis P. Lipsitt

Department of Psychology West Virginia University Morgantown, West Virginia

Department of Psychology Brown University Providence, Rhode Island

VOLUME 12

@

1978

ACADEMIC PRESS New York

San Francisco London

A Subsidiary of Harcourt Brace Jovanovich, Publishers

COPYRIGHT @ 1978, BY ACADEMIC PRESS,INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WFUTING FROM THE PUBLISHER.

ACADEMIC PRESS, INC.

111 Fifth Avenue, New York, New York 10003

United Kingdom Edition published by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road, London N W l 7DX

LIBRARY OF CONGRESS

CATALOG CARD

NUMBER:63-23231

ISBN 0-12-009712-5 PRINTED IN THE UNITED STATES OF AMERICA

Contents List of Contributors Preface

................................................

....................................................

. . .. . . . . . .

vii

.......

ix

Research between 1960 and 1970 on the Standing Height of Young Children in Different Parts of the World I. I1 . I11 . IV . V. VI . VII . VIII . IX . X.

HOWARD V . MEREDJTH Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... Comparisons Among Ethnic Groups . . . . . . . . . . . . . . . . . . . . . . . . . . ................ National and Intranational Group Comparisons . . . . . . Rural and Urban Groups Compared . . . . . . . . . . . . . . . Rural Groups Compared with Urban Socioeconomic Groups . . . . . . . . . . . . . . . . . . . Intracity and Intercity Comparisons . . . . . . . . . . . . . . . . Female and Male Groups Compared . . . . . . . . . . . . . . . Comparisons from Subgrouping for Other Variables . . . . . . . . . . . . Estimates of Population Variability and Their Uses . . . . . . . . . . . . . . . . . . . . . . . . . . . ................ Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . .............................................

2 3 11 17 20 25 36 39 43 48 50

The Representation of Children’s Knowledge DAVID KLAHR AND ROBERT S . SIEGLER ........................ 1. Introduction . . . . . . . . . . . . . . I1. From Behavioral to Cognitive Objectives . . . . . . . . . . . ......... I11 . IV . V. VI . VII . VI11. IX .

Some Criteria for Choosing a Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Balance Scale Task . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experiment I : Assessing Initial Knowledge . ........................ Experiment 2: Training on the Balance Scale ......... Revised Representations for Balance Scale Knowledge ........................ Experiment 3: Encoding Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion: Some Answers and Some Further Questions ...................... X . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

62 63

64 66 69 74 77 95 102 105 113

Chromatic Vision in Infancy MARC H . BORNSTEIN I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Fundamental Data of Color Vision: Ontogeny . . . . . . . . . . . . . . . . .

.. ...

117 123

vi

Contents

111. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

167 169

Developmental Memory Theories: Baldwin and Piaget 1. I1 . Ill . IV . V.

BRUCE M . ROSS AND STEPHEN M . KERST Introduction: Justification and Scope . . . . . . . . . . . . . . J . M . Baldwin’s Theory of Memory . . . . . . . . . . . . . . Piaget’s Theory of Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison Between Memory Theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Topical Memory Research ............................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

184 186 195 210 212 226

Child Discipline and the Pursuit of Self: An Historical Interpretation HOWARD GADLIN I . Introduction . .......................................... II . The Nineteenth Century-The Transformation of the Traditional Family . . . . . . . . . 111. The Nature of Colonial Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1V . Control in the Jacksonian Er.i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . The Ideology of the Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Child Rearing in the Modem Era . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . Child Rearing in the Modem Era-Child Discipline . . . . . . . . . . . . . . . . . . . . . . . . . . VIII . Personhood and Child Discipline in Contemporary America . . . . . . . . . . . . . . . . . . . . IX . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . X . A Note on the Role of Child Development and Psychology . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23 1 235 236 238 240 244 248 253 259 260 261

Development of Time Concepts in Children WILLIAM J . FRIEDMAN Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..................... Logical Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conventional Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experiential Time . . . . . . . . ...... ........................... V . Summary and Conclusion . . . . . . . . . . . . . . . ....................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........

I. I1. 111. IV .

267 269 280 286 294 296

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

299

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

313

Contents of Previous Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

315

List of Contributors Numbers in parentheses indicate the page5 on which the authors’ contributions begin

MARC H. BORNSTEIN Department of Psychology, Princeton IJniversity, Princeton, New Jersey (117) WILLIAM J. FRIEDMAN Department of Psychology, Ohrlin College, Oberlin, Ohio (267) HOWARD GADLIN Department of Psychology, University (fMassachusetts, AmherAt, Mussuchusetts (231) STEPHEN M. KERST Boys Town Centerfor the Study of Youth Development of Cutholic University, Washington, D.C. (183) DAVID KLAHR Depurtment of Psychology, Carnegie-Mellon University, Pittsburgh, Pennsylvaniu (61) HOWARD V. MEREDITH College of Health und Physicul Education, University of South Curolina, Columbia, South Carolina ( 1 ) BRUCE M. ROSS Boys Town Center for the Study of Youlh Development of Cutholic University, Washington, D.C. (183) ROBERT S . SIEGLER Depurtment of Psychology, Curnegie-Mellon University, Pittsburgh Pennsylvania (61)

~

vii

This Page Intentionally Left Blank

Preface The amount of research and theoretical discussion in the field of child development and behavior is so vast that researchers, instructors, and students are confronted with a formidable task in keeping abreast of new developments within their areas of specialization through the use of primary sources, as well as being knowledgeable in areas peripheral to their primary focus of interest. Moreover, there is often simply not enough journal space to permit publication of more speculative kinds of analyses which might spark expanded interest in a problem area or stimulate new modes of attack on the problem. The serial publication Advances in Child Development and Behavior is intended to ease the burden by providing scholarly technical articles serving as reference material and by providing a place for publication of scholarly speculation. In these documented critical reviews, recent advances in the field are summarized and integrated, complexities are exposed, and fresh viewpoints are offered. They should be useful not only to the expert in the area but also to the general reader. No attempt is made to organize each volume around a particular th,-me or topic, nor is the series intended to reflect the development of new fads. Manuscripts are solicited from investigators conducting programmatic work on problems of current and significant interest. The editor often encourages the preparation of critical syntheses dealing intensively with topics of relatively narrow scope but of considerable potential interest to the scientific community. Contributors are encouraged to criticize, integrate, and stimulate, but always within a framework of high scholarship. Although appearance in the volumes is ordinarily by invitation, unsolicited manuscripts will be accepted for review if submitted first in outline form to the editor. All papers-whether invited or submittedreceive careful editorial scrutiny. Invited papers are automatically accepted for publication in principle, but may require revision before final acceptance. Submitted papers receive the same treatment except that they are not automatically accepted for publication even in principle, and may be rejected. We wish to acknowledge with gratitude the aid of our home institutions, West Virginia University and Brown University, which generously provided time and facilities for the preparation of this volume. We also wish to thank Drs. John Hagen, Patricia Self, Barbara Wilcox, and James Youniss for their editorial assistance. Hayne W. Reese Lewis P. Lipsitt ix

This Page Intentionally Left Blank

RESEARCH BETWEEN 1960 AND 1970 ON THE STANDING HEIGHT OF YOUNG CHILDREN IN DIFFERENT PARTS OF THE WORLD

Howard V . Meredith UNIVERSITY OF SOUTH CAROLINA

I . INTRODUCTION II . COMPARISONS AMONG ETHNIC GROUPS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . GROUP DESCRIPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . GROUP COMPARISONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 10

111. NATIONAL AND INTRANATIONAL GROUP COMPARISONS . . . . . . A . GROUP DESCRIPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . GROUP COMPARISONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

II 11 16

IV . RURAL AND URBAN GROUPS COMPARED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . GROUP DESCRIPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . GROUP COMPARISONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17 17 19

V . RURAL GROUPS COMPARED WITH URBAN SOCIOECONOMIC GROUPS . . A . RURAL GROUPS COMPARED WITH URBAN PRIVILEGED GROUPS . . . B . RURAL GROUPS COMPARED WITH URBAN UNDERPRIVILEGED GROUPS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20 20

VI . INTRACITY AND INTERCITY COMPARISONS . . . . . . . . . . . . . . . . . . . . . . . . . . . A . INTRACITY ETHNIC GROUPS COMPARED . . . . . . . . . . . . . . . . . . . . . . . . . . B . INTRACITY SOCIOECONOMIC GROUPS COMPARED . . . . . . . . . . . . . . . . C . INTERCITY COMPARISONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

25 25 27 33

VII . FEMALE AND MALE GROUPS COMPARED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . GROUP DESCRIPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . GROUP COMPARISONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36 36 38

VIII . COMPARISONS FROM SUBGROUPING FOR OTHER VARIABLES . . . . . . . . . A . COMPARISONS FROM SUBGROUPING BY BIRTH WEIGHT AND BIRTH ORDER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . COMPARISONS FOR SINGLE AND TWIN BIRTHS, AND SMOKING WITH NONSMOKING MOTHERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

23

39 40

I

2

Howard V . Meredith

C. COMPARISONS FOR PHYSICAL ABNORMALITY, ELEVATION OF HABITAT, HEALTH STATUS, AND HEALTH CARE . . . . . . . . . . . . . . . . . . 1X. ESTIMATES OF POPULATION VARIABILITY AND THEIR USES.. . . . . . . . . . A. CHOOSING A METHOD OF ESTIMATING POPULATION VARIABILITY IN STANDINGHEIGHT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. STANDARD DEVIATION ESTIMATES OF POPULATION VARIABILITY. X. SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

43

43 45

48 50

I. Introduction The central purpose of this chapter is to compile, systematize, and compare findings from worldwide research on a delimitable aspect of child growth. Consideration is restricted to data on standing height collected primarily during the calendar years from 1960 through 1970. There is further circumscription to the segment of human ontogeny commonly designated early childhood, or the preschool years; within this frame, the focus is on standing height at ages 3 and 5 years. Viewed with explicit reference to the species, calendar period, age span, and somatic dimension under study, the broad objectives of the chapter are as follows: 1. To bring together, in ordered and conveniently usable form, a large quantity of information widely scattered in scientific journals, research monographs, service reports, and unpublished documents. 2 . To provide, by pooling metric data gathered on similar samples of children, statistics adequate to support more precise factual statements and more tenable generalizations than are obtainable from the component studies taken piecemeal. 3. To discover, through alignment and intercomparison of statistics not previously placed in juxtaposition, new facets of biologic knowledge. All sections of the chapter pertain to the standing height of contemporary child populations at ages 3 and 5 years. Sections 11-VII present a substantial array of findings on similarities and differences i n average height among tribal and national groups, rural and urban groups, intracity ethnic and socioeconomic groups, and sex groups. Section IX treats within-group variability in height, and Section VlIl assembles studies, presently sparse, on relations of height in early childhood with birth weight, birth order, maternal use of tobacco during pregnancy, single versus twin delivery, and elevation of habitat. A few of the investigations drawn upon wed measures of body length in dorsal

Standing Height of Young Children

3

recumbency; averages computed from these measures were reduced 1.5 cm as an adjustment to estimated height when standing erect (Meredith, 1960; Roche & Davila, 1974; Smith & Brown, 1970a). Several studies provided averages from standing height at ages 2.5 years, 3.5 years, and so forth; averages for ages 3 years and 5 years were secured by rectilinear interpolation. Most studies reported means for each sex separately; in many parts of the chapter it was parsimonious to combine these means and present values for groups of children taken without regard to sex. Sex differences are dealt with (Section VII) in the same manner as socioeconomic and other subgroup differences within samples. Although more than 200 studies are drawn upon, the coverage of relevant investigations is not exhaustive. Some references were located that could not be procured through interlibrary loan service or by direct overseas correspondence. Moreover, it would be unwarranted to presume that any scholar attempting a colligation of worldwide scope had been able to locate every research report available.

11. Comparisons Among Ethnic Groups A.

GROUP DESCRIPTIONS

Table I presents statistics for standing height drawn from investigations on young children of 42 contemporary ethnic groups. Each row of the table is assigned a numerical tag (Column l), thereby facilitating later reference to either the content of the row or the text description of its source. Column 2 of the table indicates the ethnic group to which the row pertains. Column 4 lists, in ascending order of magnitude, means obtained at age 3 years; means at age 5 years are given in Column 6. The numbers of children measured at the two ages are specified in Columns 3 and 5. Before comparative findings are extracted from Table I, the ethnic group represented in each row will be described, and its bibliographic source(s) cited. Tug 1-1. The data analyzed in 1-1 were collected during 1962-1964 in three Guatemalan villages situated within 50 km of Guatemala City at elevations near 2.2 km; about 95% of the subjects were “indigenous Mayan” Amerinds (Guzman, Scrimshaw, Bruch, & Gordon, 1968; Scrimshaw, Guzman, Kevany, Ascoli, Bruch, & Gordon, 1967). At one of the villages, children “received a food supplement” and “their mothers were given advice on nutrition;” another village provided “improved sanitation and medical services;” while the third lacked these special dietary or health resources. In all three villages “signs suggestive of nutritional deficiency were relatively few. ” Tag 1-2. Malcolm (1 970) obtained the statistics in 1-2 from data gathered in 1967 at villages of the Bundi tribe located at altitudes between .6 and 2.2 km in

Howard V . Meredith

4

TABLE I Mean Standing Height (cm) for 42 Ethnic Groups of Young Children Measured Mainly Between 1960 and 1970 Age 5 years

Age 3 years

Tag

Group

I- I " 1-2 1-3 1-4 1-5

Mayan Bundi Lunda Burmese Lumi Landino Angus and Sura Quechua Shi Lufa Javanese Karkar Melanesian Nivkh Mandinka Antilles black Surinam Bushnegro Surinam Indonesian Antilles Indian and black Surinam lndustani Sumatra Javanese Surinam Amerind Chuvash Eskimo Yoruba Australian Aborigine Azerbaijani Fiji Melanesian Lebanese Surinam Creole Canadian Amerind Hawaiian Japanese Texan Spanish Tatar Lithuanian Blackfeet Hutu Assiniboine and Gros Ventres Kirghiz Lebanese Armenian United States white

1-6

1-7 1-8 1-9 I- 10 1-11 1-12 1-13 I- 14 I- I5 1-16 1-17 1- I8 1- I9 1-20 1-21 1-22 1-23 1-24 1-25 1-26 1-27 1-28 1-29 1-30 1-3I 1-32 1-33 1-34 1-35 1-36 1-37 1-38 1-39 1-40

-

Sample size

Mean

300 I08 1.543 35 40 316 ca. 35 14 237

81.16 83.2 83.6 83.8 83.9 83.9 85.0 85. I 85.2

ca.

ca.

ca.

ca.

ca.

-

-

100

85.4 86.6 86.8' 87.0 87.6 88.3 88.7 89.4 89.6 89.8 89.8 89.9 90.0 90.3 90.5

62 49 43 265 90 59 200 487 I00 95 417 324 480 158

-

-

193 43 838 26 60 67

90.5 90.7 91.3 91.5 91.8 92.4

-

-

34 195 21 357 I25 565

92.6 93.3 94.0 94.5 94.5 94.7

Sample size

ca.

-

-

87 1,640 89 44 294 35 24 367 84

93.3 96.4 96.7 95.6 97.8 101.1 95.4 97.6 98.4

-

ca.

ca.

52 43 34 275 142 565 200 1,838 100

ca.

Mean

124 434 16 660 67 323 178 88 1,539 26

99.8 102.3c 100.3 101.8 103.3 102.6 104.5 104.3 99.7 101.8 102.6 102.9 103.7 104.5 105.2 105.7 103.9 1065

104.0

-

-

89 219 240 37 217 20 873 I79 593

104.6 105.0 107.5 108.8 103.6 108.9 106.7 108.6 108.8 (continued)

5

Standing Height of Young Children

TABLE 1 (continued) Age 5 years

Age 3 years

Tag 1-41 1-42

Group

Sara United States black

Sample size

Mean

-

-

164

95.4

Sample size 63 161

Mean 109.0 110.0

“Read “1-1” as Table I , row I . ”At age 3 years, taking SD,, at 4.4 cni, SE,,,, is .70 cm where n=40, .40 crn where n = 120, .25 cm where n=300, and .11 crn where n = 1500. Corresponding SE,,,,, values at age 5 years, taking SD,, as 5.0 cm, are .79, .46, .29, and . I 3 crn. Woodbury (1921) obtained SD,, values of 4.4 and 5.0 cm from data on large samples of United States white children; sample size exceeded 6500 at each age. Section IX will present findings on variability of standing height in numerous populations. cN’s and means from data collected in 1968 at Sudanese villages about 19 krn south of Khartoum are 62 and 87.0 cm for age 3 years and, for age 5 years, 67 and 102.5 cm (Sukkar, Johnson, Gadir, & Yousif, 1971).

the Madang district of east-central New Guinea. Although physical signs of malnutrition were rare, diet was appraised as insufficient in protein. Tug 1-3. The values in 1-3 were reported by David (1972) on “melanodermic aborigines” of the Lunda tribe measured during 1962-1963 at villages in Angola. The children were described as “Congolese Negroes whose basic ethnic mixture is considered a true negro type strongly influenced by pygmoid elements.” Tug 1-4. In 1961, the data yielding the results i n 1-4 were collected at several widely scattered villages in Burma (Berry, 1963~).A survey team of the Interdepartmental Committee on Nutrition for National Defense (ICNND) found no clinical evidence of caloric insufficiency, protein inadequacy, or deficiency of vitamins A, C, or D; riboflavin and thiamine intakes were low. Tug 1-5. The samples of Lumi children characterized in 1-5 were measured between 1962 and 1967 in villages located at elevations near .5 km on foothills of the Toricelli mountains in the Sepik district (Lumi subdistrict) of north-central New Guinea (Wark & Malcolm, 1969). There were dietary inadequacies in protein and calories, but overt malnutrition was infrequent. Tug 1-6. Records for the study of Ladino (mixed Spanish and Amerind) children were amassed during 1969-1972 at “villages in the department of El Progreso,” Guatemala, situated at altitudes between 0.3 km and 1.1 km (Yarbrough, Habicht, Malina, Lechtig, & Klein, 1975). Although belonging to a group in which “mild-to-moderate protein-calorie malnutrition is endemic,” the children were evaluated as “in clinically good health.” A mean of 83.4 cm was

6

Howard V . Meredith

reported from records in 49 other Guatemalan children residing at villages in which malnutrition and infection were prevalent (Blanco, Acheson, Canosa, & Salomon, 1974). Tug 1-7. The statistics in 1-7 were obtained from measures taken during 1960-1961 on children of the Sura and Angus tribes living in the Pankshin area of the Jos plateau, northern Nigeria (Collis, Dema, & Lesi, 1962). The children were “in a good nutritional state.” Tug 1-8. The values in 1-8 were secured from measures taken during 19641966 on Quechua Amerind children living in the vicinity of Nuiioa, Peru, at altitudes of 3.9 to 4.4 km (Frisancho & Baker, 1970). Diet appeared adequate; the families were largely Andes pastoralists. Tug I-9. Data for the analysis in 1-9 were gathered between 1957 and 1967 on Bantu children of the Shi tribe residing in farming villages of the Ngweshe region, along the western shore of Kivu lake, Republic of Zaire (Vis, 1969, personal communications 1974 and 1975). Tug I-10. The mean in 1-10 was computed from measures taken in 1969 on Lufa children living in the vicinity of the Lufa subdistrict administrative center, eastern highlands, New Guinea (Harvey, personal communication 1974). Tug 1-11. Measures for the study in I- 11 were taken during 1963-1965 at five villages in central Java within a radius of 60 km from Semarang (Tie, Lian, Liong-Ong, & Rose, 1967). The subjects were “in moderately good health.” Tug 1-12. The records analyzed in 1-12 were obtained in 1969 on Melanesian children residing at villages on Karkar Island, northwest of New Guinea (Harvey, 1974, personal communication 1974). Tug 1-13. Arkhipova (1967) obtained the statistics in 1-13 from data collected in 1962 on Nivkh children residing on Sakhalin Island, an island off eastern Siberia, north of Japan. Tag 1-14. During 1962-1963, measures were made on black children, “mostly Mandinkas,” living at four hinterland villages-Jali, Kantongkunda, Keneba, and Manduar-situated about 160 km from Bathurst, Gambia (McGregor, Rahman, Thompson, Billewicz, & Thomson, 1968; Thomson, Billewicz, Thompson, Illsley, Rahman, & McGregor, 1968). Tug 1-15. 1-15presents aggregate sample sizes and weighted means obtained by combining statistics from three studies. The row pertains to children of “black African descent” residing on islands of the Greater and Lesser Antilles. At age 3 years, the component means were 90.4 cm on 67 children measured during 1967-1969 in the rural district of Belair, south central St. Vincent Island (Antrobus, 1971); 89.9 cm on 70 children measured during 1963 in a rural region of Jamaica (Ashcroft, Lovell, George, & Williams, 1965); and 85.0 cm on about 130 children measured in 1961 on the islands of St. Kitts (St. Christopher) and St. Lucia (Berry, 1962~).At age 5 years, component sample sizes and means were 146 and 104.5 cm for Jamaica and about 130 and 98.7 cm for St. Kitts and

Sranding Height o j Young Children

7

St. Lucia. The Jamaican subjects were “predominantly of West African ancestry,” about 95% of the Vincentian subjects had black progenitors, and fully 80% of the subjects from the other islands were members of black families. An ICNND team appraised the diet of children on St. Kitts and St. Lucia as inadequate in protein and riboflavin. For the children of Jamaica, there were “no overt and unequivocal signs of malnutrition,” but dietary intake was low in protein and calories. Tag 1-16. Findings from three studies were poold in deriving the statistics for 1-16. At age 3 years, means of 87.1, 87.6, and 93.1 cm, respectively, were reported on 21 Bushnegroes living along the Surinam river (Luyken & LuykenKoning, 1961), 56 Bushnegroes living along the Tapanahony river (Doornbus, Jonxis, & Visser, 1968), and 13 Bushnegroes described as seminomadic residents of the interior (Van der Kuyp, 1967). At age 5 years, corresponding component means were 99.3, 101.1, and 105.0 cm, with sample sizes of 18, 34, and 90.The Van der Kuyp records were collected during 1964-1965, and the other records between 1958 and 1964. Surinam Bushnegroes are descendants of African slaves who moved to inner regions of Surinam. Tag 1-17. The statistics in I- 17 were obtained on children of Indonesian ancestry measured during 1964-1965 at rural and urban locations in Surinam (Van der Kuyp, 1967). Rural children comprised about 70% of the sample for age 3 years, and about 80% of that for age 5 years. Tug 1-18. Measures taken in 1961 on the islands of Tobago and Trinidad gave the results in 1-18 (Berry, 1952~). The sample was about 50% black, 40% Indian, and 10% mixed; diet was evaluated as inadequate in protein and riboflavin. Tag 1-19. Van der Kuyp (1967) secured the values in 1-19 from measures taken during 1964-1965 on children of Indian and Pakistani descent residing in rural regions and towns of Surinam. At ages 3 and 5 years, respectively, the percentages for children drawn from rural regions approximated 75 and 83%. Tag 1-20. The sample sizes and means in 1-20, reported by Shattock (1968), were computed from measures taken between 1960 and 1965 on offspring of Javanese immigrant families living at a rubber plantation on Sumatra; laboring families at the plantation received “good rations” and “free medical service.” Tug 1-21. Van der Kuyp (1967) obtained the statistics in 1-21 from data collected during 1964-1965 on Amerind children inhabiting the savannah zone of Surinam. Tag 1-22. The Chuvash (predominantly Turko-Tatar) records analyzed in 1-22 were amassed in 1961 on children residing at Cheboksary and on collective farms in the Kanashsky region of the Chuvash Autonomous Soviet Socialist Republic (Goldfeld, Merkova, & Tseimlina, 1965). Each age sample included nearly equal numbers of urban and rural children. Tag 1-23. Records for standing height at age 3 years were accumulated between 1961 and 1965 on 324 Eskimo children living at villages in southwest

8

Howard V . Meredith

Alaska (Heller, Scott, & Hammes, 1967). At age 5 years, a mean of 103.4 cm was obtained on 16 Eskimo males measured during 1968-1969 in the northern part of Upernavik district, Greenland; 103.4 minus .5 cm was taken as the estimated mean at age 5 years for children of both sexes (Drenhaus, SkrobakKacznski, & Jprgensen, 1974). Tug 1-24. Three samples of Yoruba children were pooled. At age 3 years, sample means were 88.0 cm for about 280 children measured during 1957-1963 at Imesi, a village in western Nigeria (Morley, Woodland, Martin, & Allen, 1968), and from measures taken between 1962 and 1970,89.9 cm for 89 children living in the poorest section of Ibadan, and 96.6 cm for 114 children living in well-to-do Ibadan homes (Janes, 1974, personal communication 1975). Corresponding means and sample sizes at age 5 years were 100.1 cm, around 240; 102.0 cm, 236; and 110.2 cm, 185. Tug 1-25. Kettle (1966) secured the statistics in 1-25 from measures on “fullblood aborigines” gathered between 1961 and 1965 at three coastal missions in the Northern Temtory, Australia. Tug 1-26. Data for computation of the mean in 1-26 were accumulated in 1961-1962 on native Azerbaijani children living at Baku, Azerbaijan Autonomous Soviet Socialist Republic (Goldfeld et al., 1965). Tug 1-27. Measures for the analysis in 1-27 were taken during 1959-1968 on Melanesian children residing in coastal villages of the Fiji archipelago (Hawley & Jansen, 1971). About 6% of the subjects showed signs of “borderline malnutrition. ’’ Tug 1-28. The subjects in 1-28, measured in 1961, were largely village children drawn from widely distributed locations in Lebanon. They were “predominantly Arabs,” although the sample included small numbers of “Armenians, Syrians, Egyptians, Palestinians, and Europeans” (Berry, 1962b). An ICNND team appraised the diet as adequate in calcium, marginal in thiamine, and deficient in iron, riboflavin, iodine, and vitamins A and D . Tug 1-29. The large samples of Creole (mixed black and mulatto) children in 1-29 were studied during 1964-1%5 “in Paramaribo, the Coronie district, and Para settlements” of Surinam (Van der Kuyp, 1967). Between 60 and 65% of the children measured at each age were urban residents. Tug 1-30. The means in 1-30 were secured from measures taken on Amerind children living at Ahousat, a Nootka reserve on “Flores Island, off the coast of Vancouver Island,” and at Anaham, a Chilcotin reserve “west of Williams Lake, British Columbia” (Birkbeck, Lee, Myers, & Alfred, 1971). Tug 1-31. The mean in 1-31 was computed from records collected during 1966-1967 on children of Japanese ancestry attending preschools in Honolulu (Smith & Brown, 1970b).

Standing Height of Young Children

9

Tug 1-32. The statistics in 1-32 were obtained on Spanish American children measured during 1%8 in several counties of Texas (McGanity, 1969, personal communication 1974). Tug 1-33. Goldfeld ef ul. (1965) reported the mean in 1-33 on Turkic Tatar children measured during 1961-1962 at Kazan, Tatar Autonomous Soviet Socialist Republic. Tug 1-34. The statistics in 1-34 were computed from records collected at Kaunas and Vilnius during 1961-1962 on native children of the Lithuanian Soviet Socialist Republic (Goldfeld et af., 1965). Tug 1-35. Fisk (1964a) reported the means in 1-35 from measures taken in 1961 on children of the Blackfeet tribe living on the Blackfeet reservation in northwest Montana. Diet was appraised as inadequate for calories, protein, calcium, iron, and vitamins A and C. Tag 1-36. The records for the study in 1-36 were accumulated during 19661967 on children of the Hutu tribe inhabiting the Butare region of Rwanda, along the eastern shore of Lake Kivu; about 70% of the children were “Bantu/Nilotic halfbreeds” and 30% “pure Nilotes” (Vis, 1969, personal communications 1974 and 1975). Tug 1-37. The subjects for the study in 1-37 were Amerinds of the Assiniboine and Gros Ventres tribes living on the Fort Belknap reservation, north-central Montana; they were measured in 1961 (Fisk, 1964d). Dietary evaluation revealed deficiencies in calories, protein, calcium, iron, and vitamins A and C. Tug 1-38. Between 1961 and 1964, data on Kirghiz (largely Mongoloid) children were collected at three urban centers in the Kirghiz Autonomous Soviet Socialist Republic. At age 3 years, sample sizes, means, and sources were 103, 93.1 cm, Naryn (Kudaiarov, 1966); 163, 94.6 cm, Dzhalal-Abad (Kozhonazarov, 1966); and 91, 95.8 cm, Frunze (Kudaiarov, 1966). Sample sizes and means at age 5 years were 197 and 104.6cm for children of Naryn, 404 and 106.7 cm for children of Frunze (Goldfeld et al., 1965; Kudaiarov, 1966) and 272 and 108.3 cm for children of Dzhalal-Abad. Naryn is situated at an elevation near 2.0 km; elevations of the other towns were given as near .8 km. Tug 1-39. The records analyzed in 1-39 were obtained on children of Armenian ancestry belonging to “upper and middle class” families residing at Beirut (S. J. Karayan, personal communication 1976). Tug 1-40. Two samples were combined at each age in 1-40. At age 3 years, component means were 94.3 cm for 186 children measured at several locations selected to yield samples “representative” of the United States white “noninstitutionalized population” in 197 1-1972 (Abraham, Lowenstein, & O’Connell, 1975), and 94.9 cm for 379 white children of all socioeconomic classes measured during 1968-1970 at urban and rural locations in 36 states and the District of Columbia (Owen, Kram, Garry, Lowe, & Lubin, 1974; Owen &

10

Howard V . Meredith

Lubin, 1973). Corresponding means and sample sizes at age 5 years were 109.3 cm, 214; and 108.5 cm, 378. Tag 1-41, The values at age 5 years in 1-4 1 were secured from measures on Sara children gathered in 1965 at preschools in Fort Archambault, Republic of Chad (Hiernaux & Asnes, 1974). The children were from homes of middle and upper socioeconomic classes; they showed no signs of malnutrition. Tag 1-42. At each age in 1-42, samples were pooled from the investigations discussed in 1-40. Means at age 3 years were 95.1 cm on 98 United States black children measured during 1971-1 972, and 95.8 cm on 66 United States black children measured during 1968-1970. Means and sample sizes at age 5 years were 110.1 cm, 96; and 109.8 cm, 65. B . GROUP COMPARISONS

Among the comparative findings of average standing height that can be drawn from Table I, are the following: 1. At age 3 years, the mean of 95.4 cm on black children of the United States exceeds the mean of 8 1.1 cm on Mayan village children of Guatemala by about 14 cm, or 17%. At age 5 years, the mean for United States black children exceeds that for Bundi village children of New Guinea by almost 17 cm, or 18%. The means obtained on Hutu rural children age 3 years and Bundi rural children 2 years older are identical; a similar relationship holds for United States black children age 3 years and Peruvian Quechua children 2 years older. 2. Lunda rural children of Angola are much shorter than black children of the United States. One-tailed significance tests a t p = .01 allow the inference: During the childhood period between ages 3 years and 5 years, average height in the Lunda rural population is more than 10 cm below that of the United States black population. 3. The means in Table I for Amerind groups are dispersed at each age across a 13 cm range. At age 3 years, means are 81.1 cm (Mayan), 85.1 cm (Quechua), 89.8 cm (Surinam Amerind), 91.5 cm (Canadian Amerind), 92.6 cm (Blackfeet), and 94.9 cm (Assiniboine and Gros Ventres). Corresponding means at age 5 years, beginning with the Quechua tribe, are 95.4, 101.8, 104.0, 108.8 cm, and 108.9 cm. Since most of these Amerind populations are represented by small samples, the variation of 13 cm among sample means may be as much as one-third larger than the greatest difference between the population means. 4. At age 3 years, means are near 84 cm for Ladino rural children of Guatemala; near 90 cm for Chuvash children of Chuvashia, Eskimo children of southwest Alaska, and Industani children of Surinam; and near 94 cm for Kirghiz urban children of Kirghizia. At age 5 years, means approximate 98 cm for Shi rural children of Zaire; 103 cm for Hutu rural children of Rwanda and Indonesian

Standing Height cf Young Children

11

children of Surinam; and 107 cm for Creole children of Surinam and Kirghiz urban children. For each of these groups the number of children measured exceeds 200.

111. National and Intranational Group Comparisons A.

GROUP DESCRIPTIONS

Presmted in Table I1 are statistics for standing height drawn from investigations on samples of contemporary young children representative of 14 nations and 14 geographical sections of countries. The term “intranational group” is used to denote those members of a nation living in a particular province, prefecture, state, department, county, or other geographical section of the country that has both urban and rural residents. The source and composition of the groups named in successive rows of Table I1 are as follows: Tug 11-1. The means in 11- 1 were derived from measures of standing height collected during 1962-1963 on a sample of Pakistani children considered to represent “all geographic and economic sections” of East Pakistan; about 30% of the subjects were urban residents and 70% lived in rural areas (Rosenberg & Reiner, 1966). Malnutrition was prevalent; the children’s diets were appraised by an ICNND team as commonly deficient in calories, protein, vitamin A , riboflavin, and iron. Tag 11-2. The means in 11-2 were secured from measures taken in 1962 at locations in several regions of Hashemite Jordan (Berry, 1963a). An ICNND nutrition team found that among Jordanian children there was considerable calorie-protein malnutrition, and insufficient intake of vitamin A and riboflavin. Tag 11-3. The statistics in 11-3 pertain to children of the Federation of Malaya; data were gathered in 1962 at representative geographic locations, urban and rural, and from all socioeconomic strata of the Malayan peninsula and the island of Panang (Fisk, 1964~).Diet was evaluated by an ICNND team as adequate in niacin, ascorbic acid, and vitamin D, but commonly inadequate in calcium, riboflavin, and thiamine. Population composition for the Federation of Malaya in 1960 approximated: Malays, 50%, Chinese, 40%, and Indians, 10%. Tag 11-4. 11-4 gives sample sizes and means obtained from a broad sampling of children made during 1965 in the savannah, plateau, rain forest, and coastal regions of the Republic of Nigeria (Darby & Edozien, 1967). Diet was appraised as “severely deficient” in riboflavin. Tag ZI-5. The means in 11-5, based on records amassed in 1965, were considered to represent the Paraguayan child population except for “the Mennonites at

Howard V . Meredith

12

TABLE I1 Mean Standing Height (cm) for 28 National and lntranational Groups of Young Children Measured Largely Between 1960 and 1970 Age 3 years

Tag

Country (whole or portion)

Sample size

Age 5 years

Mean

Sample size

Mean

~

11- 1

11-2 11-3 11-4 11-5 11-6 11-7 11-8 11-9 11- 10 11-1I

11-12 11- I3 11-14

11- 15

11- I6 11- I7 11- I8 11- I9 11-20 11-2I 11-22 11-23 11-24 11-25 11-26 11-27 11-28

East Pakistan Jordan Malaya Nigeria Paraguay Venezuela Uruguay China Republic Romania Bulgaria United States Finland East Germany Netherlands

Bolivian regions Tanzanian region Philippines region Tunisian region Brazilian states Cambodian region Indian province Singapore region Japanese prefecture Italian provinces Italian department Belgian province United States region Australian state

National samples 25 I I40 20 I 72 28 1 I86 130 1,188 >3,900

-

-

729 485

94.8 94.9

-

-

1,252

97.2

lntranational samples I66 34 433 93 158 ca. 200 79 389

ca.

82.3 83.6 85.0 85.2 88.2 89.0 90.9 92.4 93.5

1,040 42 1

82.7 83.9 85.3 85.3 85.9" 88.4 89.7 91.1 93.9 94. I

-

-

580 1,114

94.5 95.9

-

-

200

97.1 97.8 105.3 102.4 102.0 106.0 105.0

164

42 299 253 145 2,862

ca.

ca.

-

-

1.923 754 420 3,697 1,624

107.2 109.0 108.5 110.0 111.3

220 56 452 92 166 200

98.4 99.7 97.5 100.7 98.6 100.9

-

-

100

103.1 104.0 107.0 108.4 108.5 109.5

6,014 1,410 588 625 659

-

-

=A mean of 86.3 cm at age 3 years was obtained from measures taken in 1970 on 146 African children living at villages in the Lower Shire area of southern Malawi (Burgess, Burgess & Wheeler, 1973).

Standing Heighr oj Young Children

13

Sommerfeld and Filadelfia, the Japanese at Pirapo, and the Indians in the Chaco” (Centurion, Miranda, & Watkin, 1967). Tug If-6. The means in 11-6 for this 1963 study of Venezuelan children were drawn in the areas of Los Andes, Lake Maracaibo, Central Zone, Los Llanos, Crinoco-Caroni, and Isle de Margarita (Fisk, 1964e). About 65% of the subjects were mestizo, 20% white, and 15% black or Amerind. Data gathered by an ICNND nutrition team showed that diet was adequate for vitamins A, C, and niacin, but moderately substandard for protein and calories. Approximately 40% of the subjects lived in rural areas and 60% were urban residents. Tug 11-7. 11-7 displays means reported as representative of Uruguayan children in 1962; measures were accumulated from six scattered locations between Colonia Lavalleja in the north and Montevideo in the south (Berry, 1963b). The subjects were largely of Spanish and Italian descent; about 10% were offspring from white-Amerind or white-black intermarriages. The ICNND nutritionists evaluated the diet as adequate in calories, protein, riboflavin, niacin, and iron, borderline in vitamins A, C, and thiamine, and frequently inadequate in iodine. Tug 11-8. The statistics in 11-8 were obtained from data amassed during 1970-1972 at “randomly selected nurseries and kindergartens” in “rural and urban areas of Taiwan and Penghu (Pescadores);” the most prevalent signs of nutritional deficiency indicated riboflavin and niacin insufficiencies (Chen, Chiang, Huang, & Chen, 1974). Tug If-9. The large sample of Romanian children characterized in 11-9 were measured in 1964 (Tinisescu, Chiriac, Stinciulescu, Domilescu, & Jelezneac, 1970). About 44% of the subjects lived at urban centers and 56% in rural communities. Tug If-10. During 1960-1961, measures on Bulgarian children age 5 years were amassed at Sofia and throughout the country at “towns and villages determined in a stochastic manner” (Bulgarian Academy of Sciences, 1965). Subgrouped by population density, 37% of the measures in 11-10 were taken in Sofia, 38% in other towns, and 25% in villages. Tug I f - 1 1 . Statistics from two studies of United States children were combined in 11- 11: At age 3 years, component means were 95.0 cm on 445 white and black children of all socioeconomic classes measured during 1968-1970 at urban and rural locations in 36 states and the District of Columbia (Owen & Lubin, 1973; Owen et ul., 1974), and 94.6 cm on 284 white and black children measured during 1971-1972 at 35 locations selected to yield samples “representative of the total. . . noninstitutionalized population” (Abraham et ul., 1975). Listed in the same order, component means and sample sizes at age 5 years were 108.7 cm, 443; and 109.5 cm, 31 1. The composite samples were about 78% white and 22% black.

14

Howard V . Meredirh

Tug 11-12. The means in 11-12 were considered to typify the child population of Finland in the early 1960s; data accumulation occurred between 1955 and 1969 at urban centers and in rural regions (Backstrom & Kantero, 1971). Tug 11-13. The large sample in 11-13, measured in 1967, was reported as representative of East German kindergarten children age 5 years (Oehmisch, 1970). Tug 11-14. Van Wieringen (1972) secured the statistics in 11-14 from data amassed during 1964-1966 in a large-scale Netherlands survey. The mean of 1 11.3 cm at age 5 years practically coincides with the mean of 1 11.4 cm obtained by Van der Kuyp (1967) on ‘70 white children, primarily of Dutch lineage, measured during 1964-1965 in Surinam. Tug 11-15. Two samples in 11-15 were pooled at each age, one drawn in 1962 from eight departments of southwest Bolivia, and the other in 1964-1967 from rural communities and small towns of west-central Bolivia. Both included children of Amerind (largely Quechua and Aymara), Spanish, and Amerind-Spanish ancestry. The west central sample sizes and means were 44 ana 80.9 cm at age 3 years and 64 and 97.9 cm at age 5 years (Omran, McEwen, & Zaki, 1967); these children lived at altitudes varying from .2 km to 4.1 km, their diet was fairly poor, and infections were common. Sample sizes and means from the southwestem study were 122 and 83.3 cm at age 3 years and 156 and 98.6 at age 5 years (Fisk, 1964b); an ICNND team evaluated the diet of these children as frequently deficient in calcium, protein, vitamin A, riboflavin, and iodine. Tug 11-16. The mean in 11-16 was computed from data collected during 1966-1967 on Bantu children of the Pare, Shambala, and Chagga tribes living at highland and lowland villages in the Kilimanjaro and Tanga zones of northeast Tanzania (Kreysler and Schlage, 1969). Tug 11-17. Records for the study in 11-17 were accumulated between 1958 and 1966 in connection with nutrition surveys in nine regions of the Philippine Islands considered representative of “Luzon and the Visayas” (Matawaran and Gervasio, 197 1; Matawaran, Gervasio and de Gala, 1966). Tug 11-18. The statistics in 11-18 pertain to children of “Arab, Negro, and Berger” ancestry living at villages around the Chott el Djerid, a large lake in southern Tunisia bordering the Sahara desert (Boutourline, Tesi, Kerr, Stare, Kallal, Turki, & Hemaidan, 1972). The primary dietary staple of these children was unfortified wheat, “animal protein was consumed in extremely small amounts. ’ ’ Tug 11-19, The means in 11- 19 were computed from data gathered during 1963 in six states of northeast Brazil; about 85% of the children were residents of urban communities and 15% lived in rural areas (Fisk, 1965). Although “physical manifestations of nutritional deficiencies were not highly prevalent,” dietary intakes were appraised as insufficient for calories, protein, and vitamin A, and

Standing Height of Young Children

15

marginal for riboflavin and thiamine. Respiratory infections, bleeding gums, and diarrhea were fairly common. Tug 11-20. Data for the study in 11-20 were gathered during 1964-1966 on offspring of Cambodian parents residing in Pnompenh and the surrounding region (Nouth-Savoeun, 1966). Tug 11-21. The mean in 11-2 1 was obtained on Indian children measured at child development clinics and nursery schools “scattered over Agra province,” Uttar Pradesh; subjects had neither “a history of illness likely to retard growth,” nor “any serious abnormality detected on examination” (Prasad, Kumar, & Dayal, 197 1). Tug 11-22. These children in 11-22 were drawn from the south-central part of Singapore Island (Wong, Tye, & Quek, 1972). The ethnic composition of the sampling was about 68% Chinese, 16% Malayan, and 11% Indian; the socioeconomic grouping of the children’s fathers approximated 12% professional and managerial, 36% clerical and skilled, and 52% semiskilled and unskilled. Tug 11-23. In 1963, the large sample of Japanese children in 11-23 was measured in ‘‘municipal, coastal, agricultural, mountainous, and insular” areas of Kagawa prefecture (Kambara, 1969, personal communication 1973). Kambara reported that the national mean for Japanese kindergarten children measured in 1963 was 105.0 cm at age 5 years. Tag 11-24. The aggregate sample sizes and weighted means in 11-24 were derived from samples of Italian children studied between 1960 and 1968 in seven northern and central provinces. Component sample sizes, means, and sources at age 3 years were 147, 92.7 cm, Padua province (Bussadori, 1965); 350, 93.4 cm, Grosseto province (Terrosi, 1968); about 100, 93.9 cm, Pisa province (Vizzoni & Baldini, 1964); about 85, 94.4 cm, Trento province (Marzot, Colombini, & Vedovello, 1966); about 185, 94.6 cm, Teramo province (ZauliNaldi, 1963); 14, 94.6 cm, Fruili province (DeLuca, Bearzotti, & Cozzi, 1965); and 159, 95.0 cm, Aquila province (Vesi & Cantalini, 1972). From the same sources, sample sizes and means at age 5 years were 378 and 105.5 cm (Padua); 453 and 106.0cm (Grosseto); about 85 and 106.2 cm (Trento); 149 and 108.5 cm (Aquila); about 185 and 108.9 cm (Teramo); about 100 and 109.4 cm (Pisa); 67 and 11 1.4 cm (Fruili). Tug 11-25. Tonelli (1963) reported the statistics in 11-25 on children living in the Italian department of Emilia, a department situated about midway between the center of the Italian mainland and its northern boundary. Tug 11-26. Data for the study in 11-26 were collected during 1970-1971 on physically normal, healthy children measured at 20 centers of preventive medicine in the province of Liege, eastern Belgium (Geubelle, Lambrechts, Sabatier, & Baltia, 1974).

16

Howard V . Meredith

Tug 11-27. The means in 11-27 were obtained from records accumulated during 1968-1970 in 12 states: Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, and Wisconsin (Fryer, Lamkin, Vivian, Eppright. & Fox, 1972). The composite samples were considered “representative” of the “heterogeneous population” residing in the north central region of the United States. Tug 11-28. The measures analyzed in 11-28 were amassed during 1970-1971 at child health centers in urban and rural areas of Tasmania (Coy, Lewis, Mair, Longmore, & Ratkowsky, 1973). Occupation of the children’s fathers was about 15% professional and managerial; 22% skilled, nonmanual; 38% skilled, manual; and 25% semiskilled and unskilled. B . GROUP COMPARISONS

Examples of comparative findings that can be deduced from Table 11, separately and in conjunction with Table I, are as follows: 1. In the period between 1960 and 1970, averages for human standing height at age 3 years approximated 82, 85, 88, 93, 95, and 97 cm, respectively, on children of East Pakistan, Malaya, Paraguay, Romania, Finland, and the Netherlands. On using one-tailed statistical procedures to determine, a t p = .01, fiducia1 limits for population means, the ensuing generalization appears tenable: During the l%Os, average standing height at age 3 years was less for children of East Pakistan than for their age peers of Romania by 10 cm or more, and of the Netherlands by 13 cm or more. 2 . At age 5 years, averages near 97, 102, 105, 108, and 11 1 cm, in this order, were obtained for children of Jordan, Venezuela, the Republic of China, Finland, and the Netherlands. Averages near 98, 104, 107, and 110 cm characterized Bolivian children residing in southwest and west-central Bolivia, Japanese children inhabiting Kagawa province, Bulgarian children, and children of East Germany. In the early 1960s the typical Jordanian child age 5 years was no taller than the typical Dutch child at age 3 years. 3. Tables I and I1 carry averages for standing height at age 3 years near 85 cm on Shi, Malayan, and Filipino groups (1-9, 11-3, 11-17); near 88 cm on Antilles black, Paraguyan, and Cambodian groups (1-15, 11-5, 11-20); near 91 cm on Surinam Creole and Singapore groups (1-29, 11-22); and near 94 cm on Kirghiz and Italian groups (1-38, 11-24, 11-25). Similar averages at age 5 years are tabled for Ladino, Shi, and Filipino groups (1-6, 1-9, 11-17); Antilles black, Surinam Indonesian, Chuvash, Paraguayan, and Venezuelan groups (I- 15, I- 17, 1-22, 11-5, 11-6); and Surinam Creole, Kirghiz, Bulgarian, and Italian groups (1-29, 1-38, 11-10, 11-24). Sample size for each of the groups referred to in this paragraph and in the two preceding paragraphs was 200 or more.

17

Standing Height cf Young Children

IV. Rural and Urban Groups Compared A.

GROUP DESCRIPTIONS

Averages for standing height on contemporary groups of children living in rural areas and urban communities are assembled in Table 111. The upper part of the table pertains to children age 3 years, the lower part to children age 5 years, and the column at the extreme right to differences between paired urban and rural groups. Information on the content in successive rows of Table I11 is provided below.

TABLE 111 Mean Standing Height (cin) for Rural and Urban Groups of Young Children Measured Mainly Between 1960 and 1970 Rural

Tag

Group

111-1 111-2 111-3 111-4 111-5 111-6 111-7 111-9 111-10

Indian Nigerian South African Bantu Polish Czechoslovak Romanian Chuvash Russian Surinam Hindustani Surinam Creole

111-1 I 111- 12 111- I3 111-14 111-15 111-16 111- I7 111- I8 111-19 111-20

Nigerian Chuvash Bulgarian Polish Czechoslovak Russian South African Bantu Surinam Hindustani Surinam Creole Surinam Indonesian

Sample size Age

111-8

ca.

1,529 315 151

365 61 >2,200 201 ca. 440 367 300

Urban

Mean

Sample size

Mean

Difference"

3 years 19.5 87.7 89.3 92.7 94.6 92.4 88.8 93.9 89.6 91.9

324 282 265 ca. 785 1,032 > 1,700 2 I6 ca. 655 120 538

86.4 92.5 91.8 95.2 97.0 94.7 90.9 93.9 89.0 90.9

6.9 4.8 2.5 2.5 2.4 2.3 2. I .0 -.6 -1.0

49 I 211

105.1 104.5 107.9 108.4 111.0 107.8 104.3 104.5 106.6 102.0

4.9 3.7 2.6 2.4 2.2 1.3 .8 .3 .2 - .7

Age 5 years

"Urban mean minus rural mean

ca.

ca.

275 223 479 423 124 440 128 1,516 599 463

100.2 100.8 105.3 106.0 108.8 106.5 103.5 104.2 106.4 102.7

1.444

ca. ca. ca.

585 765 1,920 208 322 940 I02

18

Howard V . Meredith

Tug 111-1. The statistics in 111- 1 on Indian rural children were obtained from measures taken at villages within a radius of 50 km beyond Hyderabad; 40% of the children showed nutritional deficiency signs for vitamin A, vitamin B complex, or calorie-protein intake (Madhavan, Susheela, & Swaminathan, 1967; Rao, Singh, & Swaminathan, 1969; Swaminathan, Jyothi, Singh, Madhavan, & Gopalan, 1964). The urban values were secured using data collected in Bombay on Indian children of all socioeconomic strata; the sample was heavily weighted with children from the lower and lower middle classes (Shiddhaye, Shah, & Udani, 1972; Udani, 1963). An alternative rural-urban analysis was made; this yielded a difference between means of 7.4 cm. A rural mean of 80.1 cm resulted from pooling the study at villages around Hyderabad with studies at villages in the Lubhiana district of Punjab (Neumann, Shanker, & Uberoi, 1969) and villages in the vicinity of Poona (Phadke, Deodhar, & Kulkarni, 1971). An urban mean of 87.5 cm was obtained on combining the Bombay sample with samples studied in Delhi (Banik, Krishna, Mane, Raj, & Tasker, 1970c), Burla (Mohanta, Pande, & Praharaj, 1972), Bhopal and Gwalior (Sharma, 1970), and Kerala (Gokulanathan & Verghese, 1969). Socioeconomically, these samples were from all levels (Delhi), the lower and middle levels (Burla), the middle classes (Bhopal and Gwalior), and the middle and upper classes (Kerala). Composite sample sizes were 2100 and 1050 for rural and urban children, respectively. Tug 111-2. In 111-2, the rural mean was derived from data accumulated at villages in northern and western Nigeria (1-7, 1-24), and the urban mean from data accumulated at Ibadan (1-24) and at Lagos on Ibo and Yoruba children of all socioeconomic classes measured during 1963-1965 (Rea, 1971). Tug 111-3. Between 1968 arid 1970, records for the analyses in 111-3 were collected on rural children of northeastern and eastern Transvaal, and on urban children of Johannesburg (Richardson, 1973, personal communication 1974). Tug 111-4. Sample sizes, means, and sources used to obtain the values in 111-4 for Polish rural children were 98, 92.0 cm, Nowy Targ district (Migsowicz & P y h k , 1967); 172, 92.8 cm, Lublin district (Chrzastek-Spruch, personal communication 1974); 33, 92.5 cm, Ostroleka and Suwalki districts (Wolanski & PyBuk, 1973); and 62, 93.5 cm, Puck district (PyEuk & Wolanski, 1968). For urban children, component materials were 189, 94.4 cm, Lublin (ChrzastekSpruch & Szajner-Milart, 1973); and about 600, 95.4 cm, Warsaw (Kurniewicz-Witczakowa, Migsowicz, Mazurczak, & Eska, 1973; Wolanski, 1964). Tug 111-5. The rural values in 111-5 were computed from measures taken in 1965 on children living in a rural region near Olomouc, Czechoslovakia (Holibkova & Holibka, 1968). Sample means used in deriving the composite urban mean were 96.5 cm on 469 children measured at Brno during 1964-1972 (Gerylovova & Bouchalova, 1974); 97.4 cm on 434 children measured at Prague during 1961-1964 (Kubat, Novakova, Koufim, & Syrovatka, 1969); and 97.6

Standing Height of Young Children

19

cm on 129 children measured at Bratislava during 1961-1967 (BakiEova & SevEikova, 1973). Tugs 111-6 and 111-7. See 11-9 and 1-22. Tug 111-8. Samples of Russian children for 111-8 were combined from two studies. Component sample sizes and means were 242 and 93.8 cm for data collected in 1966 at villages of the Belgorod-Dnestrov and Tatarbunarsk districts of the Odessa region (Kravchenko, 1968) and about 200 and 94.0 cm for data gathered in 1970 at villages in the Michurinsky and Sampursky areas of Tambov province (Arkhipova, 1972). Four urban samples studied between 1960 and 1963 were pooled: their sample sizes, means, and sources were about 170, 93.8 cm, Ivan-Frankovsk (Kutsenko, 1965); 261,93.6 cm, Rostov-on-Don; 161,94.0 cm, Kursk; and 65, 95.5 cm, Ryazan (Goldfield et al., 1965). Tugs 111-9 through 111-13. See 1-19, 1-29, 111-2, 1-22, 11-10. Tug 111-14. Rural particulars in 111- 14 were 87 and 106.5 cm, Cracow county (Gastol, 1966); 58 and 104.4 cm, Suwalki and Ostroleka districts; 182 and 105.6 cm, Lublin district; 43 and 106.6 cm, Nowy Targ district; and 54 and 107.6 cm, Puck district (see 111-4). Urban subgroup sample sizes, means, and places were 308, 107.9 cm, Lublin (Chrzastek-Spruch & Szajner-Milart, 1973); and about 280, 108.9 crn, Warsaw (Kumiewicz-Witczakowa, Migsowicz, & Mazurczak, 1974). Tug 111-15. See 111-5. Urban subgroup sample sizes and means in 111-15 were 279 and 110.7 cm (Prague); about 100 and 110.7 cm (Bratislava); and 390, 11.3 cm (Brno). Tug 111-16. Refer to 111-8. The component statistics at age 5 years in 111-16 were 243 and 106.5 cm, Odessa region; about 200 and 106.6 cm, Tambov region; 265 and 107.1 cm, Ryazan; 599 and 107.5 cm, Kursk; about 370 and 107.7 cm, Ore1 (Belousov, Kardashenko, Kondakova-Varlamova, Prokhorova, & Sgromskaia, 1968); and 689 and 108.4 cm, Donetsk (Goldfeld et ul., 1965; Pokatilo, 1974). Tugs 111-17 through 111-20. See 111-3, 1-19, 1-29, and 1-17. B. GROUP COMPARISONS

Inspection and evaluation of Table 111, togelher with reference to other pertinent findings, reveals: 1. Averages for standing height of young children residing in cities are neither consistently higher nor lower than those of age peers residing in rural areas. 2. The three Indian and Nigerian comparisons, appraised statistically at p = .01, support the inference that mean height for each urban population was greater than mean height for the corresponding rural population by more than 3.5 cm.

20

Howard V . Meredith

3. Values from the samples in 111-4 and 111-14, subjected to significance tests at p = .01, allow the generalization: During the period 1960-1970, population means for height at ages from 3 to 5 years for Polish urban children surpassed those for Polish rural age peers by amounts likely between 1.4 and 3.5 cm. Direction and magnitude of the rural-urban differences appear similar for Romanian and Bulgarian young children. 4. Rural-urban differences between sample means for standing height of Surinam Hindustani and Creole children do not allow rejection of the hypothesis that the rural and urban population means for these groups were alike. Moreover, earlier large-scale studies in England (Tuxford & Glegg, 191 1) and the United States (Woodbury, 1921) found “averages for children in rural areas” were slightly “above those for children i n urban areas.” Records for standing height at age 5 years were amassed in England during 1909-1910 on 44,200 children living in 17 rural regions and 38,600 children living at 44 urban centers; the rural mean was 1.3 cm higher than the urban mean. Data were collected in 1918 on offspring under age 6 years of parents born in the United States and residing in “Iowa and eastern North-Central states”; comparison of means for 45,000 rural and 23,700 urban children showed the rural children were taller by .6 cm.

V.

Rural Groups Compared with Urban Socioeconomic Groups

In the present section, averages for standing height of children living in rural areas are compared with averages on “urban privileged” children (Table IV) and “urban underprivileged” children (Table V). The category “urban privileged” encompasses samples of urban children drawn from high income families, elite residential neighborhoods, professional and managerial occupational groups, or private schools; the category ‘ ‘urban underprivileged” includes samples of urban children drawn from low income families, slum districts, unskilled occupational groups, or the indigent social class. A.

RURAL GROUPS COMPARED WITH URBAN PRIVILEGED GROUPS

Materials complementing the contents of Table IV are as follows: Tag IF‘-I. “Rural” component sample sizes and means in IV- 1 were 1529 and 79.5 cm from data amassed at 37 villages of Andhra Pradesh in the region around Hyderabad (Rao ef al., 1969) and, from measures taken at randomly selected villages in all directions around Poona, 556 and 81.8 cm (Phadke, personal communication 1973; Phadke et al., 1971); both samples were drawn largely from “poor homes” in which food, sanitation, and personal hygiene were “far from satisfactory.” For the “urban privileged,” component sample sizes and

21

Standing Height of Young Children

TABLE IV Mean Standing Height (cm) of Rural and Urban Privileged Groups of Young Children Measured Mainly During 1960- I970 Rural

Tag

Group

1v-I IV-2 IV-3 1V-4 IV-5 1V-6 1V-7

Indian Thai Nigerian Turkish Malayan South African Bantu Russian

IV-8 IV-9 1v-I0 1v-1I IV-12 IV-13 1V-14 1V-15 1V-16

Indian Turkish Nigerian Thai Spanish Ethiopian Jamaican white South African Bantu Russian

Sample size

ca. ca. ca.

ca. ca. ca. ca.

Urban privileged

Mean

Age 3 years 2,085 80.1 185 83.0 31.5 87.7 21 86.9 4.5 83.7 151 89.3 177 93.2 Age 5 years 300 95.7 19 97.7 275 101.2 165 96.4 12,187 103.9 SO 101.3 171 106.4 128 103.5 192 108.2

Sample size

Mean

Differencen

ca. 295 156 133 ca. 70 ca. 140 178 92

94.9 93.2 96.4 93.3 88.8 93.3 96.8

14.8 10.2 8.7 6.4 5. I 4.0 3.6

ca. 280 ca. 70 196 243

107. I 107.8 110.2 104.6 110.9 104.1 109.1 105.8 109.5

11.4 10.1 9.0 8.2 7.0 2.8 2.7 2.3 1.3

100

81 76 134 122

“Urban privileged mean minus rural mean.

means were about 34 and 94.0 cm on children of “the best socioeconomic group” in Bombay (Udani, 1963); 62 and 95.0 cm on “well-privileged’’ children of wealthy families in Delhi (Banik, Nayar, Krishna, & Raj, 1972); and, on healthy children from upper and upper-middle income families of Poona, about 200 and 95.0 cm (Limaye, Chouhan, Lakhani, & Phadke, 1974). Tug W - 2 . The rural values in IV-2 were obtained by combining sample sizes and means of 95 and 8 1.5 cm on children measured in 1960 at villages scattered throughout Thailand (Berry, 1962a) and, on children measured some years later at two villages in northeast Thailand, about 90 and 84.5 cm (Migasena, Thumham, Pongpaew, Hongthong, & Harinasuta, 1974). The adjacent sample sizes and means in IV-2 were secured from records accumulated during 1967-197 1 on well-nourished children from “middle and professional class homes” in Bangkok (Khanjanasthiti, Supachaturas, Mekanandha, Srimusikapodh, Choopanya, & Leesuwan, 1973).

22

Howard V . Meredith

Tug ZV-3. For sources of the rural statistics, see 1-7 and 1-24. Sample sizes and means in IV-3 for the “urban privileged” samples were 19 and 95.5 cm on children of “upper class” families in Lagos (111-2) and, on children of “wellto-do” families in Ibadan, 114 and 96.6 cm (1-24). Tug ZV-4. Values in the rural part of IV-4 were obtained on children measured at a village 120 km from Istanbul (Neyzi, 1967) and those in the urban privileged part on children of wealthy families residing in the Nisantasi area of Istanbul (Neyzi & Gurson, 1966). Tug ZV-5. These paired means in IV-5 for “the Malay ethnic group” were secured on village children measured in 1968 at Ulu Trengganu, a district about 480 km from Kuala Lumpur, and on healthy children of above average socioeconomic status-a few of the latter were “urban elite” measured in 1969 at Kuala Lumpur, the bulk were measured in !968 at a nearby military base (Dugdale, MacKay, Lim, & Notaney, 1972; McKay, Lim, Notaney, & Dugdale, 1971). Tug ZV-6. Refer to 111-3. The “privileged” children in IV-6 attended nursery schools in Soweto, Johannesburg; they ate 10 meals each week at the schools. Tug ZV-7. The statistics on rural children in IV-7 were computed from data gathered during 1970-1971 in the Udor region of the Komi Autonomous Soviet Socialist Republic (Martirosian, 1973). From two tripartite analyses of 3 13 measures taken in 1971 at Petropavlovsk, Nekisheva (1974) reported subgroup means of 96.5 cm on 71 children of high income families, and 96.8 cm on 92 children reared under superior living conditions. Tug ZV-8. Items 3 and 4 in IV-8 were derived from measures made between 1960 and 1965 at “villages in all directions around Poona”; more than 80% of the children lived in “poor homes” (Phadke & Kulkarni, 1971). Sources for the “urban privileged” values were the same as in IV-1; their sample sizes arid means at age 5 years were 46 and 1C6.7 cm (Delhi); about 200 and 107.1 cm (Poona); about 34 and 107.1 cm (Bombay). Tug ZV-9. Same sources as IV-4. Tug ZV-10. Sources were the same as IV-3; sample sizes and means for urban privileged children in IV-10 werc 11 and 109.7 cm (Lagos); 185 and 110.2 cm (Ibadan). Tug ZV-I I. Same sources as IV-2; sample sizes and means for rural children in IV-11 were 78 and 95.0 cm (Berry, 1962a); about 90 and 97.7 cm (Migasena et ul., 1974). Tug ZV-12. The rural values in IV-12 were secured from records accumulated during 1963-1968 at villages in 48 Spanish provinces (Palacios & Vivanco, 1965; Palacios-Mateos, Garcia-Almansa, Vivanco, Fernandez, Garcia-Robles, & Moreno-Esteban, 1972), and the urban privileged values from measures taken in 1968 on “healthy and well-nourished” girls attending two schools in Madrid (Garcia-Almansa, Fernandez-Fernandez, & Palacios-Mateos, 1969). To obtain an

Standing Height of Young Children

23

estimated mean for Madrid children of both sexes, the reported mean on girls was raised .5 cm. Tug IV-13. Eksmyr (1971) reported findings in IV-13 from data collected during 1965-1967 at Ijaji, a village in a typical agricultural highland area of Ethiopia, and at two private schools in Addis Ababa. Tug 1V-14. The rural statistics in IV- 14 were obtained from measures gathered during 1965-1966 in an agricultural region surrounding the village of Lawrence Tavern, about 24 km from Kingston (Desai. Miall, & Standard, 1969; Standard, Desai, & Miall, 1969). The corresponding urban privileged values were computed from records collected in 1963 at private schools “in the wealthier suburbs of Kingston” (Ashcroft & Lovell, 1964). Tug IV-15. See IV-6. Tug 1V-16. Refer to IV-7. Statistics in 1V-16 were taken from a sample of 554 children age 5 years, subgroup means were 109.5 cm for 122 subjects classified as offspring of high income families, and 109.4 cm for 186 subjects judged to be growing under superior environmental conditions. Table IV yields the following comparative findings for the period 1960-1970. 1. Children living in rural areas were shorter i n average standing height than ethnically comparable urban children of the upper and upper-middle socioeconomic classes. 2. On the average, Indian young children residing in the rural areas sampled were shorter than their age peers of the urban privileged groups studied by more than 10 cm. 3. Nigerian and Thai young chi!dren of the upper classes were taller, on the average, than their respective age peers inhabiting rural villages by at least 7 cm. 4. Thai and South African Bantu young children of urban privileged groups were no taller than Russian rural children. Compare IV-2 and IV-8 with 111-8 and IV-7; also compare IV-11 and IV- 15 with 111-16 and IV-16. B.

RURAL GROUPS COMPARED WITH URBAN UNDERPRIVILEGED GROUPS

Proceeding to Table V, the first requisite is itemization of the complementary information basic to understanding its content. Since the rural groups in Table V were discussed in explaining Table IV, only the “urban underprivileged” samples require specification. Tugs V-I and V - 5 . The composite mean of 83.7 cm at age 3 years was derived for V-1 from sample sizes and means of about 150 and 82.6 cm for children living in slum areas of Bombay (Udani, 1963); 563 and 83.4 cm for children of the lower classes measured in 1966 at Vellore (Pereira, 1971); and 206 and 85.4 cm for children of Delhi from families in the unskilled occupational category (Banik, Krishna, Mane, & Raj, 1970a). Component values used in

24

Howard V. Meredith

TABLE V Mean Standing Height (cm) of Rural and Urban Underprivileged Groups of Young Children Measured Largely Between 1960 and 1970 Urban underprivileged

Rural

Tag

Group

v- I v-2 V-3 v-4

Indian Russian Nigerian Turkish

v-5 V-6 v-1 V-8

Indian Nigerian Russian Turkish

Sample size

Mean

Sample size

Mean

Difference'

Age 3 years

2,085 I17 ca.

315

21

80.1

93.2 87.1 86.9

ca. 915 77 I22 I06

83.1 95. I

89.0 84.7

3.6 I .9 I .3 -2.2

ca. 200 265 I43 I22

98.6 101.8 107.8 91.2

2.9 .6 -. 4 -.5

Age 5 years

ca. ca.

300 215 192 19

95.7 101.2 108.2 97.7

"Urban underprivileged mean minus rural mean

securing the weighted mean at age 5 years were 52 and 98.0 cm (Delhi); about 150 and 98.8 cm (Bombay). Tags V-2 and V - 7 . Refer to IV-7. From the two analyses cited earlier on 313 records at age 3 years, subgroup sample sizes and means were 95.1 cm for 77 children reared under poor living conditions, and 96.3 cm for 59 children of low income families. Corresponding values from the 554 measures at age 5 years were 107.8 cm for 143 children living under poor environmental conditions, and 107.1 cm for 1 1 1 children of low income families. Tags V-3 and V-6. Refer to 111-2 and 1-24. At age 3 years, sample sizes and means for subgroups were 33 and 86.6 cm for Lagos children of the lower classes and 89 and 89.9 cm for children inhabiting the poorest section of Ibadan. Listed in the same order, component values at age 5 years were 29, 100.0 cm; and 236, 102.9 cm. Tags V-4 and V - 8 . The statistics in V-4 and V-8 for urban underprivileged children were obtained from measures on progeny of unskilled factory workers living in the Murat district of Istanbul (Neyzi, Tanman, & Saner, 1965). Table V shows that during the 1960s: 1. Average standing height was less for Indian young children living in rural regions around Hyderabad and Poona than for Indian young children of low socioeconomic status living at Bombay and Delhi.

Standing Height cd Young Children

25

2. Nigerian young children of low socioeconomic status residing at Ibadan and Lagos were no shorter than their rural age peers residing in the Pankshin area of northern Nigeria and the village of Imesi in western Nigeria.

VI. Intracity and Intercity Comparisons The present section deals with two types of intracity variables, ethnic and socioeconomic, and with variations between city populations in several parts of the world. A.

INTRACITY ETHNIC GROUPS COMPARED

Table VI pertains to seven cities in each of which, during the 1960s, at least two ethnic groups of young children were studied. Particulars on the groups paired in this table are as follows. Tugs VZ-l and VI-2. Data for the analyses in VI-I and VI-2 were collected during 1968-1970. The source of the South African Bantu statistics was given in 111-3; the companion values were obtained by Richardson ( 1 973) from measures on white nursery school children residing “on the East Rand, between 10 and 30 km east of Johannesburg,” including children from “low, middle, and upper” socioeconomic levels. Tags VI-3 through Vl-6. The means in VI-3 through VI-6, all computed from records accumulated between 1961 and 1963, were reported by Goldfeld et al. (1965). Refer to 1-26, 1-33, and 1-22, Tug VI-7. The measures on both of these groups in VI-7 were gathered during 1966-1967 at preschools in Honolulu. Tag 1-31 cited the reference for the Hawaiian Japanese values. The “Hawaiian” sample was described socioeconomically as “from low and middle income families,” and ethnically as about 43% part-Hawaiian, 19% Japanese, 6% white, and 32% Chinese, Filipino, Korean, Puerto Rican, Samoan, or non-Hawaiian mixtures (Smith & Brown, 1970a). Tags Vl-8 and Vl-9. Both VI-8 and VI-9 were based on data amassed between 1962 and 1972 on offspring of parents residing in the San Francisco East Bay area and participating in the Kaiser Foundation Health Plan. Neither sample included children toward the extremes of the socioeconomic continuum, that is, there were no subjects from affluent homes or from homes in the unskilled and indigent categories. The white sample was limited to single-born children (Wingerd, personal communication 1974; Wingerd & Schoen, 1974), and the other sample was composed of children weighing about 2.5 kg at birth, 60% white and 40% black (Beck & van der Berg, 1975).

26

Howard V . Meredith

TABLE V1 Mean Standing Height (cm) on Paired Ethnic Groups of Young Children Living in Specified Cities Dunng 1960-1970

Tag

Age (years)

Mean

Sample size

Johannesburg: Bantu, South African white 91.8 69 265 104.3 133 208

v1- 1 VI-2

5

VI-3

5

V1-4

5

219

V1-5 V1-6

3 5

Cherboksary: Chuvash, Russian 90.9 208 2 I6 104.5 259 211

V1-7

3

VI-8 V1-9

3 5

VI-I0

3

Sample size

<

_I

Mean

110.6

4.6 6.3

Baku: Azerbaijani, Russian 323 105.2 282

107.4

2.2

Kazun: Tatar, Russian 105.0 286

107.0

2.0

92.4 105.6

i .5 1.1

Honolulu: Japanese, Hawaiian 60 91 .8 I38

93.0

1.2

San Francisco: white, black and white 94.5 235 3,057 3,707 109.2 263

94.9 109.8

.6

106.7

.o

404

Frunze: Kirghiz, Russian 106.7 267

96.4

Differencea

.4

“The mean in column 6 minus the mean in column 4.

Tug 1/1-10. The pair of means in VI-10 was obtained from measures taken during 1961-1963. Refer to 1-39 for particulars regarding the Kirghiz statistics. The Russian statistics were reported by Goldfeld et ul. (1965). The intracity ethnic groups juxtaposed in Table VI lead to the following findings: 1 . Of the differences in average standing height displayed, the largest are those between Johannesburg Bantu and white children. One-tailed significance tests at p = .01 allow the inference: In the late 1960s, the Johannesburg white population was taller than its Bantu population by at least 2.5 cm at age 3 years, and 4.5 cm at age 5 years. 2 . In the early 1960s, average height of Russian children age 5 years living at Baku was greater, at 99:1 probability, than that of the Azerbaijani age peers

Standing Height of Young Children

27

residing in this Asian city by an amount likely between .9 and 3.5 cm. There was a similar relation between Russian and Tatar children living at Kazan. 3. Samples of young children studied in the 1960s at Cherboksary, Honolulu, and San Francisco do not warrant acceptance of the hypothesis that in any one of these towns the ethnic groups compared differed in average standing height. Possibly larger samples would have indicated dependable differences in one or more instances, particularly in average height of Russian and Chuvash children residing at Cherboksary. B.

INTRACITY SOCIOECONOMIC GROUPS COMPARED

Intracity statistics on samples of young children drawn to represent one or more socioeconomic categories are presented in Tables VII and VIII. Table VII assembles means for standing height that allow comparisons among lower, middle, and upper socioeconomic groups i n four cities, and comparisons between lower and upper socioeconomic groups in eight cities. Assembled in Table VIII are means on groups of children studied in 19 cities, the means for a specific city relating to one of five segments of the socioeconomic continuum. Notations on the sources used in constructing these tables are as follows: Tags Vll-I and Vll-2. The values for Bombay children of low socioeconomic status in VII-1 and VII-2 were obtained by pooling samples studied in the early 1960s (Udani, 1963), and late 1960s (Shiddhaye et al., 1972). The Udani study supplied the values for Bombay children of middle class and well-to-do families. Tags Vll-3 and Vll-4. Refer to IV-4 and V-4. Tags VII-5 and Vll-6. See 111-2. Tags Vll-7 and Vll-8. The statistics in VII-7 and VII-8 were computed from measures taken during 1964-1965 (Barja, LeFuente, Ballester, Monckeberg, & Donoso, 1965). Tags Vll-9 and Vll-10. Analyses in VII-9 and VII-10 for all three socioeconomic groups were available from one study (Banik et al., 1970a), analyses for middle class children from another (Ghai & Sandhu, 1968), and analyses for “well privileged” children from a third (Banik et al., 1972). It follows that the Delhi values in Table VII for middle class children are aggregate sample sizes and weighted means derived by combining two samples, as are the values for children of the upper classes. Tags VII-11 and Vll-12. The measures used in obtaining the means in VII-11 and VII-12 were collected during 1965-1968 on children belonging to the “lowest” and “highest” socioeconomic classes of Bogod (Luna-Jaspe, Ariza, Rueda-Williamson, Mora, & Pardo, 1970 Rueda-Williamson, Luna-Jaspe, Ariza, Pardo & Mora, 1969). Tags VII-13 and Vll-14. Refer to 1-24.

Howard V . Meredith

28

TABLE V11 Mean Standing Height (cm) of Lower, Middle, and Upper Socioeconomic Groups of Young Children Measured Between 1960 and 1970 in Seven Cities ~

Lower

Middle

Age (years)

Sample size

VII-I Vll-2

3 5

ca. 235 ca. 200

83.0 99.0

VII-3 Vll-4

3 5

106

I22

84.7 97.2

Vll-5 VII-6

3 5

33 29

86.6 100.0

Lagos, Nigeria 27 89.2 30 102.0

Vll-7 Vll-8

3

5

ca. 120 ca. 120

85.6 99.8

Santiago, Chile ca. 410 90.9 ca. 410 104.5

Vll-9 VII- I0

3 5

206 52

85.4 98.0

Delhi, India 328 87.3 110 101.6

VII- 1 I VII-12

3 5

66 66

v11- 13 VI1- 14

3 5

89 236

89.9 102.0

VII- 15 VII- 16

3 5

45 30

89.8 103.2

Tag

Mean

Sample size

Mean

Bombav. Indiu ca. 90 91.2 ca. 90 107.7

Upper Sample size

Mean

Differencea

ca. ca.

34 34

94.0 109.7

11.0 10.7

ca. ca.

70 70

93.3 107.8

8.6 10.6

19 II

95.5 109.7

8.9 9.7

35 35

94.5 109.0

8.9 9.2

81 55

94.2 106.7

8.8 8.7

62 64

94.4 107.5

7.0 8.0

114 I85

96.6 110.2

6.7 8.2

87 85

92.3 105.7

2.5 2.5

Istanbul, Turkey

Bogota, Colombia 87.4 99.5

ca. ca.

Ibadan, Nigeria

Baghdad, Iraq

‘Mean for upper socioeconomic subgroup minus mean for lower socioeconomic subgroup.

Standing Height of Young Children

29

TABLE VIII Mean Standing Height (cm) of Contemporary Groups of Young Children Each Representing an lntracity Socioeconomic Subgroup ~

Age 3 years

Age 5 years

Sample size

Mean

Sample size

VIII- 1 VIII-2 VIll-3 VIII-4 VIII-5

Lower classes Indian, Vellore 563 Peruvian, Lima 39 Dominican, Santo Doming0 ca. 90 United States black, New Orleans 1 I8 United States black, Washington, D.C. 145

83.4 86. I 91.6 93.7 93.9

-

-

50 ca. 90 186 I29

99.3 105.9 107.9 108.6

VIII-6 VIII-7 VIII-8

Indian, three towns Chinese, Hong Kong Chilean, several towns

Lower and middle classes 90 208 80

83.2 90.3 90.9

98 I26 694

97.5 104.7 105.0

Vlll-9 VIII- I0 VI1I-l I

Chinese, Tainan French, Paris Ugandan, Kampala

90

91.8

-

-

91 275 379

104. I 107.0 107.6

v111- 12 VIII-13 V111- 14 VIII- 15 V111- 16 VII1- 17

Thai, Bangkok Armenian, Beirut Indian, Poona Chilean, Santiago Indian, Kerala Sara, Fort Archambault

243 I79 ca. 200 141 ca. 60 63

104.6 108.6 107.1 107.6 109.0 109.0

V111- 18 VIII-19

Ethiopian, Addis Ababa Chinese, Kingston

104.1 105.9

Tag

Ethnic group and city

Mean

Middle classes

Middle and upper classes 156 125 ca. 200

-

93.2 94.5 95.0

-

-

ca. 60

95.2

-

-

Upper classes

-

-

81

-

-

21

Tags VZI-15 and V11-16. These children were born in Baghdad; they were “mainly of Arab ancestry, with variable Kurdish, Turkish, and Persian contributions;’’ one socioeconomic class resided in “slum areas” and the other in “fashionable residential areas” (Shakir & Zaini, 1974). Tag VZIZ-Z. The mean in VIII- 1 was computed from data gathered at Vellore in 1966 on children “belonging to the lower socioeconomic groups” (Pereira, 1971).

30

Howard V . Meredith

Tug Vlll-2. Records for the study in VIII-2 were accumulated during 19661970 on children from “extremely poor homes, primarily mestizo families living in the peripheral slums of Lima” (Blanca-Adrianzen, Baertl, & Graham, 1973). The children were “in apparent good health and had no history of severe malnutrition. ” Tug Vlll-3. Hernandez (1966) reported the values in VIII-3 on children “belonging to the underprivileged class” of “la zona urbana y periferica de Santo Domingo. ” Tug VIII-4. The means in VIII-4 were secured from measures made during 1963-1965 on children of indigent families residing in New Orleans (Cherry, 1968). Tug VM-5. During 1963-1965, the data analyzed in VIII-5 were collected on children of “low income families” living in the District of Columbia (Verghese, Scott, Teixeira, & Ferguson, 1969). Tug Vlll-6. The values in VIII-6 are aggregate sample sizes and weighted means derived from combining records on children of “middle and low socioeconomic strata” measured during 1967-1968 at Burla (Mohanta et al., 1972), and children of “lower-middle socioeconomic status” measured during 1967-1968 at Bhopal and Gwalior (Sharma, 1970). The Sharma subjects showed “no signs of malnutrition or vitamin deficiency.” Tug VIU-7. Low (1971) reported the statistics in VIII-7 using records amassed between 1963 and 1967 on children from families of Chinese lineage and low to middle socioeconomic status living “in the urban areas of Hong Kong Island.” Tug Vlll-8. In 1960, measures for the study in VIII-8 were taken at several cities in Chile (Berry, 1961); the samples included “all ethnic groups except the Araucanian Indians ,” and were considered ‘‘representative of the urban population in the middle and lower income brackets.” Tug Vlll-9. During 1963-1964 the measures analyzed in this row were taken at Tainan, Taiwan, on children of Chinese middle class families (Kimura & Tsai, 1967). Tug VIII-10. The mean in VIII-I0 was obtained from data accumulated between 1958 and 1964 on physically normal children residing in a middle class suburb of Paris (Sempe, Sempe, & Pedron, 1972). Tug Vlll-11. Rutishauser (1965) secured the mean in VIII- 11 from data collected during 1962-1965 on ‘‘healthy middle class Baganda children” attending nursery schools in Kampala. (Included in the Rutishauser report was a mean of 91.1 cm at age 3 years; this mean was obtained from measures taken during 1962-1965 on 114 Baganda rural children living in areas around the child health center at Namulonge, 26 krn from Kampala.) Tug Vlll-12. Refer to IV-2 and IV-11. Tug Vlll-13. Refer to 1-38.

Standing Height oj’ Young Children

31

Tug Vlll-14. The means in VIII-14 were secured from data gathered during 1970-1972 on healthy children of “above average socioeconomic status” showing “no clinical signs of malnutrition”; see 1V-1 and IV-8. Tug Vlll-1.5. The measures for the study in VII- 15 were taken at Santiago in 1963 on kindergarten children of “upper middle” socioeconomic status (Montoya & Ipinza, 1964). Tug Vlll-16. In VIII- 16 are sample sizes and means on children from families of “high and middle” socioeconomic status living in the Cochin-EmakulamAlwaye zone of Kerala; the children were measured at a health guidance clinic (Gokulanathan & Verghese, 1969). Tug Vlll-17 and Vlll-18. Refer to 1-41 and IV-13. Tug VIU-19. The measures used in computing the mean in VIII- 19 were taken in 1963 on children of Chinese ancestry attending private schools “in the wealthier suburbs of Kingston”; many of the grandparents were born in southern China (Ashcroft & Lovell, 1964). The contents of Tables VII and VILI, conjoined with findings from other sources, indicate: 1. With respect to young children studied between 1960 and 1970 at Bombay, Delhi, Lagos, and Santiago, there was a positive relation between standing height and socioeconomic status (Table VII). Taking the four cities together, average standing height of children from poor families was less than that of children from middle class families by about 4.5 cm, and average height for children of middle class families less than that for children of well-to-do families by about 5.0 cm. 2. Children of high socioeconomic status measured during the 1960s at Bogota, Bombay, Delhi, Ibadan, Istanbul, Lagos, and Santiago were taller than their city neighbors of low socioeconomic status; at ages from 3 to 5 years, the overall difference approximated 9.5 cm (Table VII). Except at Baghdad, the differences in the last column of Table VII are at least twice those obtained by several earlier investigators in Great Britain and North America. Data at ages 3 and 7 years were collected in 1950 at Newcastle-upon-Tyne, England (Miller, Court, Walton, & Knox, 1960) and at Eugene, Oregon (Meredith, 1951). Both of these studies compared children whose fathers were in the “professional and major managerial” category with children whose fathers were “semiskilled or unskilled“ differences of 2.6 and 2.2 cm were secured. Menzies (1940), Hopkins (1947), and Weisman (1935) analyzed measures at ages between 5 and 7 years on children attending schools in the “better” and “poorer” districts of Ottawa, Ontario; London, England; and Minneapolis, Minnesota: the differences reported were 1.7, 2.5, and 3.3 cm, respectively. 3. Specific for the period 1960-1970 and for ages between 3 and 5 years, Turkish children of wealthy families residing in Istanbul were no taller in average

32

Howard V . Meredith

standing height than American black children of indigent families residing in New Orleans (VII-3, VII-4, VIII-4). Indian children of the “professional and major managerial strata” living at Delhi under “well-privileged” conditions were no taller than American black children living in “poverty areas” of several states or in “low income families” at Washington, D.C. (VII-9, VII- 10, VIII-5). Dominican children of “the underprivileged class’ ’ in Santo Doming0 were no shorter than Chilean middle class children in Santiago (VII-7, VII-8, VIII-3). Ethiopian and Chinese children attending private schools in Addis Ababa and Kingston, respectively, were shorter than Russian children of low income families living under poor environmental conditions at Petropavlovsk (IV- 13, v1n- 18, v-7). 4. During the 1960s, average standing height at age 3 years on children of low socioeconomic status was near 83 cm in Bombay and Vellore, near 87 cm in Bogod and Lagos, near 90 cm in Ibadan, and near 94 cm in Washington, D.C. (Tables VII and VIII). A mean near 80 cm was cited earlier on Indian rural children (In-l), and a mean roughly 10 cm higher than this was reported as “representative” of the Puerto Rican rural population (Fernandez, Burgos, Asenjo, & Rosa, 1969). The Puerto Rican report, based on data gathered during 1963-1965 at isolated communities, gave a mean of 89.6 cm, and evaluated the rural diet as deficient in calories, calcium, vitamin A, riboflavin, and iron. Several analyses of measures taken in the United States during 1968-1972 yielded means between 93 and 94 cm for children of low socioeconomic status. On “poverty area children” measured in health projects of the United States Children’s Bureau, sample sizes and means were 261 and 93.0 cm for white subjects and 517 and 93.6 cm for black subjects (Systems Development Project Staff, 1968). On children of “lower income families” measured in eight states-California, Kentucky, Massachusetts, Michigan, New York, South Carolina, Washington, and West Virginia-sample sizes and means were 320 and 93.1 cm for white subjects and 245 and 94.2 cm for black subjects (United States Center for Disease Control, 1972). On Mexican children of migrant farm families in Colorado (Chase, Kumar, Dodds, Sauberlich, Hunter, Burton, & Spalding, 1971) and “low income families” in Texas (United States Center for Disease Control, 1972); sample sizes and means were 63 and 93.1 cm (Colorado); 47 and 93.7 cm (Texas). A mean of 94.0 cm was obtained from a sample (TI = 95) considered to represent United States black and white children in families “below the poverty level” (Abraham er al., 1975). For children of low socioeconomic status age 5 years, averages were between 97 and 100 crn at Bogod, Bombay, Delhi, Istanbul, Lagos, Lima, and Santiago (Tables VII and VIII), and above 107 cm from studies on black children at New Orleans (VIII-4) and Washington, D.C. (VIII-5) and the studies on black and white children cited above. Specific sample sizes and means on “poverty area,” “lower income,”

33

Standing Height of Young Children

and “poverty level” children, listed in the same order as above, were 280, 107.1 cm; 538, 109.3 cm; 442, 107.2 cm; 329, 108.7 cm; and 88, 107.9 cm. C.

INTERCITY COMPARISONS

Table IX brings together statistics for standing height of contemporary groups of young children residing in 25 cities. The table is illustrative, not exhaustive; as will be disclosed later, 50 or more city populations were sampled between 1960 and 1970. Information complementing the successive rows of Table IX follows: Tug I X - I . The small samples of Iranian children in IX- 1 were measured at the capital of Fars province (Forbes, personal communication 1972; Forbes, Ronaghy, & Majd, 1971). TABLE IX Mean Standing Height (cm) of Young Children Measured During 1960-1970 at 25 Cities Ape 3 years

Tag

City and country

IX- I IX-2 LX-3 1X-4 IX-5 IX-6 IX-7 IX-8 IX-9 IX-I0 1x-11 IX-12 IX-13 IX-14 1X-15 1X-16 1X-17 IX-18 IX-19 IX-20 IX-21 1x-22 IX-23 1X-24 1X-25

Shiraz, Iran Delhi, India Lagos, Nigeria Santiago, Chile Naples, Italy Astrakhan, U.S. S . R. Yaroslavl, U.S.S.R. Brussels, Belgium Rome, Italy Sofia, Bulgaria London, England Courtrai, Belgium St. Louis, U.S.A. Budapest, Hungary San Francisco, U.S.A. Moscow, U .S.S.R. Murmansk, U.S.S .R. Riga, Latvian S.S.R. Sydney, Australia Warsaw, Poland Petrozavodsk, U.S.S.R. Amsterdam, Netherlands Lvov, Ukrainian S.S.R. Brno, Czechoslovakia Prague, Czechoslovakia

Sample size

ca. ca.

Age 5 years

Mean

31 504 79 565 240 332 683

85.5 86.6 89.4 90.0 91.8 92.0 93. I

-

-

-

-

498 160

93.2 93.6

-

-

205 414 3,057 818 272 200 1,890 ca. 600 I90 I08

94. I 94.2 94.5 94.5 94.6 94.7 95.0 95.4 95.8 96. I

-

-

469 434

96.5 97.4

ca.

Sample size

ca. ca.

ca.

40 124 70 565 230 229 397 546 600 714 160

1,221 205 520 3,707

Mean

98.1 101.6 102.4 103.8 105.3 105.6 105.9 107.4 107.8 108.1 107.8 108.9 109.0 109.2 109.2

-

-

255 200 2,919 ca. 280 252 92 ca. 280 390 219

108.0 109.8 109.3 108.9 108.1

ca.

110.6

110.8 111.3 110.7

34

Howard V . Meredith

Tug IX-2. The values in IX-2 were taken from the 1970 report cited in VII-9 and VII- 10. Another report, based essentially on the same data, gave means of 86.6 cm on 541 children age 3 years, and 100.9 cm on 135 children age 5 years (Banik, Krishna, Mane, & Raj, 1970b). Tug IX-3. See 111-2. Tug 1 x 4 . Refer to VII-7 and VII-8. Tug I X - 5 . IX-5 presents statistics on children of Naples measured in 1963 (Tatafiore, 1965). Tugs IX-6 u n d I X - 7 . The means in IX-6 and IX-7 were reported by Goldfeld et al. (1965); the records used were accumulated during 1960-1962 at the two European cities specified. Tug 1x4. Twiesselmann ( 1969) collected the measures for the analysis in IX-8 during 1960-1961. Tug IX-9. These values i n IX-9 were obtained from data gathered during 1966-1967 in the schools of Rome (Gabrielli, Lucchetta, Chiurazzi, DiGiorgi, & Cuni, 1967). Tug I X - 1 0 . See 11- 10. Tug I X - 1 1 . The records analyzed in IX- 1 1 were made between 1960 and 1965 on children from families living in the central region of London (Tanner, Whitehouse, & Takaishi, 1966). From measures amassed in 1959 at representative London schools, Scott (1961) secured a mean of 108.3 cm on 1120 children age 5 years. Tug IX-12. The mean in IX-12, computed from data collected during 19671968, was reported by Franckx (1%9). Tug IX-13. The sample sizes and means in IX- 13 pertain to children born and reared at St. Louis, Missouri, and measured between 1967 and 1972. Ethnic composition of the samples was 64% white and 36% black children at age 3 years (Jordan & Spaner, 1972) and 60% white and 40% black children at age 5 years (Jordan & Spaner, 1974). Tug IX-14. Data collection for the study in IX-14 took place in 1968-1969 (Eiben, Hegedus, Banhegyi, Kiss, Monda, & Tasnadi, 1971). Data at age 5 years were collected simultaneously at Kormend, yielding a mean of 109.4 cm on a sample of 55 children (Eiben, personal communication 1969). Tug IX-15.See VI-8. Tug I X - 1 6 . IX-16 is a composite mean derived by pooling records on 355 Moscow children measured in 1962 (Goldfeld et ul., 1965) and on 463 Moscow children measured during 1963-1964 (Kogan, 1969). Tug I X - 1 7 . The statistics in IX- 17 were secured from data collected in 1969 at age 3 years, and in 1964 at age 5 years (Lapitskii, Belogorskii, Nemzer, Pogorely, Sinopalnikov, & Zolotareva, 1970; Lapitskii, Belogorskii, Nemzer, & Zolotareva, 1967).

Standing Height of’Young Children

35

Tug IX-18. Goldfeld ef ul. (1965) reported the statistics in IX-18 on native Latvian children measured during 1962-1963. Tug IX-19. The means in IX-19 were obtained on large samples of children measured between 1970 and 1972 at child health centers, preschools, and kindergartens in the “Sydney metropolitan area” (Jones & Hemphill, 1974; Jones, Hemphill, & Meyers, 1973). Tug IX-20. Refer to 111-4 and 111- 14. Tug IX-21. This analysis in IX-21 was made from measures taken in 1966 (Kuz’min, 1970). Tug IX-22. The samples analyzed in IX-22 were measured in 1963 (Oppers, 1964). Tug IX-23. The mean in IX-23, based on data collected during 1970-1971, was reported by Zhaglina ( 1974). Tugs ZX-24 and IX-25. See 111-5. The statistics in Table IX, colligated with findings on the standing height of young children in other cities, warrant the following statements: 1. At age 3 years, the mean in IX-2 for Indian children of Delhi is 10.8 cm, or 1 I%, lower than that for children of Prague. Generalizing at 99: 1 probability: During the 1960s, the Indian population of children age 3 years living at Delhi was shorter in average standing height than the Czech population of age peers living at Prague by an amount likely between 9.8 and 11.8 cm. Compared with the “all India” mean of 84.5 cm at age 3 years, obtained from 5700 measures gathered in several Indian states during 1956-1965 (Indian Council of Medical Research, 1968), the Delhi mean i n IX-2 is 2.1 cm higher. 2. Average standing height at age 3 years for city children of the 1960s was near 90 cm at Santiago, Chile, and Lagos. Nigeria; near 92 cm at Naples, Italy, and Astrakhan, Soviet Union; near 94 cm at St. Louis, Missouri, and Budapest, Hungary; and near 95 cm at Sydney, New South Wales, and Warsaw, Poland. Means between 92 and 94 cm were obtained at several cities in the Soviet Union. Specific sample sizes, means, times of data collection, and places are: 267, 92.2 cm, 1962, Saratov; 512, 92.8 cm, 1960-1961, Minsk; 185, 92.9, 1962, Pskov; 261, 93.6 cm, 1960, Rostov-on-Don; 161, 94.0 cm, 1962, Kursk (Goldfeld et ul., 1965); about 400,92.7 cm, 1962-1963, Kemerovo (Kaganovich, Kraeba, & Bogachanov, 1966); and about 170, 93.8 cm, 1963, Ivano-Frankovsk (Kutsenko, 1965). Means above 95 cm, in addition to the five in Table IX, are 95.6 cm on 51 West German children measured at Kiel (Spranger, Ochsenfarth, Kock, & Henke, 1968), and 97.6 cm on 129 Czechoslovak children measured at Bratislava (111-5). 3. At age 5 years, the means in Table IX are higher for Lvov than for Santiago by 7 cm, higher for Brno than for Naples by 6 cm, higher for Amsterdam than for Astrakhan by 5 cm, and higher for Riga than for Yaroslavl by almost 4 cm. It is

Howard V . Meredith

36

highly probable that during the 1960s the population of children age 5 years at Santiago, Chile was at least 6 cm shorter in average standing height than the population at this age in Brno, Czechoslovakia. 4. Means between 106 and 108 cm were obtained from investigations in the 1960s at Brussels (IX-8), Kursk (111-16), Lublin (111-14), Murmansk (IX-l7), Ore1 (III-l6), Rome (IX-9), Ryazan (111-16), and several additional cities in the Soviet Union. Particulars, following the same procedure as at age 3 years, are: 332, 106.0 cm, 1960, Tomsk; 292, 106.3 cm, 1961, Orenberg; 217, 106.6 cm, 1962, Pskov; 638, 107.2 cm., 1962, Blagovyeschensk-on-Amur; and 201, 108.0 cm, 1962, Saratov (Goldfeld et al., 1965). Means of 105.7 and 107.4 cm, respectively, were secured on 178 Italian children measured during 1960-1961 at Bari (Pirk & Meli, 1962), and on 124 Italian children measured during 19621964 at Bolzano (Dattoli, 1965). 5. Continuing for age 5 years, means between 108 and I10 cm characterized children at Budapest (IX-l4), Courtrai (IX-l2), Donetsk (111-16), Dzhalal-Abad (I-38), Kormend (IX- 14), Petrozavodsk (IX-21), Riga (IX- 18), San Francisco (IX-l5), Sofia (IX-lo), St. Louis (IX-l3), Sydney (IX-l9), and Warsaw (IX-20). Materials from five other studies yielding means in this category are: about 500, 108.0 cm, 1960-1962, Vienna (Stracker, 1963); 502, 108.3 cm, 1961, Kalinin (Goldfeld et af., 1965); about 170, 108.3 cm, 1963, Ivano-Frankovsk (Kutsenko, 1965); about 225, 108.5 cm, 1969, Gorki (Dorozhnova, 1973); and 554, 108.6 cm, 1971, Petropavlovsk (Nekisheva, 1974).

VII. Female and Male Groups Compared The present section is based entirely on averages for standing height computed from measures on 200 or more children of a particular age and sex. A.

GROUP DESCRIPTIONS

Table X displays paired means for males and females of 16 populations sampled at age 3 years, and 18 populations sampled at age 5 years. Source identifications are as follows: Tags X - 1 and X - 2 1 . See 1-29. Tags X-2 a n d X - 2 2 . The values in X-2 and X-22 were obtained on “poverty area children” measured during 1968 in connection with health projects of the United States Children’s Bureau (Systems Development Project Staff, 1968). Tags X - 3 a n d X - 1 7 . See 1-3. Tags X - 4 and X - 2 3 . Studies made at Minsk and Yaroslavl provided data for the analyses in X-4 at age 3 years (Goldfeld et al., 1965).,The measures used

37

Standing Height of Young Children

TABLE X Mean Standing Height (cm) of Male and Female Young Children Measured Largely Between 1960 and 1970 ~~

Males

Tag

Group

Sample size

x-1

X-16

Surinam Creole United States black Angolan Lunda Russian Romanian Indian Filipino Dutch Spanish Bulgarian United States white Tasmanian Czech Australian Chinese Finnish

414 263 74 3 598 > 1,980 834 217 627 776 677 1,576 560 460 96 I 650 245

X-17 X- 18 X-19 x-20 x-2I x-22 X-23 X-24 X-25 X-26 X-27 X-28 X-29 X-30 X-31 X-32 x-33 x-34

Angolan Lunda Hungarian East German Filipino Surinam Creole United States black Russian Italian Bulgarian Spanish Japanese Chinese Dutch United States white Belgian Australian Surinam Industani Finnish

799 276 953 236 757 279 1,298 808 1,002 5,941 3,118 1,532 820 1,896 913 1,440 777 208

x-2 x-3 x-4 x-5 X-6 x-7 X-8 x-9

x-10 x-ll x-I2 X-I3 X-14

X-15

Females

Mean

Sample size

Mean

Difference"

Age 3 years

91 .0 93.7 83.6 93.3 93.9 84.3 85.7 97.6 91 -0 93.3 95 .0 96.5 97.6 95.7 93.4 96.0

424 254 800 597 > 1,980 789 216 625 739 699 1,481 554 443 929 538 240

91.4 93.5 83.5 92.8 93.1 83.4 84.8 96.7 89.9 92.2 93.9 95.3 96.2 94.3 91.4 93.7

- .4

84 I 244 895 216 782 259 1,130 833 92 I 6,246 2,896 1,330 804 1,811 854 1,479 739 212

96.4 109.1 109.8 97.3 106.3 109.0 107.5 107.2 105.8 103.5 103.6 104.5 110.8 108.7 107.9 108.7 103.6 107.9

-.I .I .3 .4

.2 .3 .5

.8 .9 .9 .9 1.1 1.1 1.1

I .2 1.4 I .4 2.0 2.3

Age 5 years

"Mean for males minus mean for females.

96.3 109.2 110.1 97.7 106.8 109.5 108.1 107.9 106.6 104.4 104.5 105.4 I 1 1.7 109.6 108.9 109.8 104.8 109.1

.5 .5

.6 .7 .8 .9 .9 .9 .9 .9 1 .0 1.1

1.2 1.2

Howord V . Meredith

38

at age 5 years (X-23) were gathered at Blagovyeschensk-on-Amur, Donetsk, Ka!inin, and Kursk (Goldfeld er ul., 1965; Pokatilo, 1974). Tug X-5. See 11-9. Tug X - 6 . The means in X-6 were obtained from measures taken at Delhi (Banik er ul., 1970c), Vellore (VIII-l), and Maharashtra villages in the vicinity of Poona (Phadke er ul., 1971). Tugs X - 7 and X - 2 0 . See 11- 17. Tags X - 8 and X-29. See 11-14. Tugs X - 9 and X - 2 6 . See IV- 12. Tugs X - 1 0 and X-25. During 1960-1961, records for these analyses were accumulated at Sofia and other urban communities in Bulgaria (11- 10). Tags X-11 a n d X - 3 0 . See VI-8. Tug X-12. See 11-28, Tag X - 1 3 . Measures taken at Brno and Prague were used in computing the weighted means in X-13 (111-5). Tugs X-14 and X-32. See IX- 19. Tags X - 1 5 and X-28. See 11-8. Tugs X - 1 6 andX-34. See 11-12. Tug X - 1 8 . See IX-14. Tug X - 1 9 . See 11-13. Tug X - 2 4 . Data in X-24 were pooled from studies in the department of Emilia (11-25), the province of Grosseto (11-24), and Rome (IX-9). Tug X-27. See 11-23. Tug X - 3 1 . See IX-8 and IX- 12. Tug X - 3 3 . See I- I 9. 13.

GROUP COMPARISONS

The column in Table X farthest to the right shows the amount by which each of the means for males exceeds its adjacent mean for females. It is found: 1. Collectively, the 34 paired samples of young children indicate that during the 1960s males were taller than females by aboat .8 cm. This overall figure, considering the strength of its foundation, can be taken as a reasonably precise estimate of the human sex difference in standing height for the age period from 3 to 5 years. 2. For the 16 differences at age 3 years, the obtained average is 1 .O cm. This value appears spuriously inflated, due principally to the differences of 2.0 and 2.3 cm i n X-15 and X-16, respectively. The Chinese study (X-15) reported means higher for males than for females by 1.3 cm at age 4 years, and .9 cm at age 5 years; in the Finnish study, obtained differences were 1.5 cm at age 2 years, 2.0 cm at age 4 years, and 1.2 cm at age 5 years.

Standing Height of Young Children

39

3 . In several instances, the hypothesis of identical population means for standing height of males and females in early childhood (the null hypothesis) cannot be rejected. Specifically, findings from statistical tests do not allow the inference of a highly dependable population sex difference in height for any of the following groups: East Germar,, Filipino, Hungarian, Lunda, Russian, Surinam Creole, or United States black. From samples of Lunda children as large as those in X-3, a difference between obtained means of at least .9 cm is a prerequisite to positing, at p = .01, differential male and female populations for standing height at age 3 years. 4. The samples of Dutch, Spanish, and United States white children lend consistent support to the generalization that during the 1960s preschool age females of these populations were slightly shorter than coeval males. Obtained differences are clustered between .9 and 1.1 cm for these groups at both ages and, at one age, for Australian, Belgian, Bulgarian, Chinese, Indian, and Japanese groups. Differences between 1.2 and 1.5 cm are shown for the Czechoslovak, Finnish, Surinam Industani, and Tasmanian groups.

VIII. Comparisons from Subgrouping for Other Variables A.

COMPARISONS FROM SUBGROUPING B Y BIRTH WEIGHT A N D BIRTH ORDER

1 . Birth Weight Four studies reported during the 1960s supported the generalization that there is a positive relationship between body weight at birth and standing height i n early childhood. Banik ef al. (1970b), using measures of standing height accumulated between 1964 and 1967 on Indian children of Delhi, obtained means of 85.1 cm at age 3 years and 98.5 cm at age 5 years for those subjects weighting 2.5 kg or less at birth. Means for the subjects weighing more than 2.5 kg at birth were higher by 2.1 and 3.3 cm at ages 3 and 5 years, respectively. Subgroup sample sizes were 155 and 386 at age 3 years and 39 and 100 at age 5 years. Mortison (1969), analyzing data on United States black “poverty area children” measured in 1968, secured means that were lower on children whose birth weights were 2.5 kg or less than on children weighing above 2.5 kg at birth by 1.0, 1.2, and .3 cm at ages 3, 4, and 5 years. The sample sizes were 99 and 375 at age 3 years, 72 and 334 at age 4 years and 67 and 316 at age 5 years. At age 3 years, records for standing height collected during 1961-1967 at Buffalo, New York gave a mean 4.6 cm higher on 93 white children whose birth weights were above 2.5 kg than the mean for 160 white children with birth

Howard V . Meredith

40

weights below 2.5 kg (Cruise, 1973). Means near 91.5 cm at age 3 years, and 104 cm at age 5 years, were reported from data gathered during 1963-1970 at Montreal, Quebec on white children who weighed less than 2.5 kg at birth and were singleton offspring of mothers with normal pregnancy periods (Fitzhardinge & Steven, 1972). These means, based on sample sizes of 96 and about 75, are lower than the means for white children in IX- 15 by 2.9 and 5.2 cm at ages 3 and 5 years, respectively. Body weight distributions for viable neonates of different ethnic populations vary considerably. Birth weights below 2.5 kg were found for 29% of 2695 Indian infants delivered during 1959-1961 at Delhi (Ghosh & Beri, 1962), 12% of 2997 United States black infants born in 1949 at Baltimore (Taback, 1951), and less than 7% of 14,390,000 United States white born during 1957-1%0 (Meredith, 1970).

2 . Birth Order Wingerd (personal communication 1974) classified in three birth rank categories a large number of standing height records at age 5 years on white children measured at San Francisco between 1964 and 1972. The means obtained were 109.5 cm for 1243 first-born children, 109.0 cm for 1783 children of second or third birth orders, and 108.8 cm for 681 children of fourth or higher birth ranks. Earlier studies in the United States, England, and Canada reported a similar negative association between birth rank and height in early childhood (Meredith, 1950). First-born children were taller than children of second to fourth birth orders by .8 cm for Iowa City children age 5 years, 1.5 cm for Middlesbrough children age 6 years, and .8 cm for Toronto children age 6 years. The difference obtained by Wingerd on comparing means at age 5 years for ranks 1 and 4 or more (.7 cm) is the same as that obtained in the Toronto study on comparing the mean for rank 1 with that for ranks 5 and above. The negative relationship of standing height and birth order found at childhood ages does not hold in infancy. It i s well established that at birth body length and birth order are positively related (Meredith, 1950). B.

COMPARISONS FOR SINGLE WITH TWIN BIRTHS, AND SMOKING WITH NONSMOKING MOTHERS

1 . Twin Pregnancy Wilson (1974) analyzed data amassed between 1960 and 1972 on white twins age 3 years “drawn from the entire socioeconomic range found in the metropolitan Louisville (Kentucky) area.” The obtained mean of 92.8 cm on 360 children was 1.7 cm lower than that secured at this age from 3057 white singleton births in San Francisco (IX- 15).

Stuntling Height

c?f’ Young Children

41

2. Maternal Smoking Wingerd and Schoen ( 1974) investigated maternal tobacco smoking during pregnancy in relation to standing height of offspring at age 5 years. They obtained means of 109.4 cm on 2404 children of mothers who did not use tobacco during pregnancy; 109.0 cm on 562 children whose mothers, during pregnancy, smoked from 1 to 14 cigarettes daily; and 108.5 cm on 741 children of mothers who smoked 15 or more cigarettes daily. Using measures of standing height amassed in 1965 on 1 1,500 children 7 years of age residing in different parts of England, Scotland, and Wales, Goldstein (1971) secured a mean 1.4 cm greater for offspring of nonsmoking mothers than offspring of mothers who smoked over 10 cigarettes daily throughout pregnancy. A similar relation between the height of children and maternal smoking during their prenatal development has been found in other studies (Meredith, 1975). C.

COMPARISONS FOR PHYSICAL ABNORMALITY, ELEVATION OF HABITAT, HEALTH STATUS, AND HEALTH CARE

I . Down’sSyndrome On Dutch children with Down’s syndrome (Mongolism) residing in NoordBrabant province, Swaak (1967) reported means of 87.4 cm at age 3 years and 98.4 cm at age 5 years. These means, based on sample sizes of 39 and 40, respectively, are lower than those at corresponding ages from the 1964-1966 Netherlands survey (11-14) by 9.8 and 12.9 cm.

2. Elevation of Habitat Two studies made during the 1960s lend support to the generalization that standing height in early childhood and elevation of abode are inversely related. Kirghiz children were measured during 1961-1964 in a town (Naryn) situated at an elevation near 2.0 km, and at two towns (Frunze and Dzhalal-Abad) whose elevations were near .8 km (1-38). Means for children of Naryn were lower than composite means for children of Frunze and Dzhalal-Abad by 1.9 cm at age 3 years and 2.7 cm at age 5 years. Sample sizes were 103 and 197 for the higher town and 254 and 676 for the towns located at elevations near .8 km. Kambara (1969) analyzed measures taken in 1963 on Japanese children age 5 years residing i n mountainous and coastal areas of Kagawa prefecture (11-23). Obtained means were 102.6 cm for about 360 children inhabiting mountainous areas, and 103.9 cm for about 690 children inhabiting coastal regions. Compare 1-8 and 1-35.

3 . Health Status During the 1960s, several studies were reported on young children of similar health status. When brought together these studies indicate that, with health

42

Howard V . Meredith

status held fairly constant, young children of different ethnic groups varied greatly in average standing height. At age 3 years, means near 85 cm were obtained on Indian children of Bhopal and Gwalior shcwing “no signs of malnutrition or vitamin deficiency” (V111-6), and on Nigerian Angus and Sura children “in a good nutritional state” (1-7); near 90 cm on Indian children of Agra province who were without “any history of illness likely to retard growth” (11-21) and Jamaican children showing “no unequivocal signs of malnutrition” (I- 15); near 93 cm on “well-nourished” Thai children (IV-2) and Italian children of Padua province living ‘‘under reasonably good conditions” (11-24); and near 97 cm on Russian (IV-7) and Yoruba (1-24) children reared under good to superior living conditions. There is a difference of 12 cm between the lowest (84.6 cm) and highest (96.6 cm) averages of this series. At age 5 years, reported means were near 89 cm for “healthy” Malayan children (IV-5), near 95 cm for Quechua Amerind children whose “diet appeared adequate” (I-8), near 99 ern for Peruvian children “in apparent good health” (V111-2), near 101 cm for Maharastra children showing “no signs of malnutrition or vitamin deficiency” ( V I I I ~ I )near , 105 cm for “well-nourished’’ Thai children (IV-I l ) , near 107 cm for “healthy” children of Poona manifesting “no clinical signs of malnutrition” (IV-l), near 109 cm for Sara children of Chad exhibiting “no signs of malnutrition” (I-41), and near 11 1 cm for “healthy and well-nourished” Spanish chiidren (IV- 12). This series of averages yields a range exceeding 20 cm.

4 . Health Care Guzman ct al. (1968) compared Mayan Amerind children reared in Guatemalan villages under the three conditions specified in 1- 1 . Average standing height at age 3 years was found to approximate 79 cm at the village deficient in dietary and health resources, 8 1 cm at the village providing “improved sanitation and medical services,” and 84 cm at the village where “children received a food supplement and their mothers were given advice on nutrition.” Means of 88.0 and 100.1 cm at ages 3 and 5 years, respectively, were reported by Morley el al. (1968) on Yoruba single-born children reared in ,a village that provided “the best medical services possible to any sick child” and “food supplements” to any child showing signs of malnutrition. The number of children measured at each age was more than 200. As indicated in 1-24, these means are 8 and 10 cm lower than the means obtained on Yoruba children of well-to-do Ibadan families. On 32 Indian children age 3 years measured at a child development center in New Delhi, a mean of 96.1 crn was secured (Khurana, Agarwal, Manwani, & Srivastava, 197 1 ) . The subjects were “apparently healthy”; they received health care, including immunizations, and their mothers were given advice on nutrition. Compared with the mean of 88.6 cm obtained by Banik ei al. (1970~)for 124

Standing Height of Young Children

43

Delhi children of the upper-middle and upper socioeconomic classes, this mean is higher by 7.5 cm. The living conditions of children residing at Petropavlovsk were classified by Nekisheva (1974) as poor, moderately satisfactory, and superior. From a total sample of 313 children at age 3 years, the mean (96.8 cm) for standing height of 92 children living under superior conditions was 1.7 cm higher than that for 77 children living under poor conditions. At age 5 years, the sample total was 554, and the mean for height of children living under superior conditions was higher than for children living under poor conditions by 1.6 cm. Assuming the difference of 5 cm in average standing height of the Guatemalan “control” and “nutritional guidance” groups is due entirely to improvement in health care, this difference is far less than the 12 cm difference between Mayan children reared with health guidance and the Russian children of Petropavlovsk reared under superior conditions.

IX. Estimates of Population Variability and Their uses A.

CHOOSING A METHOD OF ESTIMATING POPULATION VARIABILITY IN STANDING HEIGHT

If standing height were measured on all of the young children at a given age in a specified population, and the frequency distribution of the measures on this population were found to be a replica of the Gaussian or normal distribution, the mean and standard deviation of the measures would precisely describe the population in respect to standing height. Rarely does an investigator find it possible to measure the entire membership of a population. Instead, a sample is drawn from the population, each member of the sample is measured, and the measures are analyzed statistically to obtain estimates of population parameters. All of the means presented in earlier sections of this chapter are estimates of population central tendency, and all of the standard deviations that will be included in this section are estimates of population variability. Human biologists study samples for the purpose of gaining knowledge about populations and, on occasion, to construct aids (e.g., growth charts) that pediatricians, school health personnel, and others can use to enhance their services within populations. Measures of standing height taken on large samples of young children selected randomly from a population commonly yield frequency distributions that differ little from the Gaussian model. This assertion can be substantiated by comparing series of percentiles obtained directly from ordered measures of height with corresponding percentiles obtained through use of the standard deviation.

Howard V . Meredith

44

Studies on large samples of Dutch children (11- 14) and Australian white children (IX- 19) analyzed variability in terms of the standard deviation and in terms of percentiles, 3, 10, 25, 75, 90,and 97. Table XI reproduces percentiles from these reports and aligns them with percentiles secured by computing values above and below each age-sex-height mean at points .67, 1.28, and 1.88 sigma. The table shows almost complete agreement between the empirical percentiles (Method A) and the percentiles derived on the assumption that standing height TABLE X1 Percentiles (cm) for Standing Height of Young Children Obtained from Ordered Measures (A) and from Standard Deviations (B) Males

Females Method B

B minus A

Dutch children age 3 years 90.3 .I 89.8 .o 91.7 92.6 95.0 .o 94.0 100.2 .2 99.2 .I 101.6 102.6 -.I 103.8 104.9

89.4 91.7 94. I 99.3 101.7 104.0

- .4

Australian children age 3 years 87.6 .o 85.9 -.I 88.5 90.2 92.8 -.2 91.3 .o 91.5 98.6 101.2 .2 100.2 103.8 .I 103.3

85.5 88.3 91.2 97.4 100.3 103. I

- .4 -.2 -.l

97

87.6 90.3 93.0 98.6 101.0 103.7

3 10 25 75 90 97

102.9 105.8 108.4 114.6 117.5 120.2

Dutch children age 5 years 102.9 .o 101.8 -.I 104.9 105.7 108.6 .2 107.5 114.8 .2 113.9 117.7 .2 116.6 120.5 .3 119.6

102.0 104.8 107.7 113.9 116.8 119.6

3 10 25 75 90 97

100.6 103.5 106.3 113. I 116.2 119.2

Australian children age 5 years 100.6 .o 99.1 .o 102.3 103.5 106.5 .2 105.1 .o 112.1 113.1 I 16.1 -.I 115. I 119.0 -.2 118.2

99.3 102.3 105.3 112. I ,115.1 118.1

Percentile

Method A

3 10 25 75 90 97

90.2 92.6 95.0 100.0 102.5 105.0

3 10

25 75

90

Method B

B minus A

Method A

.o .I .I .I .2

-.I .I -.2

.2

-.I .2

.o .2 .O

- .4

.o .2

.o .o

-.I

Standing Height of Young Children

45

distributions at these ages closely approximate the Gaussian distribution (Method B). Some research reports in the 1960s included percentiles 10, 25, 75, and 90, each rounded to the nearest centimeter, and standard deviations rounded to the nearest millimeter. These statistics were supplied in studies of Italian children at ages 3 and 5 years (11-25), and Belgian children at age 5 years (IX-12). Sixteen percentiles from the Italian study (four on each sex at each age) and eight from the Belgian study were paired with percentiles derived from the standard deviations. Taken to the nearest centimeter, there was agreement in 21 of the 24 comparisons. With two methods suited to the same task, the appropriate course is to choose the one having the greater reliability and mathematical robustness. It follows that in early childhood the standard deviation is the preferred method to use when estimating the variability of a population in standing height, or when obtaining percentiles in the process of constructing graphic or tabular frames of reference for standing height. B . STANDARD DEVIATION ESTIMATES OF POPULATION VARIABILITY

It is a statistical commonplace that large samples yield more dependable estimates of population parameters than small samples. The 111 estimates of population variability in Table XII, with eight exceptions where sample size is between 135 and 188, are based on sample sizes exceeding 200. The sources of the estimates are as follows: Tags XII-I and XII-23. See IX-14. Tags XII-2 and XII-16. See 11- 12. Tags XII-3 and XII-17. See X-4. Tags XII-4 and X I I - 2 2 . See IX-24 and IX-25. Tags XII-5 and X I I - 2 1 . See 11- 14. Tags X U - 6 and XII-29. See 1-22. Tags XII-7 and X I I - 3 2 . See 11-24 (Bussadori study) and 11-25. Tags XII-8 and XII-18. See 1-6. Tags XII-9 and X I I - 3 4 . The population estimates in XII-9 and XII-34 were obtained using measures taken during 1958-1960 at the towns of Katowice, Szczecin, and Warsaw (Miesowicz, 1964). Tags XII-10 and XII-28. See IX- 19. Tags XII-11 and XII-20. See 11-8. Tag XII-12. Owen and Lubin (1973) reported the findings in XII- 12. Refer to

1-40. Tags XII-13 and XII-35. See 1-3. Tags X U - I 4 a n d X I I - 3 7 . The samples used to obtain the values in XII- 14 and XII-37 were considered representative for children of “all India”; data collection

46

Howard V . Meredith

took place between 1956 and 1965 in several states (Indian Council of Medical Research, 1968). Tugs XII-15 and XII-36. See 1-29. Tug XII-19. XII- 19 carries variability values based on measures amassed during 1970 under the auspices of the Japanese Ministry of Education (Kimura, personal communication 1973).

TABLE XI1 Standard Deviations (cm) for Standing Height of Young Children

Females

Males

Sample size

Tag

Group

x11- I x11-2 XI1-3 XII-4 XII-5 XII-6 XII-7 Xll-8 XII-9 XII- 10

XII- 12 XlI-13 X11- I4 XII- 15

Hungarian Finnish Russian Czechoslovak Dutch Chuvash Italian Ladino Polish Australian Chineseh United States white Lundah Indian'' Surinam Creole

177 245 598 460 627 213 301 170 305 961 5180 206 1,042 3,057 A14

XII- 16 XII-17 XI1-18 XII-19 XII-2G XII-21 XII-22 Xll-23 XII-24 XII-25 XI1-26 XII-27 XII-28

Finnish Russian Ladino Japanese' Chinese" Dutch Czechoslovak Hungarian Kirghiz East German Surinam lndustani Azerhiajani Australian

2108 1,298 144 10,:Ilo 1,964 804 327 276 469 1.906 938 I88 I.440

Standard deviation

Sample size

Standard deviation

Both sexes Sample size

Standard deviation

Age 3 years

XII- I 1

3.6 3.9 4.1 3.8 3.9 3.7 4.0 3.6 4.1 4.3 4.1 4.4 5.6 6.6 7.2

237 240 597 443 625 204 267 I46 31 I 929 872 173 1,024 2,956 424

3.1 3. I 3.6 3.9 3.9 4.2 4.0 4.4 4.2 4.7 4.3 4.5 5.4 6.3 6.8

414 485 1,195 903 1,252 417 568 316 616 1,890 1,852 379 2,066 6.013 838

3.7a 3.7 3.9 3.9 3.9 4.0 4.1 4.1 4.2 4.6 4.6 4.6 5.5 6.5 7.0

212 1.130 I 50 10.104 1,616 820 342 244 404 1,791 900 135 1,419

3.9 4. I 4.8 4.5 4.1 4.1 4.8 4.6 4.1 4.9 5.2 4.4 5.0

420 2,428 294 20,414 3,580 1,624 669 520 873 3,697 1,838 323 2.919

4.2 4.4 4.5 4.5 4.7 4.7 4.7 4.1 4.8 4.9 4.9 4.9 5.0

Age 5 vears

4.4 4.1 4.2 4.6 4.6 4.7 4.7 4.1 4.8 4.8 4.2 5.1 4.9

fcontinued)

Standing Herghr of Young Children

47

T A B L E XI1 (continued) Females

Males

Both sexes

-

Tag

Group

XII-29 Xll-30 Xll-31 XI1-32 XII-33 XII-34 XII-35 Xll-36 Xll-37

Chuvash Belgian Bulgarian Italian United States white'" Polish Lunda" Surinam Creole Indian"

Sample size

Standard deviation

22 I 913 1.002 807 2.169

50 5 0 5 3 51

300

5 3

742 757 3,413

hO

56

63 67

Sample size

213 854 92 I 759 2,596 300 790 782 2,940

Standard deviation

4.7 5.1 5.2 5.4 5.5 5.7 6.0 6. I 6.4

Sample size

434 1,767 1,923 1,566 4,765 600 1,532 1,539 6,353

Standard deviation

5.0 5.1 5.3 5.3 5.5 5.6 6.0 6.2 6.6

"For SD=4.0 em, SE of SD is .20 cm where n=200, .14 em where n=400, .09 cm where n=loOO, and .05 em where n=3000. For SD=6.0 e m , corresponding SEs are .30, .21, .13, and .08 em. *Age 3.5 years. 'Age 5 . 5 years. "Age 4.5 years.

Tugs XII-24 through XII-27. See 1-38, 11-13, 1-19, and 1-26, respectively. Tag XII-30. See IX-8 and IX- 12. Tug XII-31. See 11- 10. Tug XII-33. The standard deviations in XII-33 were obtained on data gathered during 1937-1939 in a Federal anthropometric project conducted in 16 states and the District of Columbia (O'Bnen, Girshick, & Hunt, 1941). Inspection and statistical evaluation of Table XI1 yield the following: 1. Estimates of variability i n standing height for different populations of young children are not consistently higher on one sex than the other. Among the 37 pairs of sex specific standard deviations, I7 are higher for males than females, 16 higher for females than males, and 4 are of equal magnitude for both sexes. 2. At age 3 years, 77% of the 30 standard deviations for the sexes separately are between 3.5 and 5.0 cm. For the sexes separately at age 5 years, 80% of the 44 standard deviatibns are between 4.0 and 5.5 cm. 3. Standard deviation estimates for both sexes together (Table XII, cclumn farthest right) are between 3.7 and 4.6 cm on the nine white populations represented at age 3 years. On the 12 white populations at age 5 years, the delimiting values are 4.2 and 5.6 cm. Other populations for whom standard deviation estimates are within these limits are Central American Ladino, South American Industani, European Chuvash, and Asian Azerbaijani, Chinese, Japanese, and Kirghiz.

48

Howard V . Meredith

4. At age 3 years, population standard deviation estimates for standing height are low (3.7-3.9 cm) for Czechoslovak, Dutch, Finnish, and Hungarian children in Europe, and high (5.5-7.0 cm) for Lunda children in Africa, Indian children in Asia, and Creole children in South America. Low (below 4.8 cm) and high (6.0 cm and above) variability in standing height is found for the same populations at age 5 years. 5. For the two sexes together, statistical significance tests a t p = .01 yield no dependable differences at age 3 years in variability of standing height among the eight populations of European children represented. There are dependable differences, at 99: 1 probability, between each of the European populations and the populations of Indian, Lunda, and Creole children. Standing height at age 5 years is more variable in the Indian, Lunda, and Creole populations than in the populations of East Germany, Japan, Republic of China, Soviet Union, and the Netherlands. There are no statistically dependable differences at this age among the Azerbaijani, Chinese, Czechoslovak, Dutch, East German, Hungarian, Japanese, Kirghiz, Ladino, and Surinam Industani populations.

X. Summary This chapter centers on recent advances in knowledge pertaining to the standing height of preschool children. It assembles and orders research findings from more than 200 investigations made in different parts of the world between 1960 and 1970. The ordering is problem oriented; comparisons and syntheses are presented for human standing height at ages 3 and 5 years in relation to ethnic and national origin; urban and rural habitat; socioeconomic and health status; sex, birth size, and birth order; and maternal pregnancy variables. Selected findings on children of the 1960s are as follows: 1. At ages 3 and 5 years, average height was more than 10 cm, or 11%, less for Lunda children of Angola than for United States white children. There were similar differences at age 3 years between Mayan children of Guatemala and Kirghiz children of the Kirghiz Soviet Socialist Republic, and at age 5 years between Shi children of Zaire and United States black children. 2. Average height at age 3 years was near 90 cm for black children of St. Vincent and Jamaica islands, Chuvash children of the Chuvash Socialist Republic, Eskimo children of Alaska, Industani children of Surinam, and Yoruba children of Nigeria. At age 5 years, an average of about 102 cm characterized Amerind children of Surinam, Indonesian children of Surinam, children of Paraguay, and children of Venezuela. 3. Bundi children of New Guinea were no taller in average height at age 5 years than Hum children of Rwanda age 3 years. The same relation was found for Jordanian children age 5 years and Dutch children age 3 years. Compared with

Standing Heighr of Young Children

49

children of East Pakistan, Malaya, and the Philippines, children of Finland, East Germany, and the Netherlands were taller in average height by more than 10 cm. 4. Indian and Nigerian children living at urban centers were more than 3 cm taller in average height than their respective age peers residing in rural regions. Urban-rural differences were near 2 cm for Bulgarian, Czechoslovak, Polish, and Romanian children. There were no differences in average height of rural and urban children belonging to the Creole, Hindustani, and Indonesiah ethnic populations of Surinam. 5. Indian children living in rural areas of Hyderabad state and Bombay province were, on the average, more than 10 cm shorter than upper class urban children of Bombay and Delhi, and about 3 cm shorter than children of low socioeconomic status at the same cities. Thai children residing in middle and upper class homes at Bangkok were no taller than Russian rural children living in the Udor region of the Komi Soviet Socialist Republic. 6. For young inhabitants of Santiago, Chile, and Lagos, Nigeria, average standing height was between 2 and 5 cm less on children of poor families than those of middle class families, and between 4 and 7 cm less on children of middle class families than those of well-to-do families. Turkish young children of wealthy families residing at Istanbul were no taller than American black children of indigent families living at New Orleans. 7. Intercity comparisons showed that at age 3 years children of Santiago, Chile, were about 6 cm shorter in average height than their age peers at Prague, Czechoslovakia. At age 5 years, average height was similar (near 109 cm) for children of Budapest, Hungary, Courtrai, Belgium, St. Louis, Missouri, Sydney, Australia, and Warsaw, Poland. 8. For Creole children of Surinam, black children of the United States, Lunda children of Angola, Hungarian children of Budapest, and East German children, standing height averages at ages 3 and 5 years were practically alike on males and females. Males were between . 8 and 1.2 cm taller than females in child populations of Belgium, Bulgaria, India, Japan, the Netherlands, Romania, Spain, and Tasmania. 9. At age 3 years, a study in the United States showed that twins were shorter than children from singleton pregnancies, and studies in Canada, India, and the United States showed that children weighing less than 2.5 kg at birth were slightly shorter than children who at birth weighed 2.5 kg or more. At age 5 years, studies in the United States found that first-born children were slightly taller than children of later birth order, and children of mothers who did not use tobacco during pregnancy were slightly taller than children whose mothers smoked 15 or more cigarettes daily throughout pregnancy. 10. Large samples drawn from more than 30 populations showed that variability in standing height was not systematically higher on one sex than the other. At age 3 years, standard deviations were near 4 cm in Chuvash, Czechos-

50

Howard V . Meredith

lovak, Dutch, Italian, Ladino, Polish, and Russian populations. Standard deviations at age 5 years were near 5 cm in Australian, Azerbaijani, Belgian, Bulgarian, Chinese, Chuvash, Czechoslovak, Dutch, East German, Hungarian, Italian, Kirghiz, and Surinam In’dustani populations. Greater variability, standard deviations near 6 cm at both ag,es, was found in populations of Indian, Lunda, and Creole children. ACKNOWLEDGMENTS Gratitude is expressed to the following persons who assisted with literature search, reference procurement, provision of unpublished material, language translation, verification of statistics, or manuscript typing: M. C. Bushey, H. Chrzastek-Spruch, 0. Eiben, R . Eksmyr, S . J. Foman, A. P. Forbes, J. E. Goettsch, R . G . Harvey, M. D. Janes, T. Kambara, S. J . Karayan, K. Kirnura, V. B. Knott, L. A. Malcolm, W. J. McGanity, E. M. Meredith, F . J. Miller, M. A. Ned, 0. Neyzi, M. V. Phadke, B. D. Richardson, M. Sempe, J. Spranger, M. Steffens, M. C. Swaminathan, A. M. Thomson, H. L. Vis, J. Wingerd, and N. Wolanski. For generous support in carrying forward the project, appreciation is tendered to Dean Warren K. Giese, Professors John H. Spurgeon, Steven N. Blair, and R . G . Sargent, all of the College of Health and Physical Education, University of South Carolina.

REFERENCES Abraham, S . , Lowenstein, F. W., & O’Connell, D. E. Preliminaryfindings of thejrst health and nutrition examination survey, United States, 1971 -I 972: Anthropometric and clinical findings (DHEW Publ. No. HRA-75-1229). Rockville, Md.: National Center for Health Statistics, 1975. Antrobus, A. C . K . Child growth and related factors in a rural community in St. Vincent. Journal of Tropical Pediatrics and Environmental Child Health, 197 I , 17, 188-210. Arkhipova, G . P. Physical development of children and adolescents of the minority nationalities of Sakhalin Island. Gigiena i Sunifariya. 1967, 32, 107-109. Arkhipova, G . P. Physical development of the rural population of Tambov province. Sovetskoe Zdravookhranenie, 1972. 31, 28--31. Ashcroft, M. T . , & Lovell, H. G. Heights and weights of Jamaican children of various racial origins. Tropical and Geographical Medicine, 1964, 4, 346-353. Ashcroft, M. T., Lovell, H. G . , George, M . , & Williams, A. Heights and weights of infants and children in a rural community of Jamaica. Journal of Tropical Pediatrics, 1965, 11, 5 6 4 8 . Backstrom, L., & Kantero, R.-L. Studies on growth of Finnish children from birth to ten years: 11. Cross-sectional studies of height and weight. Acta Paediatrica Scandinavica, Supplementurn, 1971, 220, 9-12. BakiEovi, Z . , & Sevfikova, A. Dynamika r a m mestskych a vidieckych deti od narodenia do troch rokov. Bratislavske L e E r s k Lisiy, 1973, 59, 84-94. Banik, N . D. D., Krishna, R . , Mane. S. I . S., & Raj, L. Longitudinal growth pattern of children during preschool age and its relationship with different socioeconomic classes. lndian Journal of Pediatrics, 1970, 37, 4 3 8 4 4 7 . (a) Banik, N. D. D., Krishna, R . , Mane, S . 1. S . , & Raj, L. A longitudinal study of physical growth of children of different birth weight groups from birth to 5 years. lndian Journal offediatrics, 1970, 37, 95-101. (b)

Standing Height o j Young Children

51

Banik, N. D. D., Krishna, R . , Mane, S . I. S . , Raj, L., & Taskar, A. D. A longitudinal study of physical growth of children from birth up to 5 years of age in Delhi. lndian Journal ofMedical Research, 1970, 58, 135-142. (c! Banik, N. D. D., Nayar, S . , Krishna, R.. & Raj, L. The effect of nutrition on growth of preschool children in different communities in Delhi. Indian Pediatrics, 1972, 9, 460-466. Barja, I . , LaFuente, M. E., Ballester, D., Monckeberg, F., & Donoso, G. Peso y talla de preescolares chilenos urbanos de tres niveles de vida. Revista Chilena de Pediarria, 1965, 36, 525-529. Beck, G . J., & van der Berg, B. J . The relationship of the rate of intrauterine growth of low-birthweight infants to later growth. Journal offediarrics, 1975, 86, 504-51 I . Belousov, A. Z . , Kardashenko, V. N.. Kondakova-Varlamova, L. P., Prokhorova, M. V., & Spromskaia, E. P. Dynamics of the physical development of children and adolescents in the city of Grel. Sovetskoe Zdravookhrunenie. 1968, 27, 25-28. Berry, F. B . Chile: Nutrition survey, 1960 (Report, Interdepartmental Committee on Nutrition for National Defense). Washington, D.C.: U.S.Government Printing Office, 1961. Berry, F . B. Kingdom of Thailand: Nurririon survey, 1960 (Report, Interdepartmental Committee on Nutrition for National Defense). Washington, D.C.: U . S . Government Printing Office, 1962. (a) 8erry, F. B. Republic of Lebanon: Nutrition survey. 1961 (Report, Interdepartmental Committee on Nutrition for National Defense). Wa\hington, D.C.: U . S . Government Printing Office, 1962. (b) Berry, F. B. West lndies (Trinidad, Tobago, St. Lucia, St. Christopher, Nevis, and Anguilla): Nurririon survey, 1961 (Report, 1nterdepartmen:al Committee on Nutrition for National Defense). Washington, D.C.: 1J.S. Government Printing Office, 1962. (c) Berry, F. B. Hashemite Kingdom of Jordan: Nirrririon survey, 1962 (Report, Interdepartmental Committee oil Nutrition for National Defense). Washington, D.C.: U.S. Government Printing Office, 1963. (a) Berry, F . B. Republic of Uruguay: Nutrition surve,y, 1962 (Report, Interdepartmental Committee on Nutrition for National Defense). Washington, D.C.: U . S . Government Printing Office, 1963. (b) Berry, F. B. Union of B u r m : Nutririon survey, 1961 (Report, Interdepartmental Committee on Nutrition for National Defense). Washington, D.C.: U.S. Government Printing Office, 1963. (Cf

Birkbeck, J . A , , Lee, M . , Myers, G . S . , & Alfred. B . M. Nutritional status of British Columbia Indians: 11. Anthropometric measureinents, physical and dental examinations at Ahousat and Anaham. Canadian Journal ojPublrc Health. 1971, 62, 4 0 3 4 1 4 . Blanca-Adrianzen, T . , Baertl, J . M . , & Graham, G. G. Growth of children from extremely poor families. American Journal of Clinicul Nuwitinn. 1973, 26, 926-930. Blanco, R. A., Acheson, R. M . , Canosa. C . , & Salomon, J . B. Height, weight, and lines of arrested growth in young Guatemalan children. American Journal of Physical Anthropology, 1974, 40, 3947. Boutourline, E., Tesi, G . , Kerr, G . I?.. Stare, F. J . , Kallal, Z., Turki, M., & Hemaidan, N. Nutritional correlates of child development i n southern Tunisia: I . Linear growth. Growth, 1972, 36, 4 0 7 4 2 4 . Burgess, H. J. L., Burgess, A., & Wheeler. E. F. Results and appraisal of a nutrition survey in Malawi. Tropical and Geographical Medicine, 1973, 25, 372-380. Bussadon, G. Valori anthropometrici ed auxologici di una popolazione infantile in alcuni comuni della provincia di Padova (Anno scolastico 1962-1963). Minerva Pediatrica, Monograph Series, 1965, pp. 122-127.

52

Howard V . Meredith

Centurion, C., Miranda, H., & Watkin, D. Republic of Paraguay: Nutrition survey, 1%5 (Report, United States National Center for Chronic Disease Control, Public Health Service). Washington, D.C.: U.S. Government Printing Office, 1967. Chase, H. P., Kumar, V., Dodds, J. M., Sauberlich, H. E., Hunter, R. M., Burton, R. S., & Spalding, V. Nutritional status of preschool Mexican-American migrant farm children. American Journal of Diseases of Children, 1971, 122, 3 16-324. Chen, M. L., Chiang, C. H., Huang, C. S . , & Chen, J. S. A study on the nutritional status and physical growth of Chinese childrem and young adults. Journal of Formosan Medical Association, 1974, 73, 374-386. Cherry, F. F. Growth from birth to five years of New Orleans underprivileged Negro children. Bulletin of the Tulane University Medical Faculty, 1%8, 24, 233-240. Chrzastek-Spruch, H., & Szajner-Milart, 1. Wskazniki rozwoju fizycznego dzieci i mJodziezy miasta lublina od 0 od 17 lat. Pediatria Polska, 1973, 48, 1495-1500. Collis, W . R. F., Dema, I., & Lesi, F. E. A. Transverse survey of health and nutrition, Pankshin Division, Northern Nigeria. West .4frican Medical Journal, 1962, 11, 131-154. Coy, J. F., Lewis, I. C., Mair, C. H., Longmore, E. A., & Ratkowsky, D. A. The growth of Tasmanian infants from birth to three years of age. Medical Journal of Australia, 1973, 2, 12-18. Cruise, M. 0. A longitudinal study of the growth of low birth weight infants: 1. Velocity and distance growth, birth to 3 years. Pediatrics, 1973, 51, 620428. Darby, W. J., & Edozien, J.Republic qfNigeria, Nutrition survey, 1965 (Report, Nutrition Section, Office of International Research, National Institutes of Health, United States Department of Health, Education, and Welfare). Washington, D.C.: U.S. Government Printing Office, 1967. Dattoli, A. Standards della statura a del peso dai 4 al 12 anni in bambini del gruppo etnico italiano di Bolzano. Rivista ltaliana di Medicina e Igiene della Scuola, 1%5, 11, 1 1 1-126. David, J. H. S. Height growth of melanodermic natives in northeastern Lunda (Angola). South African Journal of Medical Science. 1972, 37, 49-60. DeLuca, G., Bearzotti, L., & Cozzi, M. Valori auxologici medi normali del bambino friulano della Provincia di Udine. Minerva Pediatrica, Monograph Series, 1%5, pp. 209-210. Desai, P., Miall, W. E., & Standard, K. L. A five-year study of infant growth in rural Jamaica. Wesr lndies Medical Journal, 1969, 18, 210-221. Doornbos, L., Jonxis, J. H. P., & Visser, H. K. A. Growth of Bushnegro children on the Tapanahony River in Dutch Guyana. Humun Biology, 1968, 40, 396415. Dorozhnova, K. P. Shifts in physical development and changes of frame in preschool age children in connection with acceleration. Gigiena i Sanitariya. 1973, 38, 27-32. Drenhaus, U., Skrobak-Kacznski, J., c Jorgensen, J. 9 . Uber den sakularen trend der akzeleration bei gronlandischen Eskimos au!i dem nordlichen bezirk von Upernavik. Zeirschrifr fur Morphologie und Anthropologie, 1974, 65, 293-304. Dugdale, A. E., MacKay, D. A., Lim, R. K. H., & Notaney, K. H. Growth charts based on measurements of Malay preschool children. Medical Journal of Malaya, 1972, 27, 85-88. Eiben, O., Hegediis, G., Banhegyi, M., Kiss, K., Monda, M., & Tasnadi, 1. Budapest; ovodhsok es iskolasok resti fejlettskge (1968-1969). Budapest: Eotvos Lorand University, 1971. Eksmyr, R . Anthropometry in Ethiopi,an private school children. Nutrition and Metabolism. I97 1, 13, 7-20. Fernandez, N. A,, Burgos, J. C., Asernjo, C. F., & Rosa, 1. R . Nutrition survey of five rural Puerto Rican communities. Boletin de la Asociucion Medica de Puerto Rico, 1969, 61, 42-52. Fisk, S. C. Blackfeet Indian Reservation: Nutrition survey, 1961 (Report, Interdepartmental Committee on Nutrition for National Defense and Division of Indian Health, U.S. Public Health Service). Washington, D.C.: U S . Government Printing Office, 1964, (a)

Standing Height of Young Children

53

Fisk, S. C. Bolivia: Nutrition survey, 1%2 (Report, Interdepartmental Committee on Nutrition for National Defense). Washington, D.C.: U.S. Government Printing Office, 1964. (b) Fisk, S. C. Federation of Malaya: Nutrition survey, 1962 (Report, Interdepartmental Committee on Nutrition for National Defense). Washington, D.C.: US. Government Printing Office, 1964. (C) Fisk, S. C. Fort Belknap Indian Reservation: Nutrition survey, 1961 (Report, Interdepartmental Committee on Nutrition for National Defense and Division of Indian Health, U.S. Public Health Service). Washington, D.C.: U . S . Government Printing Office, 1964. (d) Fisk, S. C. Venezuela: Nutrition survey, 1963 (Report, Interdepartmental Committee on Nutrition for National Defense). Washington, D.C.: U.S. Government Printing Office, 1964. (e) Fisk, S. C. Northeast Brazil: Nutrition survey, I963 (Report, Interdepartmental Committee on Nutrition for National Defense). Washington, D.C.: U.S. Government Printing Office, 1965. Fitzhardinge, P. M . , & Steven, E. M. The small-for-date infant: 1. Later growth patterns. Pediatrics, 1972, 49, 671-681. Forbes, A. P., Ronaghy, H. A , , & Majd, M. Skeletal maturation of children in Shiraz, Iran. American Journal of Physical Anthropology. 197 1, 35, 449454. Franckx, H. Lengte en gewicht bij scholieren uit het kortrijkse. Archives Belges de Medecine Sociale, Hygiene, Medicine du Travail et Medecine Ligale, 1969, 27, 109-1 14. Frisancho, A. R., & Baker, P. T. Altitudeand growth: A study of the patterns of physical growth of a high altitude Peruvian Quechua population. American Journal of Physical Anthropology, 1970, 32, 219-292. Fryer, B. A,, Lamkin, G. H., Vivian, V. M., Eppright, E. S., & Fox, H. M. Growth of preschool children in the north central region. Journal ofthe American Dietetic Association, 1972, 60, 30-37. Gabrielli, G., Lucchetta, G., Chiurazzi, L., DiGiorgi, R., & Curri, P. Indagine auxologica sulla popolazione scolastica del comune di Roma: Nota I. Correlazione peso-statura in un campiore di 6,000 soggetti tra i 4 ed i 13 anni. Pediatria Internazionale, 1967, 17, 257-281. Garcia-Almansa, A . , Fernandez-Femandez, M. D., & Palacios-Mateos, J. M. Patrones de crecimiento de 10s nirios espafioles normales. Revista Clinica Espariola, 1969, 113, 4 5 4 8 . Gasto), B. Badania nad dynamizmen wzrastania dzieci wiejskich przedszkoli powiatu krakowskiego. Folia Medica Cracoviensia, 1966, 8, 191-210. Gerylovova, A , , & Bouchalova, M. The relation between children’s and parents’ heights in the age-range 0-6 years. Annals of Human Biology, 1974, 1, 229-232. Geubelle, F., Lambrechts, L., Sabatier, J., & Baltia, A. Poids et taille des enfants et adolescents sains de la province de Liege. Revue Medicale de Liege, 1974, 29, 582-590. Ghai, 0. P., & Sandhu, R. K. Study of physical growth of Indian children in Delhi. Indian Journal of Pediatrics, 1968, 35, 91-108. Ghosh, S., & Ben, S. Standard of prematurity for North Indian babies. Indian Journal of Child Health, 1962, 11, 210-215. Gokulanathan, K. S., & Verghese, K. P. Socio-cultural malnutrition: Growth failure in children due to socio-cultural factors. Journal of Tropical Pediatrics, 1969, 15, 118-124. Goldfeld, A . Y., Merkova, A. M., & Tseimlina, A. G. Materials on the physical development of children and adolescents in cities and rural localities of the U.S.S.R. Leningrad: Meditsina, 1965. Goldstein, H. Factors influencing the height of seven year old children: Results from the National Child Development Study. Human Biology, 1971, 43, 92-1 11. Guzman, M. A., Scrimshaw, N. S., Bruch, H. A , , & Gordon, J. E. Nutrition and infection field study in Guatemalan villages, 1959-1964: VII. Physical growth and development of preschool children. Archives of Environmental Health, 1968, 17, 107-1 18.

54

Howard V . Meredith

Harvey, R. G . An anthropometric study of growth and physique of the populations of Karkar Island and Lufa subdistrict, New Guinea. Philosophical Transactions of the Royal Society of London, Series B , 1974, 268, 279-292. Hawley, T. G., & Jansen, A. A. J. Height and weight of Fijians in coastal areas from one year till adulthood. New Zealand Medical Journal, 197 1 , 73, 346-349. Heller, C. A., Scott, E. M., & Harmnes, L . M. Height, weight, and growth of Alaskan Eskimos. American Journal of Diseases ojChildren, 1967, 113, 338-344. Hernandez, L. E. Crecimiento y desarrollo normal de nifio pobre dominicano. Boletin de la Oficina Sanitaria Panamericana, 1966, 61, 27-39. Hiernaux, J., & Asnes, D. Le croissance des enfants Sara de 3 a 10 ans a Fort-Archambauh (Republique du Tchad). Bulletin et MPmoih de Soc dAnthropologie de Paris, 1974, 13, 427453. Holibkova, A , , & Holibka, V. Die korperliche entwicklung der kinder der Olomoucer landlichen. Acta Universitatis Palackianae Olomucensis Facultatis Medicae, 1968, 51, 127-138. Hopkins, J. W. Height and weight of Ottawa elementary school children of two socioeconomic strata. Human Biology, 1947, 19, 68-82. Indian Council of Medical Research. Growth and physical development of Indian infants and children: All India. New Delhi: Stalistics Division of Indian Council of Medical Research, 1968. Janes, M. D. Physical growth of Nigerian Yoruba children. Tropical and Geographical Medicine. 1974, 26, 389-398. Jones, D. L., & Hemphill, W. Height. weight and other physical characteristics of New South Wales children. Part 2 . Children underfive years of age. Sydney: Health Commission of New South Wales, 1974. Jones, D. L., Hemphill, W., & Meyers, E. S . A. Height, weighiand otherphysical characteristics of New South Wales children. Part 1 , Children age 5 years and over. Sydney: New South Wales Department of Health, 1973. Jordan, T. E., & Spaner, S . D. Biological and ecological influences on development at 24 and 36 months of age. Psychological Reports, 1972, 31, 319-332. Jordan, T. E., & Spaner, S . D. Biological and ecological influences on development at 48 and 60 months of age. Psychological Reports, 1974, 34, 119-126. Kaganovich, D. I . , Kraeba, N. A , , & Bogachanov, N. D. Physical development and hemoglobin level in children living i n Kemerovo age 1-7 years. Pediatrya (Moscow), 1966, 45, 81-82. Kambara, T. Studies on growth and nutritional intake of school children in Kagawa Prefecture: Regional differences in growth. Shikoku Igaku Zasshi, 1969, 25, 570-582. Kettle, E. S. Weight and height curves for Australian Aborigines infants and children. Medical Journal of Australia, 1966, 1, 972-971. Khanjanasthiti, P., Supachaturas, P., Mekanandha, P., Srimusikapodh, V.. Choopanya, K., & Leesuwan, V. Growth of infants and preschool children. Journal of the Medical Association of Thailand, 1973, 56, 88-100. Khurana, V., Agarwal, K. N., Manwani, A. H., & Srivastava, G . Physical growth in first five years. Indian Pediatrics, 1971, 8, 331 -335. Kimura, K., & Tsai, C. M. Comparative studies of the physical growth in Formosans: I . Height and weight. Zinruigaku Zassi, 1967, 75, 11-18. Kogan, R. B. Assessing the high level of physical development of young children. Zdravookhranenie Rossiiskoi Federatsii, 1969, 13, 29-32. Kozhonazarov, K. K. Physical development of children in nursery schools of the town of DzhalalAbad in the Oshsk region. Sovei'skoe Zdravookhranenie Kirgizi!, 1966, 6 , 18-22. Kravchenko, A. G. Physical development of rural children 3 to 7 years old in the Odessa district. Sovetskoe Zdravookhranenie. 1968, 27, 33-39.

Standing Height of Young Children

55

Kreysler, J . , & Schlage, C. The nutrition situation in the Pangani basin. In H. Kraut & H. D. Cremer (Eds.) lnvestigatiom into health and nutrition in East Africa. Miinchen: Weltforum Verlag, 1969, 85-178. & Syrovatka, A. NEkteradatao &tu skupiny dEti piedsikolKubat, K., Novakova, M., Kouiim, .I., niho vtku dlouhodobc sledovant. Ceskoslovenska Pediatrie, 1969, 24, 256-263. Kudaiarov, D. K. Principal indices of physical development of children 3 to 7 years old Permanent inhabitants of high mountainous (Naryr) and low mountainous (Frunze) regions of Kirghizia. Sovetskoe Zdravookhranenie Kirgizii, 1966, 6, 12-15. Kumiewicz-Witczakowa, R., Miqsowicz, I . , Mazurczak, T. Evaluation of certain features of somatic development of Warsaw children aged up to 6.5 years using centile grades and tables. Pediatria Polska, 1974, 49, 427437. Kurniewicz-Witczakowa, R., Miqsowicz, I . , Mazurczak, T., & Eska, H. J. Wskatniki rozwoju fizycznego dzieci warszawskich od 0 do 36 miesiaca tycia. Problemy Medycyny Wieku Rozwojowego, 1973, 3, 99-1 17. Kutsenko, V. V. Some indices of the physical development of Ivano-Frankovsk kindergarten children. Pediatrya (Moscow), 1965, 44,81-82. Kuz’min, E. M. Physical development of kindergarten pupils of Petrozavodsk. Gigiena i Sanitariya, 1970, 35, 105-107. Lapitskii, F. G., Belogorskii, V . Y., Nemzer, M. P., Pogorely, I. A , , Sinopalnikov, 0. V . , & Zolotareva, M. P. Basic indices of physical development of children in the second and third years in Murmansk. Zdravookhranenie Rossiiskoi Federarsii. 1970, 14, 28-3 I . Lapitskii, F. G., Belogorskii. V. Y . , Nemzer, M. P . , & Zolotareva, M. P. Physical development of children in nursery schools of Murmansk (transpolar region). Pedidrya (Moscow), 1967, 46, 62-63. Limaye, C . , Chouhan, C. N., Lakhani, S . M . , & Phadke, M. V . Optimal growth standards from Poona. Indian Pediatrics, 1974, 11, 673476. LOW,W. D. Stature and body weight of southern Chinese children. Zeirschrft fur Morphologie und Anthropologie, 1971, 63, 1 1 4 5 . Luna-Jaspe, H., Ariza, J., Rueda-Williamson, R., Mora, J . O . , & Pardo, F. Estudio seccional de crecimiento, desarrollo y nutricion en 12,138 nifios de Bogota, Colombia: 11. El crecimiento de niiios de dos clases socioeconomicas durante sus primeros seis aiios de vide. Archivos LatinoArnericanos de Nutricidn, 1970, 20, 151-165. Luyken, R., & Luyken-Koning, F. W. M. Studies on the physiology of nutrition in Surinam: 1X. Somatometrical data. Documenfa de Medicina Geographica et Tropica, 1961, 13, 123-130. Madhavan, S . , Susheela, T. P . , & Swaminathan. M. C. Association of growth status and the prevalence of nutritional deficiency signs among preschool children. Indian Journal qf Medical Research, 1967, 55, 497-503. Malcolm, L. A. Growth and development of the Bundi child of the New Guinea highlands. Human Biology, 1970, 42, 293-328. Martirosian, R. B. Physical development of preschool and school children of rural regions of the Komi A . S . S . R. Zdravookhranenie Rossiiskoi Federatsii, 1973, 17, 25-26. Marzot, G., Colombini, M., & Vedovello, R. Studio statistico sulla variabilita della statura e del peso in soggetti della Val de Fiemme (provincia de Trento). Clinica Pediarrica (Bologna), 1966, 48, 219-228. Matawaran, A. J . , & Gervasio, C. C. Average heights and weights of Filipinos. Philippine Journal of Nutrition, 1971, 24, 74-92. Matawaran, A. J . , Gervasio, C. C., & de Gala, A. B. Preliminary report on the average height and weight of some Filipinos. Philippine Journal of Nutrition. 1966, 19, 2 9 4 9 . McGanity, W. J . Nutrition survey in Texas. Texas Medicine, 1969, 65, 40-49.

56

Howard V . Meredith

McGregor, I. A,, Rahman, A. K., Thompson, B., Billewicz, W. Z., & Thornson, A. M. The growth of young children in a Gambian village. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1968, 62, 341-352. McKay, D. A,, Lim, R. K. H., Notaney, K. H., & Dugdale, A. E. Nutritional assessment by comparative growth achievement in Malay children below school age. Bulletin of the World Health Organization. 197 1, 45, 233-342. Menzies, F. Report on the average heights and weights of elementary school children in the County of London in 1938. London: London County Council, 1940. Meredith, H. V. Birth order and body size: 11. Neonatal and childhood materials. American Journal of Physical Anthropology. 1950, 8 , 195-224. Meredith, H. V. Relation between socioeconomic status and body size in boys seven to ten years of age. American Journal of Diseases of Children, 1951, 82, 702-709. Meredith, H. V. Methods of studying physical growth. In P. H. Mussen (Ed.),Handbookof research methods in child development. New York Wiley, 1960. Meredith, H. V. Body weight at birth of viable human infants: A worldwide comparative treatise. Human Biology, 1970, 42, 217 -264. Meredith, H. V. Relation between tobacco smoking of pregnant women and body size of their progeny: A compilation and synthesis of studies. Human Biology, 1975, 47, 451472. Miqsowicz, I. Porowanie rozwoju fizycnego dzieci i dodziety z trzech r6tnych Srodowisk miejskich. Prace i Muteriab Nlzukowe, 1964, 2, 125-137. Miqsowicz, I., & Pytuk, M. Rozwtjj fizyczny dzieci nowotarskich i helskich. Prace i Materialy Naukowe, 1967, 10, 19-32. Migasena, P., Thumham, D. I . , Pongpaew, P., Hongthong, K . , & Harinasuta, C. Anthropometric studies in preschool children of northeast Thailand. Journal of Nutritional Science and Vitaminology, 1974, 20, 127-132. Miller, F. J . W., Court, S. D. M., Walton, W. S., & Knox, E. G . Growing up in Newcusde-uponTyne. London: Oxford University Press, 1960. Mohanta, K. D., Pande, T. H., & Praharaj, K. C. Anthropometric measurements of children of western Orissa. Indian Journal of Pediatrics, 1972, 39, 12-14. Montoya, C . , & Ipinza, M. Peso y es(.aturade pre-escolares santiaguinos pertenecientes a dos estratos sociales diferentes. Revista Chilena de Pediatria, 1964, 35, 269-277. Morley, D. C., Woodland, M., Martin, W. J., & Allen, I. Heights and weights of West African village children from birth to the age of five. West African Medical Journal, 1%8, 27, 8-13. Mortison, J. Low birth weight study: Height and weight tables and charts. Systems Development Project, Study Series, 1969, 9/3(4a). Nekisheva, Z. I. Effect of certain social factors on the physical development of preschool children attending kindergartens. Gigiena i Sanitariya, 1974, 2, 58-61, Neumann, C. G., Shanker, H., & Uberoi, I. S. Nutritional and anthropometric profile of young rural Punjabi children. Indian Journal of Medical Research, 1969, 57, 1122-1 149. Neyzi, 0. Somatik gelizme indekslerinin tatbiki degeri. Haseki Tip Bulteni, 1967, 5, 1 4 1 . Neyzi, O., & Gurson, C. T. Physical measurements on two groups of Istanbul children: Normal and undernourished. Stockholm: Swedish Medical Research Counsil, 1966. Neyzi, O., Tanman, F., & Saner, G. Preliminary results of a child health survey in an underdeveloped area of Istanbul. Mimeographed report, 1965. Nouth-Savoeun, Confriburion a I'itude de la croissance physique des enfants cambodgiens de la naissance u 14 ans. Unpublished M.D. thesis, Universite Royale de Phnom-Penh, 1966. O'Brien, R., Girshick, M. A,, & Hunt, E. P. Body measurements of American boys and girls for garment and pattern construction (Bureau of Home Economics, United States Department of Agriculture, Publ. No. 366). Washington, D.C.: U.S. Government Printing Office, 1941.

Standing Height of Young Children

57

Oehmisch, W. Die enwicklung der korpermusse bei kindern und jugendlichen in der Deutschen Demokratischen Republik: Ergebnisse einer reprasentativen untersuchung in den jahren I9671 1968. Berlin: Deutschen Akademie fur Arztliche Fortbildung, 1970. Omran, A. R., McEwen, W. J . , & Zaki, M. H. Epidetniological studies in Bolivia. New York: Research Institute for the Study of Man, 1967. Oppers, V . M. Groeidiagrammen van amsterdamse kinderen. Op Grond van Cijfers, 1964, 2, 5 1 4 5 . Owen, G. M., Kram, K. M., Gamy, P. J . , Lowe, J. E . , & Lubin, A. H. A study of nutritional status of preschool children in the United States, 1968-1970. Pediatrics, 1974, 53, 597-648. (Supplement) Owen, G. M., & Lubin, A. H. Anthroponietric differences between Black and White preschool children. American Journal of Diseuses of Children. 1973, 126, 168-169. Palacios, J . M., & Vivanco, F. Datos de talla y peso de 128,000 nifios espafioles. Revista Clinica Espariola, 1965, 99, 230-238. Palacios-Mateos, J . M., Garcia-Almansa, A,, Vivanco, F., Fernandez, M. D., Garcia-Robles, R., & Moreno-Estaban, B. El crecimiento de 10s niiios espaiioles desde el nacimiento hasta 10s cinco aiios. Revista de Sanidud e Higiene Publicu, 1972, 46, 1093-1 105. Pereira, S. M. Growth and development of children in underprivileged communities, Tamil Nadu. Indian Journal of Medical Research, 197 I , 59, 156-163. (Supplement) Phadke, M. V., & Kulkarni, H. D . Growth and development in the underprivileged sections of the Bombay area. Indian Journal of Medical Research, 1971, 59, 164-176. (Supplement) Phadke, M. V., Deodhar, N. S . , & Kulkarni, H. D . The rural preschool child-his growth and nutritional status. Indian Journal of Medical Research, 197 I , 59, 748-755. Pire, E., & Meli, G. Rilevazioni auxologiche sulla popolazione scolastica della citta di Bari. Revista ltaliana de Medicina e Igiene della Sruola. 1962, 8, 307-318. Pokatilo, V. M. Dynamics of the physical development and physical training of kindergarten children. Gigiena i Sanitariya, 1974, 12, 93-94. Prasad, R., Kumar, R., & Dayal, R. S. Physical growth and development from one to five years. Indian Pediatrics, 1971, 8, 105-1 19. PyEuk, M., & Wolanski, N. Niektore wJasciwosci fizjologiczne dzieci i d o d z i e f y jako wynik adaptacji do rotaych warunkow Srodowiskowych: Doniesienie 11. Prace i Materialy Naukowe, 1968, 11, 129-151. Rao, N. P., Singh, D., & Swaminathan, M. C. Nutritional status of preschool children of rural communities near Hyderabad City. Indian Journal of Medical Research, 1969, 57, 2132-2146. Rea, J. N. Social and economic influences on the growth of preschool children in Lagos. Human Biology, 1971, 43, 4 6 4 3 . Richardson, B. D. Growth standards: An appraisal with special reference to growth in South African Bantu and White preschool children. South African Medical Journal, 1973, 47, 699-702. (Supplement) Roche, A. J . , & Davila, G. H. Differences between recumbent length and stature within individuals. Growth, 1974, 38, 313-320. Rosenberg, I . H . , & Reiner, M. L . Nutrition survey of E m Pakistan (Report, Ministry of Health, Government of Pakistan, with University of Dacca and Nutrition Section, Office of International Research, National Institutes of Health, United States Department of Health, Education, and Welfare). Washington, D.C.: U.S. Government Printing Office, 1966. Rueda-Williamson, R., Luna-Jaspe, H . , Ariza, J., Pardo, F., & Mora, J. 0. Estudio seccional de crecimiento, desarrollo y nutricion en 12,138 niiios de Bogoti, Colombia. Sociedad Colombiana de Pediarria y Puericultura, 1969, 10, 335-339. Rutishauser, 1. H. E. Heights and weights of middle class Baganda children. Lancet, 1965, ii, 565-567.

58

Howard V . Meredith

Scott, J. A. Report on the heights and weights of school pupils in the County of London in 1959. London: London County Council, 1961. Scrimshaw, N. S . , Guzman, M. A , , Kevany, J. J., Ascoli, W., Bruch, H. A,, & Gordon, J. E. Nutrition and infection field s.tudy in Guatemalan villages, 1959-1964. Archives of Environmental Health, 1967, 14, 787-801. SempC, P., SempC, M., & P d r o n , G. Croissance et maturation osseuse. Paris: Theraplix, 1972. Sharma, J. C. Physical growth and development of fhe Maharastrians, Lucknow: Ethnographic and Folk Culture Society, 1970. Shakir, A,, & Zaini, S. Skeletal niaturation of the hand and wrist of young children in Baghdad. Annals of Human Biology, 1974, 1, 189-199. Shattock, F. M. Heights and weights in a Javanese labour population of a north Sumatran rubber estate. Tropical and Geographical Medicine. 1968, 20, 147-158. Shiddhaye, S., Shah, P. M., & Udani, P. M. Physical growth and development of pre-term children during the first five years of life: Longitudinal study. Indian Pediatrics, 1972, 9, 282-289. Smith, D. S . , & Brown, M. L. Anthropometry in preschool children in Hawaii. American Journal of Clinical Nutrition, 1970, 23, 932-939. (a) Smith, D. S . , & Brown, M. L. Growth measures of a selected group of preschool children in Honolulu. Hawaii Medical Journal, 1970, 29, 375-377. (b) Spranger, J., Ochsenfarth, A,, Kock, H. P., & Henke, J. Anthropometrische normdaten im kindesalter . Zeitschrift fur Kinderheilkunde, 1968, 103, 1 - 12. Standard, K. L., Desai, P., & Miall, W. E. A longitudinal study of child growth in a rural community in Jamaica. Journal ofBiosocia1 Science, 1969, 1, 153-176. Stracker, 0. A. Korpermasse der kinder und jugendlichen im jahre 1962. dsterreichische Arztezeitung, 1%3, 18, 1-24, Swaak, A. J. Een onderzoek bij 125 kinderen met het syndroom van Langdon Down (mongoloidisme). Nederlands Tddschrift voor Geneeskunde, 1967, 111, I 10-120. Swaminathan, M. C., Jyothi, K. K . , Singh, R., Madhaven, S . , & Gopalan, C. A semi-longitudinal study of growth of Indian children and the related factors. Indian Pediatrics, 1964, 1, 255-263. Systems Development Project Staff. Height and weight charts and tables on poverty area children. Systems Development Project, Study Series. 1968, 819. Sukkar, M. Y., Johnson, D., Gadir, A. M. A,, & Yousif, M. K . The nutritional status of children in rural Khartoum. Sudan Mediral Journal, 1971, 9, 23-38. Taback, M .Birth weight and lenglh of gestation with relation to prematurity. Journal of rhe American Medical Association. 1951, 146, 897-901. Tinisescu, G., Chiriac, I . . Stinciulescu, E., Domi!escu, M., & Jelezneac, I . Nivelul dezvolt2rii fizice a copiilor de 0-3 ani din republica socialists romlnia. Pediatrirr (Bucharest), 1970, 19, 97-108. Tanner, J . M., Whitehouse, R. H., & Takaishi, M. Height, weight, height velocity, weight velocity: British children, 1965. Archives of Disease in Childhood, 1966, 41, 454-471, 613-635. Tatafiore, E. Indagini auxologiche sull’accelerazione dell’acaescimento e sulla statura definitiva. Minerva Pediatrica. Monograph Series, 1965, pp. 69-72. Terrosi, F. lndagine auxometrica nel bambino della provincia di grosseto. Archivio Italiano di Pediatria e Puericulrura, 1968, 26, 148-160. Thomson, A. M., Billewicz, W. Z . , Thompson, B., Illsley, R., Rahman, A. K . , & McGregor, I. A. A study of growth and health of young children in tropical Africa. Transacrions ofthe Royal Society of Tropical Medicine and Hygiene, 1968, 62, 333-340. Tie, L. T., Lian, 0. K., Liong-Chg, T. W., & Rose, C. S. Health, development, and nutritional survey of preschool children in central Java. American Journal of Clinical Nutrition, 1967, 20, 1260-1 266.

Standing Height oJ Young Children

59

Tonelli, E. Studio statistic0 sulla variabilita della statura e del peso nei soggetti emiliani di eta compresa fra 1 e 14 anni. Giornale di Igiene e Medicina Preventiva, 1963, 1, 3-17. Tuxford, A. W., & Glegg, R. A. The average height and weight of English school children. British Medical Journal, 191 I , 1, 1423-1424. Twiesselrnann, F. Diveloppement biometrique de l'enfant a I'adulre. Brussels: Presses Universitaires de Bruxelles, 1969. Udani, P. M. Physical growth of children in different socioeconomic groups in Bombay. Indian Journal of Child Health, 1963, 12, 593-6 I 1 . United States Center for Disease Control. Ten-State nutrition survey, 1968-1970. 111. Clinical, anrhropometry. dental (United States Department of Health, Education, and Welfare, Publ. No. HSM-72-8131). Washington, D.C.: U.S. Government Printing Office, 1972. Van der Kuyp, E. Body weights and heights of the Surinam people. Voeding, 1967, 28, 435469. van Wieringen, J . C. Seculaire groeiverschuiving: Lengte en gewicht surveys 1964-I966 in Nederland in historisch perspectief. Leiden: Netherlands Instituut voor Praeventieve Geneeskunde TNO, 1972. Verghese, K. P., Scott, R . B . , Teixeira, G . , & Ferguson, A. D . Studies in growth and development: XII. Physical growth of North American Negro children. Pediatrics, 1969, 44, 243-247. Vesi, G., & Cantalini, C. Indagine auxologica sul bambino aquilano della scuola matema (dal3" al6" anno di vita). Minerva Pediatrica, 1972, 24, 223-237. Vis, H. L. Protein deficiency disorders. Postgraduate Medical Journal, 1969, 45, 107-1 15. Vizzoni, L., & Baldini, G. lndagine statistica sulle alterazion dell'accrescimento del bambino della province di Pisa. Rivista di Clinica Pediatrica, 1964, 73, 185-198. Wark, L., & Malcolm, L. A. Growth and development of !he Luni child in the Sepik District of New Guinea. Medical Journal of Australia, 1969, 2, 129-136. Weisman, S. A. Contour of the chest in children: Environment. American Journal of Diseases of Children, 1935, 49, 52-59. Wilson, R . S . Growth standards for twins from birth to four years. Annals of Human Biology, 1974. 1, 175-188. Wingerd, J., & Schoen, E. J . Factors influencing length at birth and height at five years. Pediatrics. 1974, 53, 737-741. Wolanski, N. Ocena rozwoju fizycznego dziecka w wieku do trzech lat. Prace i Muterialy Naukowe, 1964, 2, 95-124. Wolanski, N., & PyBuk, M. (Eds.). Studies in human ecology. Warsaw: Polish Academy of Sciences, 1973. Wong, H. B., Tye, C . Y., & Quek, K. M. Anthropometric studies on Singapore children: I . Heights, weights and skull circumference on preschool children. Journal of the Singapore Paediatric Sociezy, 1972, 14, 68-89. Woodbury, R . M. Statures and weights of children under six years of age (United States Children's Bureau, Publ. No. 87, United States Department of Labor). Washington, D.C.: U.S. Govemment Printing Office, 1921. Yarbrough, C., Habicht, J.-P., Malina, R . , Lechtig, A , , & Klein, R . E. Length and weight in rural Guatemalan Ladino children: Birth to seven years of age. American Journal of Physical Anthropology, 1975, 42, 439447. Zauli-Naldi, G. Dati auxologici normali del bambino dai 3 ai 12 anni della provincia di teramo. Aggiornamento Pediarrico, 1963, 18, 69-75. Zhaglina, A. K. Anthropometric and hemodynamic indices of preschool children. Gigiena i Sanitariya, 1974, 2, 108-109.

This Page Intentionally Left Blank

THE REPRESENTATION OF CHILDREN’S KNOWLEDGE

David Klahr and Robert S . Siegler CARNEGIE-MELLON UNIVERSITY

I . INTRODUCTION

....

62

I1 . FROM BEHAVIORAL TO COGNITIVE OBJECTIVES . . . . . . . . . . . . . . . . . . . . .

63

III . SOME CRITERIA FOR CHOOSING A REPRESENTATION . . . . . . . . . . . . . . . .

64

IV . BALANCE SCALE TASK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . REPRESENTATION OF CHILDREN’S KNOWLEDGE ABOUT THE BALANCE SCALE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . ASSESSING THE ACCURACY OF THE REPRESENTATIONS . . . .

66 67 68

V . EXPERIMENT I: ASSESSING INITIAL KNOWLEDGE . . . . . . . . . . . . . . . . A . METHOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C . EVALUATION OF DECISION TREE REPRESENTATION . . . . . . . . . . . . . .

13

VI . EXPERIMENT 2: TRAINING ON THE BALANCE SCALE TASK . . . . . A . METHOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

74 74 76

69 70 71

VII . REVISED REPRESENTATIONS FOR BALANCE SCALE KNOWLEDGE . . . . . . A . PRODUCTION SYSTEM REPRESENTATION ..................... . . . . B . EVALUATION OF THE PRODUCTION SYSTEM REPRESENTATION C . PROTOCOL ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D . REVISED PRODUCTION SYSTEM FOR A MODEL nr CHILD . . . . . . E. PRODUCTION SYSTEM FOR MODEL IlIA ....................... F . PRODUCTION SYSTEM INTERPRETATION ...................... G . EVALUATION OF REPRESENTATIONS FOR JAN’S KNOWLEDGE . . . . .

11 77 81 82 85 89 89

VIII . EXPERIMENT 3: ENCODING HYPOTHESIS . . . . . . . . . . . . I. . . . . . . .... A . RECONSTRUCTION PARADIGM . . . . . . . . . . . . . . . . . . . . . . . . . . .... B . SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

95 96

94

101

61

62

David Klahr and Robert S . Siegler

IX. DISCUSSION: SOME ANSWERS AND SOME FURTHER QUESTIONS . . . . . . . A. WHERE WE HAVE BEEN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. WHERE WE ARE GOING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

102 102

103

X. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

105

APPENDIX A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

106

APPENDIX B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

110

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

113

I.

Introduction

In this chapter we will discuss two related issues. One issue concerns the ways that children from 5 to 17 years perform a scientific induction task. We will summarize a series of experiments designed to investigate questions about initial knowledge, instructional effectiveness, and individual differences in both initial performance and responsiveness to instruction. The second issue is methodological: its focus is not on what we can say about children’s knowledge of a task, but rather on how we can say it. That is, the second issue we will address is the representation of children’s knowledge. The two issues are related simply because the researcher’s decision about how to represent knowledge plays a central role in guiding both the kind of theory that gets formulated and the kind of experiment that gets run. We have found this to be the case in our own studies, and we believe that it might be worthwhile to direct attention to some properties of different representations and criteria for choosing among them. Our discussion will move back and forth between general conceptual issues and some very specific examples of both empirical techniques and theoretical statements. We will start by describing the historical trend in instructional psychology that has made the representation of knowledge a central issue, and then we introduce criteria that we believe might be useful in choosing and evaluating different representations. In addition to a set of evaluative criteria, we will list five central questions for research in developmental and instructional psychology. Next we will introduce a specific task that has interesting psychological and instructional properties: a variant of Piaget’s balance scale problem. We will present a formal model-using a particular representation-for different levels of knowledge that children might have about how to do the task. The task will provide the concrete reference for the rest of our chapter.

The Representation of Children’s Knowledge

63

Having described the formal properties of the task, and some predictions about the performance of different aged children on it, we will then describe our first experiment. Based upon the results of the experiment, we will evaluate the initial hypotheses, as well as examine the merits and limitations of the representation in which the initial models are stated. The initial representation and the associated experiment enable us to make certain predictions about the effects of an instructional sequence. In the second experiment we will explore some instructional issues, and this in turn will reveal some limitations of the initial representation. In particular, we find that older and younger children who are initially classified by our models as having identical task-specific knowledge show a striking differential responsiveness to instruction. This presents a serious challenge to our initial representation of children’s knowledge. It is clear that the initial formulation does not tell the whole story about differences in task-specific knowledge. A revised representation of the knowledge required to perform at different levels is introduced. The representation is a production system and some of the general properties of production systems are discussed. Then we will present an analysis of the problem-by-problem performance of two children during a training sequence, and formulate a more detailed production system model of the knowledge of one of them. The model is actually run as a computer simulation and its results are compared with the child’s performance. This fine-grained analysis suggests that the initial encoding of the stimulus may be a crucial difference between older and younger children, and that this may account for the results of Experiment 2. This encoding hypothesis states that differential responsiveness to training between 5- and 8-year-old children is due to differences in the way they encode the balance scale dimensions. The explanatory power of the hypothesis is illustrated in detail in Experiment 3. Finally, we will summarize the preceding discussion in terms of how well it answers our initial set of questions. Then we will briefly discuss the several types and levels of knowledge that might be important in instructional investigation. Different levels of aggregation of both model and data are obviously appropriate for different scientific questions. The direct and explicit consideration of some ways to represent knowledge will provide useful guidelines for further empirical work.

11. From Behavioral to Cognitive Objectives The goal of any instructional effort is the production of new knowledge in the learner. Over the last 15 years of instructional research there has been an increasing emphasis on stating such goals as clearly as possible. The trend was to move

64

David Klahr and Robert S . Siegler

from an emphasis on simply describing educational means-the sequence of instructional activities-to a prior statement of the desired ends of instruction. The elaboration of behavioral objectives was perhaps the most extensive formalization of this trend (Mager, 1962). Behaviors were typically specified in great detail, although the underlying processes were not. However, even behavioral objectives have implicit in them an underlying cognitive theory: Behavior is simply an observable indicator of underlying cognitive processes. One well-known normative model for instruction (Glaser, 1968) stresses the need to determine the learner’s initial state as well as the desired end state. In the original formulations of this approach, both initial and final learner states were described primarily in terms of tasks and subtasks arranged in a Gagne-like hierarchy. There was little mention of how one might characterize the underlying psychological processes that acted upon them to produce the task behavior in question. As that approach has developed, however, it has focused increasingly upon such cognitive representations (see Resnick, 1976, for a summary of this trend in the area of elementary mathematics instruction). Perhaps the strongest statement of the desirability and feasibility of describing the learner in terms of internal psychological representations is Greeno’s view of “cognitive objectives.” Greeno ( 1976) argued that cognitive psychology has now developed powerful and flexible methods for the representation of knowledge. Using an example from instruction in elementary fractions, Greeno showed how two different views of the conceptual content of the subject matter can be represented explicitly by two quite distinct cognitive structures, which in turn lead to differential predictions about problem difficulty, problem-solving strategies, and optimal procedures for instruction. Without such a representation, these predictions might never have been made. This is essentially the same point stressed by Klahr and Wallace (1976) with respect to the need for explicit and precise models in cognitive development: “A theory of transition can be no better than the associated theory of what it is that is undergoing that transition” (p. 14). According to this view, the first step in the formulation of developmental theories is the creation of a precise model of the initial and final form of the cognitive process under investigation. Studies by Baylor and Gascon (1974), Young (1973), Klahr and Wallace (1976), and several others, summarized in Siegler (1978), provide developmental models of this type.

111. Some Criteria for Choosing a Representation What considerations might guide us in the choice of a representation? What kinds of representations are available, and what are their relative merits? Knowledge representation has become an important topic in the emerging field of

The Representution o j Children’s Knowledge

65

“cognitive science” (Bobrow & Collins, 1975), and some initial attempts to address it can be found in Bobrow ( 1975), Becker ( 1975), Moore and Newell (1974), and Reddy and Newell (1974). These efforts constitute the first steps toward a full-fledged theory of representation, and they have already yielded a reasonable set of dimensions with which to characterize different representations. Although such taxonomic systems allow us to classify representations, they do not make any statements about their relative merits. Regardless of the final location of a representation along the dimensions of importance, the ultimate evaluation of the quality of a representation depends upon the set of questions being addressed. We believe that in the area of instruction and development the important questions are: Question 1: What are the differences in knowledge that underlie different levels of task performance? Question 2: What are the alternative strategies that might result in any given level of task performance? Question 3: For a given level of performance, what is the optimal level of difficulty for an instructional sequence? Question 4: What are the critical features of an instructional sequence that enable it to have any effect? Question 5: When and why will two learners at the same initial performance level learn differently from the same instructional sequence? Or, to summarize our concerns: What do children know about a task, how do they learn about it, and why do some know more and/or learn more than others? Given this set of questions, there are four criteria that we believe to be most important in choosing a representation: 1. The representation must be sufficient to account for behavior. Thus, it must have a clear mapping onto the empirical base it is supposed to account for. 2. It should be amenable to multiple-level analyses. That is, it should be easy to aggregate and disaggregate the grain of explanation. For the design of wellcontrolled experiments or curriculum design, the representation will have to be stated in terms of averages across many subjects; it must be a modal form. For detailed study of individual strategies and component processes, it must be capable of disaggregation without drastic revision. 3. The representation should conform to the relevant properties of the human information-processing system as determined by laboratory studies of human processing capacities. 4. The representation should have “developmental tractability” (Klahr & Wallace, 1970). It should allow us to state both early and later forms of competence and provide an easy interpretation of each model as both a precursor and successor of other models in a developmental sequence (see Resnick, 1976, for a similar viewpoint).

66

David Klahr and Robert S. Siegler

IV. Balance Scale Task The type of balance scale used throughout our investigation consisted of a two-arm balance, with several pegs located at equal intervals along each arm. Small circular disks, all of equal weight, were placed on the pegs in various configurations (as shown in Table I), while the balance was prevented from tipping. The subjects’ basic task was to predict the direction in which the balance scale would move if it were allowed to. In order to answer some of the questions listed above, several variations on this basic theme were introduced. These included: asking children to explain their predictions; allowing the scale to move to its equilibrium position (thus providing feedback about the accuracy of the predictions); observing an experimenter-controlled series of configurations and their effects; constructing one’s own configurations; and reconstructing initial configurations from memory. (A more complete report of these experiments is presented in Siegler, 1976.) The basic physical concept that underlies the operation of the balance scale is torque: The scale will rotate in the direction of the greater of the two torques acting on its arms. The total torque on each arm is determined by summing the individual torques produced by the weights on the pegs, and the individual torques are in turn computed by multiplying each weight by its distance from the fulcrum. Since the pegs are at equal intervals from the fulcrum, and the weights are all equal, a simpler calculation is possible. It consists of computing the sum of the products of number of weights on a peg times the ordinal position of the peg from the fulcrum. This is done for each side, and the side with the greater sum of products is the side that will go down. (If they are equal, the scale will balance .) The components of this knowledge are acquired over a remarkably long span of experience and education; even 5-year-olds often know that balances such as teeter-totters tend to fall toward the side with more weight, while many 16-yearolds do not know the appropriate arithmetic computations for determining the balance’s behavior (Jackson., 1965; Lee, 1971; Lovell, 1961). It even seems likely that most college-educated adults could not easily state the physical principles that underlie the sum-of-products algorithm. Furthermore, for many configurations, there are shortcuts that eliminate the need to do any arithmetic computation (e.g., identical configurations on each arm will balance; if both weight and distance are greater on one side, that side will go down). Note that the balance scale task shares a common property of many scientific problems: The universal rule for generating correct predictions is easy to describe, and once known, it is easily remembered and executed. However, the formulation of the rul-ither by induction from empirical examples or by deduction from general physical principles-is quite difficult.

67

The Representation of Children’s Knowledge

A.

REPRESENTATION OF CHILDREN’S KNOWLEDGE ABOUT THE BALANCE SCALE

Siegler (1976) suggested that the different levels of knowledge that children have about this task could be represented in the form of binary decision trees (see Fig. 1). The model of mature knowledge (Model IV, Fig. 1D) was suggested by a task analysis of balance scale problems; the models of less sophisticated knowledge (Models 1-111, Figs. 1A-C) were derived from the empirical results of Inhelder and Piaget (1958) and Lee (197 l ) , and from our own pilot studies. A child using Model I considers only the number of weights on each side: If they are the same, the child predicts balance, otherwise he predicts that the side with the greater weight will go down. For a Model II (Fig. 1B) child, a difference in weight still dominates, but if weight is equal, then a difference in distance is sought. If it exists, the greater distance determines which side will go down, otherwise the prediction is balance. A child using Model III (Fig. 1C) tests both

Model

Model

I9

m

(C)

(0)

Fig. 1A-D. Decision tree representations for Models I-IV of balance scale predictions. A , Model I ; B, Model II; C . Model I l l ; D . Model IV. D = distance; W = weight.

68

David Kluhr and Robert S . Siegler

weight and distance in all cases. If both are equal, the child predicts balance; if only one is equal, then the other one determines the outcome; if they are both unequal, but on the same side with respect to their inequality, then that side is predicted to go down. However, in a situation in which one side has the greater weight, while the other has the greater distance, a Model 111 child, although recognizing the conflict, does not have a consistent way to resolve it. This child simply “muddles through” by making a random prediction. Model IV represents “mature” knowledge of the task: Since it includes the sum-of-products calculation, children using it will always make the correct prediction. Note, however, that if they can base their prediction on simpler tests, they will do so. B . ASSESSING THE.ACCURACY OF THE REPRESENTATIONS

It is possible to determine which, if any, of these four models accurately characterizes a child’s knowledge about the balance scale task by examining his pattern of predictions for six types of problems (see Table I for an example of each type): (1) balance problems, with the same configuration of weights on pegs on each side of the balance; (2) weight problems, with unequal amounts of weight equidistant from the fulcrum; (3) distance problems, with equal amounts of weight different distances from the fulcrum; (4) conflict-weight problems, with more weight on one side and “more distance” (i.e., occupied pegs further from the fulcrum) on the other, and the configuration such that the side with more weight goes down; (5) conflict-distance problems, similar to conflict-weight, except that the side with more distance goes down; (6) conflict-balance problems, like other conflict probllems, except that the scale remains balanced. Children whose knowledge corresponded to different models would display dramatically different patterns of predictions on the six types of problems just listed. Those using Model I would consistently make correct predictions on balance, weight, and conflict-weight problems, and they would never be correct on the other three problem types. Children using Model I1 would behave similarly to those using Model I ton five of the six problem types, but they would correctly solve distance problems. Those following Model 111 would consistently make accurate predictions on weight, balance, and distance problems, and would perform at a roughly chance level on all conflict tasks. Those using Model IV would solve all problems of all types. To the extent that there is a correlation between age and the level of the model which best represents a child’s knowledge, there should be clear developmental patterns for each problem type. The most interesting is the predicted decrement in performance on conflict-weight problems. Children using Models I or I1 will get these problems right even though they do not see them as conflict problems, whereas children using Model 111 will attend to the conflicting cues of weight and distance, but they will have to muddle through, and their resulting predictions

The Representution of Children’s Knowledge

69

TABLE 1 Predictions for Percentage of Correct Answers and Error Patterns on Posttest for Children Using Different Models Models

Problem type

Predicted developmental trend

1

11

111

IV

Balance

100

100

100

100

No change-all at high level

children

Weight

100

I00

loo

100

No change-all at high level

children

Distance

0 (Should say “balance”)

I00

100

100

Dramatic improvement with age

100

100

33 100 (Chance responding)

Decline with age Possible upturn in oldest group

Conflict-distance

0 (Should say “rightdown”)

0 (Should say “right d o w n ” )

33 loo (Chance responding)

Improvement with ape

Conflict-balance

0 (Should say “rightdown”)

0 (Should say ’* right-down”)

33 100 (Chance responding)

Improvement with age

w w

w w w

Conflict-weight

will be at a chance level of performance. Another prediction, shown in Table I , is that performance on distance problems should improve dramatically with age. The youngest subjects, using Model I will err on every problem, while children using Models 11, 111, or IV will never err. By a similar logic each of the problem types yields a predicted developmental course, the results of which are shown in Table I. (See Siegler, 1976, for a complete analysis.)

V.

Experiment 1: Assessing Initial Knowledge

The purpose of Experiment 1 was to assess the validity of the foregoing analysis for a group of children spanning a wide age range.

David Klahr and Robert S . Siegler

70

A.

METHOD

Subjects were 120 female !students from a private school in Pittsburgh. Fifteen students from each of eight grade levels were grouped as shown at the top of Table 11. Materials included a wooden balance scale, 10 different colored metal weights, and two wood blocks. The balance scale’s arm was 80 cm long, with four pegs on each side of the fulcrum. The first peg on each side was 7.6 cm from the fulcrum and each subsequent peg was 7.6 cm from the peg before it. The arm could swing freely from the point of attachment to the fulcrum, 10 cm above the fulcrum’s base. Each metal weight weighed 40 gm, measured 2.5 cm in diameter, and had a hole in its middle so that it would fit on the pegs; as many as six weights could be placed on any one peg. The two blocks of wood, each 11.4 cm high, could be placed under the arm of the balance scale to prevent it from moving regardless of the configuration of the metal weights on the pegs. Children’s knowledge was assessed through a 30-item test. On each problem the experimenter started with an empty balance, the arms of which were supported by the two wooden blocks. Then the metal weights were placed on the TABLE I1 Developmental Trends Observed and Predicted on Different Problem Types in Experiment I“ ~~~

Grade K-1st Age (years) 5-6 Number of 73 each type Mean age (months)

4th-5th 9-10 120

8th-9th 13-14 169

1 Ith-12th

16-17 207

Predicted developmental trend (from Table 1)

Problem type 4

Balance

94

99

99

100

No change-All at high level

children

4

Weight

88

98

98

98

No change-All at high level

children

4

Distance

9

78

81

95

Dramatic improvement with age

6

Conflict-weight

86

14

53

51

Decline with agePossible upturn for oldest

6

Conflict-distance

I1

32

48

SO

Improvement with age

6

Conflict-balance

7

17

26

40

Improvement with age

46

61

62

67

Weighted mean %

“Percentage of problems predicted correctly

The Represenrarion of Children’s Knowledge

71

pegs on the two sides of the balance scale, and the child was asked to predict which side would go down or whether the scale would balance if the two wooden blocks, underneath the arms of the balance, were not there. Among the 30 items were four balance, four weight, four distance, six conflict-weight, six conflictdistance, and six conflict-balance tasks of the types shown in Table I; they were presented in the same random order for each child. Children were tested individually in a quiet room in their school. The experimenter’s initial instructions were: Today we are going to play with this balance scale. The balance scale has these pieces of wood that are all the same distance from each other [pointing to the pegs] and these pieces of metal that all weigh the same.

At this point the children were encouraged to hold the weights to see that they weighed the same amount and to observe the equal distances between adjacent pegs. Children’s knowledge was then assessed by presenting them with the 30 problems described above. The problems were introduced with the following instructions: Let’s see what you know about the balance scale. I’ll put the weights on the pegs in different ways and you tell me whether this side would go down or this side would go down or they would both stay like they are now if I took the wood blocks away. The balance scale won’t actually move, but you tell me how the scale would go if the pieces of wood were not there.

Following this test, children were asked to explain their responses.’ Children spent between 15 and 30 minutes on the entire task. B.

RESULTS

The percent of correct predictions for each problem type by each age group is shown in Table 11. A 4 (age) by 6 (problem type) analysis of variance revealed that both main effects and their interaction were significant (D < .OOl). Note that the developmental patterns are very close to those predicted in Table I. In particular there is a dramatic improvement in distance problems and a decrement i n conflict-weight problems. The conflict-weight problems never did show an upturn, although performance appears to have leveled off for the older age groups. Not apparent in Table I1 is the substantial consistency that existed in performance on items within each problem type. Only on conflict-weight problems did accurate prediction decrease with age, and within this category such decrements ‘See Siegler (1976) for the criteria used to classify explanations.

72

David Kluhr and Robert S. Siegler

occurred on all six problems. The magnitude of the improvement over age on the four distance problems was unmatched by that on any of the 26 other items. On all eight of the balance and weight items, but on no other tasks, was the developmental trend minimal. With one class of exceptions, the four models make exact predictions about which of the three possible responses (left-down, right-down, balance) the subject will make on each one Qf the 30 problems. (The exception class contains the 18 conflict problems for Model 111; here the prediction is a lack of consistency, i.e., essentially chance responding.) Thus, we can compare the response pattern of each child to the predicted patterns for each of the models, and classify the child according to which, if any, model she was using in making her predictions. Using very strict criteria that had vanishingly small probabilities of misclassifying a random responder, it was possible to classify 107 of the 120 children. The results are shown in Table 111. Children’s explanations were also used to determine which model the child was using. The criteria for classifying according to explanations were derived from a literal interpretation of the models. Altogether, 117 of the 120 children’s explanations fit one of the four models. As shown in Table IV, the two classifications-one derived from children’s predictions, the other from their explanations-were highly correlated (r = .89,p < .001). All of the 23 children judged to be using Model I by the predictions data were judged as using Model 1 by the explanations criterion, and all eight of the children classified as using Model IV on the predictions measure-and only those eight-were classified as using Model IV on the explanations measure as well. On the other hand, many children were classified as using Model I1 by the predictions measure who were placed in Model 111 by the explanations measure. One interpretation of this discrepancy between the explanations and predictions criteria is that there were some children who used Model 111 tests, but consistently resolved the conflict by relying on the weight cue. Further evidence that children knew more about the balance scale than is revealed by the predictions classification comes from an analysis of the content of their explanations.

TABLE 111 Percentage of Children in Each Age Range Fitting Each Model

Age (years) 5-6

9-10 13-14 16-17 Total

( n = 30) ( n = 30) ( n = 30) ( n = 30) ( n = 120)

Model I

Model I1

Model 111

Model IV

77 10 10

0 30 23.3 20 18.3

0 40 56.1 63.3 40

6.1 3.3 16.7 6.1

0

24.2

0

Unclassified

23 13.3 6.1 0 10.8

13

The Represeritcition of Children’s Knowledge

TABLE 1V Number of Children in Experiment 1 Fitting Each Model-Predictions and Explanations Criteria

Model Classified by Explanations Criterion

I

I

Model Used: Classified by Predictions Criterion

I

I1

111

IV

0 0 0 8

I I1

23

1

0

0

1

111

0 0

7 13 0

IV

46 0

Fully one-third of the children advancing Model 111 explanations cited the ratio properly of conflict-balance problems (e.g., one on the third peg equals three on the first peg), but not the composition rule necessary for a Model IV placement. C. EVALUATION OF DECISION TREE REPRESENTATIONS

How well do the decision trees used in Fig. 1 represent children’s knowledge on the balance scale task? The representation appears to fare well on the first, second, and fourth criteria listed earlier. With regard to the first criterion, the models are clearly sufficient to account for the predictions data: The problem type by model analysis provided an exhaustive and unambiguous mapping between behavior and theory. With regard to the second criterion, the data can be analyzed at the level of either individual subjects or group averages, and also can be considered at either the level of response patterns or sheer number of correct responses. Regarding the fourth criterion, the formal relationship between the models is one of strict inclusion: Model I tests are included in Model 11, etc. This logical structure predicts an invariant developmental sequence (although we could not test this directly in a cross-sectional study). In terms of the third criterion-integration of psychological parameters-the merits of the representation are less clear. Although a number of implications could be drawn, few explicit statements about psychological, as opposed to logical, properties of the knowledge required to do the task have yet been advanced. Another orientation from which to evaluate the representation is to ask how well it answers the five questions about instruction and development that were listed above. Thus far, it has answered only Question 1: The difference between high and low performers is represented by differences among the four models. The models are silent on Question 2, which addresses the issue of alternative paths to the same performance. This inadequacy was most noticeable in the discussion of the idiosyncracies that are masked by the “muddle through” category on Model 111. Since the models do not have any representation for their own

David KIahr and Robert S . Siegler

14

induction, they are unable to !jay anything about Question 4, which asks about critical features of the instructional sequence. However, the models do suggest some straightforward ways to empirically investigate Question 3 (How difficult should an instructional sequence be?), and they imply that there should be no differences in responsiveness to instruction, thus providing an assertion that refutes the premise of Question 5 (When and why will differential learning occur?). In Experiment 2, these issues suggested by Questions 3 and 5 were addressed.

VI.

Experiment 2: Training on the Balance Scale Task

In the second experiment, 5- and 8-year-olds were equated for performing at a level not beyond Model I. Then they were provided with experience on either distance or conflict problems. or with one of two control procedures. Distance problem experience focused on the type of problems solvable by Rule I1 but not by Rule I; it thus was geared one step above the learners’ initial level. Conflict problem experience, emphasizing problems not understood even qualitatively until Rule 111, was intended to be two or more steps advanced. According to Piagetian theory, the fit between a child’s existing knowledge and the new information presented is a critical determinant of when, how much, and what kind of learning will occur (Piaget, 1971). Support for this view has been found by Turiel (1966) and Blatt (197 1) in the area of moral development, and by Kuhn (1972) in class-inclusion training. Therefore, we predicted that our Model I children would benefit from distance problems, while they would learn little, if anything, from conflict problems. As we already noted, there is nothing in the models that would predict differential responsivity to instruction of older and younger children. But both intuition and empirical evidence support the notion that older children are more adept than younger ones at mastering many novel problems on which taskspecific knowledge is equally lacking (cf. Siegler, 1975; Siegler & Liebert, 1974, 1975). Thus, there were no clear grounds on which to base a prediction about age differences in response to the training sequences. A.

METHOD

Experiment 2 included three segments: pretest, experience, and posttest.

I. Pretest The pretest consisted of eight items: two weight, two distance, two conflictweight, and two conflict-distance. The tasks and apparatus were similar to those used in Experiment 1; on each trial, the child was shown a configuration and

The Represenrution of Children's Knowledge

75

asked to predict which of the three possible outcomes would occur if the wood blocks were removed. There was no feedback during the pretest.

2 . Experience All experiential conditions except the bias control (see below) included 16 trials on which children were presented a randomized sequence of various types of balance scale problems. Children were asked to predict what would happen and why they thought so; then the wood blocks supporting the scale were removed so that the prediction was confirmed or disconfirmed. After a 10-second interval the weights were removed and placed on the scale in a different arrangement. Conflict problem experience involved presentation of six conflict-weight, six conflict-distance, one distance, two balance, and one weight problem. Distance problem experience included 12 distance, 2 balance, and 2 weight problems. Thus, each experiential condition included 12 problems of the type being emphasized; the additional four problems of other types were intended to prevent children from acquiring strategies too narrowly suited to the demands of the majority of items. Within the control condition there were two subgroups: the exposure control and the bias control. The exposure condition was designed to control for the possibility that any experience with the balance scale could improve performance; children in this condition were presented a sequence composed of 14 weight and 2 balance items that would familiarize them with the balance scale's workings but would not directly engender knowledge of' Models I1 or 111. However, this control procedure might itself bias children toward a greater reliance on Model I than if they had been left untutored. Therefore, a bias control was included in which children simply received the pretest and posttest. Within each age group's control condition, one-half of the children were assigned to the exposure control and one-half to the bias control. 3 . Posttest The posttest included a randomly ordered, no-feedback presentation of 24 items, four each of balance, weight, distance, conflict-weight, conflict-distance, and conflict-balance types. The pretest took approximately 10 minutes, the experience 25 minutes, and the posttest 15 minutes. Eight-year-olds were given the three parts in succession; 5-year-olds were given the pretest one day and the experience and posttest in a second session within the next 48 hours. 4 . Participants Sixty children, 30 5-year-olds and 30 8-year-olds, all with less than Model I1 knowledge, were randomly assigned within age and sex to the three treatment groups.2 All groups had equal numbers of males and females except for the 'See Siegler (1976) for the details of this selection procedure.

David Klahr and Robert S. Siegler

16

8-year-old control group, which included four boys and six girls. The mean CA of kindergartners was 70 months (range = 66-75 months), while the mean CA of third graders was 106 months (range = 101-1 17 months). The experimenter, a 22-year-old female research assistant, served for all children. B. RESULTS

Responses to the 24-item posttest were classified according to a scheme similar to the one used in Experiment 1. (There were no differences between the two control groups, so the data from both of them were combined.) As shown in Table V , 45 of the 60 children behaved according to the models: 21 using Model 1, 17 using Model 11, and 7 using Model 111. A Chi-square test indicated that significant differences were present in the type of rules used by children in the six age-by-experience groups (x' = 45.54, df = 1, p < .001). More specific analyses revealed that 5-year-olds more often used Model I and 8-year-olds more often Models I1 or 111 (x' = 12.91, df = 1, p < .001), and that children exposed to the control procedure more often used Model I, while those exposed to conflict or to distance problems more often used Model I1 or 111 (x2 = 13.20, df = 1, p <.001). An interactive relationship between type of experience and age was also apparent. Fisher Exact tests indicated that among 5-year-olds, experience with distance problems led to more adoptions of Models I1 and 111 than did experience with conflict problems or the control conditions (p < .01). As can be seen in Table V, the effect was almost exclusively to promote attainment of Model 11; no condition led to many children attaining Model 111. Among the 8-year-olds, however, both distance and conflict problem experience led to more adoptions of TABLE V Number of Children Using Different Models-Experiment Age group

Model I

Model 11

Model I11

2 Unclassifiable

~~

5-Year-olds Control Distance training Conflict training Total

8 3 5 16

8-Year-olds Control Distance training Conflict training Total Grand total

3 8 2 13 21

17

15

The Representntion of Children’s Knowledge

I1

Models I1 and 111 than did the control procedures (p < .001), and conflict problem experience led to greater use of Model I11 than did the distance problems and control conditions (p < .01). In summary, then, Table V shows that both age groups can learn from training that is only one level beyond their current level (i.e., distance training). However, given training that is two levels beyond (i.e., conflict training), the 5-year-old children learned nothing, while the 8-year-olds benefited substantially. Thus, it is clear that older and younger children derived different lessons from the same experience, even when they had identical initial predictive knowledge about the task.

VII.

Revised Representations for Balance Scale Knowledge

These empirical results raise questions that reveal some of the limitations of the decision tree representation used thus far to represent children’s knowledge of balance scale tasks. Since the four models purport to represent all of what a child knows about the task, they predict that children classified according to one of the models should be identical on all task-related performance, including learning about the task. Thus, they predict that the differential responsiveness to experience with conflict problems that we observed between the two age groups should not have occurred. Of course, the models make this prediction by default, since they have no representation of the learning process as such. That is, they contain no representation of the way that positive and negative information obtained during the training sequence is treated, nor about the ways in which the models might undergo transformation from one level to the next. Another limitation of the representation is that it allows no way to describe the many different means utilized by subjects to amve at the same end. We have already alluded to this in our discussion of the Model 111 explanations data, and now it is time to address it directly. We need a representation that can account for not only the logical form of the decision rules used to make predictions, but also the psychological properties of the rules. That is, we need a representation that enables us to clearly indicate the perceptual and mnemonic demands of actually using the decision rules. In this section we will introduce such a representation for children’s knowledge about this task, and we will present examples of the kinds of questions the representation enables us to ask. Then, in the next section, we will describe an experiment that provides some answers to these questions. A.

PRODUCTION SYSTEM REPRESENTATION

In Fig. 2 we have restated the four models of Fig. 1 as production systems. [See Newel1 (1973) for an extensive introduction, and Klahr (1976b) for some

David Klahr and Robert S . Siegler

78

Model

I P1: ((Same W ) --> (Say " b a l a n c e " ) ) P2: ( ( S i d e X more W) --> (Say "X down"))

Model I I

P1: ((Same W ) -->

(Say " b a l a n c e " ) ) P2: ( ( S i d e X more W ) --> (Say " X down")) P3: ( (Same W ) ( S i d e X more 0) --> (Say " X d o u n " ) )

M o d e l 111

P1: P2: P3: P4: P5:

(Sam@ W ) --> (Side (Same (Side (Side

(Say " b a l a n c e " ) )

X more W) --> (Say "X down")) W ) ( S i d e X more 0) --> (Say " X down")) X more W) ( S i d e X less D) --> muddle t h r o u g h ) X more W ) ( S i d e X more 0 ) --> (Say "x d o u n " ) )

Model I V

P1: ( (Same W ) --> (Say " b a I ance") 1 PZ: ( ( S i d e X more W ) --> (Say " X down") 1 P3: ( (Saiiie W ) ( S i d e X more 0) --> (Say " X doun") ) P 4 ' : ( ( S i d e X more W ) ( S i d e X less 0) --> ( g e t Torques)) P5: ( ( S i d e X more W) ( S i d e X more 0) --> (Say "x d o u n " ) ) P6: ((Same Torquo) --> (Say " b a l a n c e " ) ) P7: ( ( S i d e X more Torque) --> ( s a y " X d o u n " ) )

T r a n s i t i o n a l requirements Productions Opera t o r s

I -> I 1

add P3

I1

111

add P4, P5

I l l -> I V

motli f g P 4 ; add P6, P7

->

add d i s t a n c e e n c o d i n g and compar i son

add t o r q u e computat i o n and compar i son

Fig. 2 . Production system ( P ) representations for Models 1-W. D = distance; W = weight. Written in a special language called PSG. See text for further explanation.

examples from cognitive devellopment.] A production system consists of a set of rules-called productions-written in the form of condition-action pairs; the conditions are symbolic expressions for elements of knowledge that might be present at some instant. A production system operates via a recognize-act cycle. During the recognition cycle, all the condition sides of all the productions are compared with the current contents of the immediate knowledge state. We will refer to this immediate knowledge as the contents of working memory (WM). It can be interpreted as primary or short-term memory (Waugh & Norman, 1965), M-space (Pascual-Leone, 1970), short-term plus intermediate-term memory (Bower, 1975; Hunt, 1971), or more generally as the currently activated portion of long-term memory, or simply as the current state of awareness of the system.

The Representrition of Children’s Knowledge

79

The productions whose conditions are matched by elements in WM are placed into the conflict set, a conflict resolution principle is applied, and one production fires. The act cycle executes the actions that are associated with the fired production. Then the next recognition cycle commences. Thus, the conditions are tests on the momentary state of WM. A sequence of condition elements on the left side of a production is interpreted as a test for the simultaneous existence of the conjunction of the individual knowledge elements. If, for a given production, all the condition elements happen to be true at some instant, we say that the production is “satisfied.” If only one production is satisfied, then it “fires”: the actions associated with it, written to the right of the arrow (see Fig. 2 ) are taken. These actions can modify the knowledge state by adding, deleting, or changing existing elements in it, or they can correspond to interactions with the environment-either perceptual or motor. If more than one production is satisfied at a given moment, then the system needs to invoke some conflict resolution principle. In the systems shown here all conflicts are assumed to be resolved such that special cases have priority over general cases. For example, suppose that the two productions in the conflict set are: PI: (a b-+ x) P2:(b -+ y) P1 is a special case of P2, since P2 is satisfied whenever P1 is satisfied, but not vice versa. That is, P2 is satisfied when element b is in WM, but P1 is satisfied only when both b and a are present. The special case conflict resolution principle will choose P1. [Further discussion of conflict resolution i n production systems can be found in McDermott and Forgy (in press), Newell (1973), Newell and McDermott (1975), and Rychener (1976).] Consider, for example, Model I1 in Fig. 2 . It is a production system consisting of three productions. The condition elements in this system are all tests for sameness or difference in weight or distance. The actions all refer to behavioral responses. None of the models in Fig. 2 contain a representation for any finer grain knowledge, such as the actual amount of weight or distance, or the means used to encode that information. Nor is there any explicit representation of how the system actually produces the final verbal output. It is simply assumed that the system has access to encoded representations of the relational information stated in the conditions. We will return below to further consideration of the way that this information becomes available to the system. Returning to Model 11, notice that on any recognize cycle, only one production will fire. If the weights are unequal, then P2 will fire; if the weights are equal and the distances are not, then both P1 and P3 will be satisfied, but since P3 is a special case of P 1, the conflict resolution principle will choose P3 to fire; finally, if both weights and distances are equal, then only P1 will be satisfied and it will fire. (The numbers attached to

80

David Klahr and Robert S. Siegler

the productions [e.g., PI, P2, etc.] are not supposed to have any psychological meaning. They serve simply as labels for the reader; note that a production maintains its label across the four models.) We can compare the four models to determine the task facing a transition model. At the level of productions the requisite modifications are straightforward: a transition from Model I to Model I1 requires the addition of P3; from Models I1 to III, the addition of P4 and P5; and from Models 111 to IV, the addition of P6 and P7 and the modification of P4 to P4’. (This modification changes the action side from random muddling through to “get torques”.) We can compare the four models at a finer level of analysis by looking at the implicit requirements for encoding and comparing the important qualities in the environment. Model I tests for sameness or difference in weight. Thus, it requires an encoding process that either directly encodes relative weight, or encodes an absolute amount of each and then inputs those representations into a comparison process. Whatever the form of the comparison process, it must be able to produce not only a same-or-different symbol, but if there is a difference, it must be able to keep track of which side is greater. Model I1 requires the additional capacity to make these decisions about distance as well as weight. This might constitute a completely separate encoding and comparison system for distance representations, or it might be the same system except for the interface with the environment. Model 111 needs no additional operators at this level. Thus, it differs from Model I1 only in the way it utilizes information that is already accessible to Model 11. Model IV requires a much more powerful set of quantitative operators than any of the preceding models. In order to determine relative torque, it must first determine the absolute torque on each side of the scale, and this in turn requires exact numerical representation of weight and distance. In addition, the torque computation would require access to the necessary arithmetic production systems to actually do the sum of products calculations. Although we have compared the four models at two distinct levelsproductions and operators-the levels are not really that easily separated. Missing from these models is a set of productions which would indicate the interdependence: productions that explicitly determine which encoding the system will make. That is, in these models, there are almost no productions of the form: (want to compare weights) (attend to stimulus and notice weight). The sole exception to this occurs in P4’ in Model IV. When this model is confronted with a nonconflict problem, either P I , P2, P3, or P5 will fire on the first recognize cycle. However, if it is a conflict problem, then P4’ fires, and the system attempts to “get torques.” The result of this unmodeled action, as described above, would be to produce a knowledge element that could satisfy either P6 or W on the next recognize cycle.

-

The Representation of Children’s Knowledge

B.

81

EVALUATION OF THE PRODUCTION SYSTEM REPRESENTATION

Each of the four production system models in Fig. 2 makes precisely the same prediction as its counterpart i n the decision tree representation of Fig. I . Thus, on the first of the evaluative criteria listed above-accounting for behavior-the production system model fares as well as the decision tree model. With respect to the second criterion-multiple-level analysis-and the fourth-developmental tractability-the production systems are somewhat more explicit than the decision trees about the requirements for both the encoding operations and the rules (i.e., productions) that utilize the symbolic elements produced by the operators. They also clarify the developmental differences between models in terms of these two kinds of entities. The major advantage of the production system representation lies in its integration of general psychological principles-the third of our evaluative criteria. Production systems of the type used here incorporate a theory of the control structure and general representation that underlies a broad range of human problem-solving ability (Newell & Simon, 1972). As Newell (1973) put it: “The production system itself has become the carrier of the basic psychological assumptions-the system architecture of. . . [the production system] is taken to be the system architecture of the human information processing system” (p. 516). Thus, models written in this form can be viewed as variants within a general psychological theory, and to the extent that such a general theory is consistent with the empirical results from experimental psychology, then these models are also consistent with them. With respect to the five questions listed earlier, the production systems have enabled us to be very explicit about Question 1 (differences that underlie performance), and in particular about the important role of encoding operators. They have indicated some potential sources of variation for each level of performance (Question 2), although since they are written as modal types, this is merely suggestive at this point. Similar comparisons of the relative efficacy of the two forms of representations for answering the other three questions yield the same result. Thus, while the new representation does not provide much of an advantage over the old for understanding the results of Experiment 2, it does provide some guidance about where to look for an explanation-in the encoding of the stimulus. In order to model the conditions under which one or another aspect of the stimulus is attended to and encoded, we would need to augment the models in Fig. 2 with productions like P4’. These productions would transform the models from simple discrimination nets into active problem solvers, and they would enable us to make predictions about such things as eye movements and solution latencies for different classes of problems. However, before we can make such

82

David Klahr and Robert S. Siegler

an extension, we must first determine the varieties of possible encoding schemes that subjects are actually using. As a first step in that direction, we undertook a detailed examination of the problem-by-problem protocols of a few children in a training sequence. C.

PROTOCOL ANALYSIS

Several children, ranging in age from 5 to 10 years, were unsystematically selected to be run individually in a conflict training sequence. They were given instructions about the balance scale and about the fact that there were rules underlying the balance scale’s behavior that they could discover if they “watched carefully and thought about it.” In addition, following their prediction on each trial, they were asked to state their reasons for the prediction. Then the blocks were removed, the children observed the scale’s movement and if they were incorrect, they were again asked, “Why do you think that happened?” These entire sessions were videotaped, and then all the verbal comments, as well as major physical activities, were transcribed into the form shown in Appendix A. At the beginning of each problem, there is an indication of the problem number, the configuration, and the elapsed time (in minutes and seconds) since the start of the session. Problem numbers T I , T3, etc., correspond to items from the training sequence, and problem numbers E7, E8, etc. (see lines 11300 and 14700) are from an exploratory session which followed the training sequence. In the exploratory session, the children were edcouraged to build interesting problems or to explain to the experimenter what kinds of problems would achieve certain outcomes. The problem configuration is indicated by a numerical code that is a near-pictorial representation of the problem. In TI (line 00400) the code 0001/2000 indicates one weight on the first peg (from the fulcrum) on the left side, and two weights on the first peg on the right side. In T3 (line 02200), the code 0100/1OOO indicates a single weight on the third peg on the left, and a single weight on the first peg on the right. Excerpts from the protocol of Lisa, a 5-year-old female, are shown in Appendix A. The protocol provides a rich data source from which to select “observations.” However, in this discussion we will focus only on those aspects that indicate the kind of encoding of distance and weight that Lisa appears to use. Lisa was first given the standard instructions and pretest described earlier. Her response pattern did not conform to any of the four models. However, if Model I were modified such that heavy things went up instead of down, then she was a perfect Model I subject. The first problem in the training sequence confirms this interpretation (lines 00400-01200). Lisa knows which side has more weight, but her prediction is based upon the assumption that more weight goes up. However, when confronted with the contrary evidence, she changes the “sign” of the correlation between weight and direction of tipping. This single-feedback trial

The Representation of Children’s Knowledge

83

was sufficient: For the remainder of this half hour session, she never again errs in her understanding of the direction of the effect of weight differences. As we will see, the correct encoding of distance and its effect required a much longer series of trials. The second training problem (not shown) was a balance problem, so T3 (0100/1000) was the first instance in which Lisa received feedback indicating that equality of weight is not a reliable predictor. Her own verbalization of the problem captures her puzzlement: “Well why are they both the same thing [same weight] and one’s up and one’s down?” (line 3300). Another distance problem followed immediately (T4: 0020/0020), and Lisa’s first response is to say balance, but she quickly corrects herself, having detected the distance difference. Her encoding of distance is correct in that it is based on the fdcrum, rather than the end points, as the zero reference point (lines 0440004500). However, she incorrectly associates greater distance with the side that goes up rather than the opposite, in the same way that she initially had the sign wrong for weight effects. This is her first attempt to utilize distance information, and she gets negative feedback. At this point she might abandon distance as a useful cue, or she might-as she did with weight-simply change the sign of the relation. As we will see, she does neither. T5 was a complex distance problem (0101/1100) (not shown), and T6 (0102/ 2010) a balance problem, neither one of which yielded a useful protocol. In T7 (0200/2000) we return to a distance problem. It is clew from the protocol that Lisa is still attempting to use distance (lines 09600-10000). She still encodes direction of distance from the f u l m m correctly, but she has not changed her erroneous assumption about the effect of this difference. Note also that she has not yet made any statement about absolute amount of distance; all her statements are about relative distance. In order to focus on the issue of distance encoding, we skip over about 15 minutes of conflict training in which the problems were mainly complex conflict-weight and conflict-distance (i .e., two or more pegs occupied on each side) from which no clear pattern emerged. We pick up the protocol again in an excerpt from the exploratory phase in which Lisa was allowed to construct problems according to various experimenter requests or hints. In E7 (lines 11300-14400), she has been asked to construct some problems such that she will not be quite sure what the result will be. In general, Lisa does no such thing, and instead tends to construct problems about which she is very confident. Thus, her initial configuration is 0003/0004, a problem in which both weight and distance indicate that the right side will go down. Then the experimenter modifies it to a distance problem (0004/0004), and Lisa apparently forgets all about distance differences, reverting to a Model I prediction of “balance” (lines 12000-12300). With a little prompting from the experimenter (lines 13500-14000), she invokes a (post hoc) distance explanation (lines 14100-14200). Notice that the distance

84

David Klahr and Robert S . Siegler

description is not just a relative judgment, but instead is stated in terms of two absolute (albeit approximate) quantities. It appears that, even after almost 30 minutes of experience with the balance scale, Lisa knows that distance is an important factor, but she has not yet developed a reliable rule about the effect of distance differences. Then, over the next 2-minute period, she begins to demonstrate a stabilizing grasp of this concept. First she creates a balance problem and makes the correct prediction (lines 14900-15900). Then a new experimenter enters, and feigning ignorance, asks how the scale works. Lisa creates (0003/0003) and predicts correctly, and for the right reasons (lines 16800-17300). Then, at the experimenter’s request, she correctly creates a balance problem (0003/3000). It is interesting that she does this in the “easiest” way, given the configuration from which she was starting, but it is also the case that this is the same balance configuration that was used in the preceding problem. Then she creates a distance problem such that the scale tips in a desired direction (19000-19400) and gives the correct explanation and, finally, she initiates yet another balance problem, one unlike any she has ever seen before (3000/0003). Recall that this protocol analysis was undertaken after a discussion of the production system representation of knowledge about the balance scale (Fig. 2). In that representation, we tried to emphasize the differences between the encoding of information about the environment (the undefined operators) and the combination rules [cf. Gelman’s (1972a, 1972b) operator+stimator distinction, and Klahr and Wallace’s (1973) operator-rule dichotomies] for acting on that information (the productions). The protocols tell us something about the nature of the representations that are being used by the child, and hence something about the encoding operators that produce them. It is clear that Lisa extracts information from the training series that will enable her to improve both the encoding operators and the combination rules. With respect to weight, she has no difficulty in formulating an appropriate encoding based on counting the number of weights. Although there is an initial error with respect to the relation between weight differences and the direction of the scale, this is quickly corrected and remains stable for the rest of lhe session. Distance encoding follows quite a different course. Initially it is ignored. Then differences i n distance are noted, but their effect is quite unstable in the face of negative feedback, and as we saw, they are occasionally ignored well into the training sequence. However, it appears that by the very end of the exploratory trials, an appropriate encoding of distance, and a concomitandy appropriate rule for utilizing it (at least on disunce problems), has been formulated. Learning about the balance scale, then, would seem to require much more than is suggested by a comparison of adjacent models in Fig. 1. The production system representation of Fig. 2 has enabled us to make explicit the difference between encoding operators and decision rules, and it has guided our search for

The Representtrtion of Children‘s Knowledge

85

instances of both of these kinds of learning in the protocol. The analysis suggested that there is a point in the development of knowledge about this task during which the dimensions may be encoded in idiosyncratically incorrect ways, and that the form of the encoding may depend upon trial-to-trial feedback. In the next section we will introduce a model that attempts to capture these phenomena for an individual subject. D.

REVISED PRODUCTION SYSTEM FOR A MODEL III CHILD

Thus far, the production system representation has been used only to suggest some of the complexities of learning about the task. In this section we will work toward the creation of a production system model of a single child’s behavior during a training sequence. The representation will be more than suggestive, for it will be specific enough to run as a computer simulation. The simulation will serve two purposes. First, it will demonstrate the sufficiency of the model to account for the data it purports to explain. Second, the particular simulation language in which the model is stated is based upon, and incorporates in its structure, very specific assumptions about the nature of the human informationprocessing system. Thus, the model to be described here is a particular instance of a much broader theory of human problem solving. Our subject, Jan, was a female second-grader, age 7 years, 11 months. Her performance on an 8-item pretest and a 16-item training series is shown in Table VI. In Table VI, each row corresponds to a problem. The columns indicate, respectively, problem number, problem configuration, problem type (Distance (D), Balance (B), Conflict-Weight (CW), Weight (W), etc.), Jan’s response (Left (L)- or Right (R)-down, Balance (B)), feedback from the scale (if the subject’s prediction was inconsistent with what the scale did, it is indicated by a -), predictions from three of the previously described models (Models IV, 11, and I), and finally, two columns corresponding to the model to be described in this section. The first of these columns-IIIA-contains the model’s prediction, and the second contains the value of a variable criterion that is used to make the prediction. For example, Problem 7 has three weights on the first peg on the left and two weights on the third peg on the right; it is a conflict-distance problem. Jan predicted that the left side would go down, but as Model IV (which is always correct) predicted, the right side went down so the subject got negative feedback. The other three models shown here (Models 11, I , and IIIA) all make the same prediction as the subject: left-down. The number at the bottom of each of the four model columns shows the number of mismatches between Jan’s predictions and the model’s. Jan’s responses to the pretest make her a perfect Model I1 subject. Her responses during the training sequence provide a poor fit to Models I, 11, and IV. Recall that the criterion for fitting Model I11 was that the responses be essentially

David KIahr and Robert S.Siegler

86

TABLE VI Jan on Training Sequence, and Predictions from Four Models

Problem

Prediction

Number Configuration Type S2 Feedback Model IV Model I1 Model I Pretest I 2 3 4 5 6 7 8

100~100

0 I01300 1001200 010~020 0201002 2001400 1001200 030(020

Training series 1 020010200 2 002010200 3 002013000 4 000310100 5 0200(0400 6 0102120I0 7 000310020 8 0100~0200 9 0040~1020 II= 000112000 12 001311020 13 012012200 14 020011300 15 0002~0010 16 002311110

D CW CD W D CD CD W

D

B CD

cw cw B CD

cw cw W CD CD

cw CD CW

Model IIIA Criterion

L R

R R R R R L

L B R L L B L L L R L L L R R

B B

R R L R R L R R L

L B R L R B L R L R L R R L L

70

6b

76

L B L L R B

“Problem 10 was omitted. Abbreviations: B = balance; C == conflict; D = distance; W ONumber of mismatches between Jan and model.

=

L B R L L B L L L R L L L R L

R L R B L R L R L R R L L

weight; R

W

D W D W

D W

Ib

=

right; L

=

left.

random for conflict problems. Thus, although the “muddle through” prediction of Model III does not make: an exact prediction on any trial, it predicts the absence of a consistent pattern over the set of conflict problems. And indeed, this is what we find in Table VI: On 5 of the 11 conflict problems Jan responds as if she were relying on the weight cue, and on the other six she conforms to the distance cue. Thus, we could simply classify Jan as a Model 111 subject and leave it at that. Such an interpretation has several deficiencies. First, the classification scheme itself is unsatisfactory when compared with the others. Model I11 subjects get so

The Represenrution of Children’s Knowledge

87

classified as a residual category, by the absence of any pattern in their responses to conflict problems, while all other classification is based on the Occurrence of things that were predicted to happen, rather than the absence of things that should not. In addition to this “taxonomic” weakness, Model 111’s “muddle through” prediction tells us nothing about the psychological processes that actually operate when subjects detect conflict but do not yet know how to deal with it correctly. We have already cited some of the idiosyncratic strategies that different subjects bring to bear on this situation. Finally, it is important to emphasize that Table V1 represents responses during a training sequence, a situation in which the child was presumably attempting to integrate the feedback from the balance scale’s actual behavior with her current hypothesis about how it worked. None of the four models described thus far have any mechanism to represent and utilize such information. Thus, the model to be described represents our first steps toward remedying these deficiencies. Jan was run under the same conditions as Lisa, and an analysis of her trial-bytrial explanations provided the initial evidence for the model that we eventually formulated. The most striking feature of her comments was the way she appeared to represent distance and weight on conflict problems. Both of them were treated as dichotomous: More than two weights was treated as “big,” otherwise weight was “little,” and if the third or fourth peg were occupied, then distance was “big,” otherwise it was “little.” Rather than present another lengthy protocol analysis here, we will show just two examples of this dichotomous encoding of distance. On Problem 12 (0013/1020), the child predicts left-down; upon seeing the result, she says: Oh, now I think I know why. . . . I think I know because. . . it’s supposed to be a rule that they usually go down more if they’re on that side [pointing to the extreme right of the balance scale]. So that one went down cause it’s two there [pointing far right] and none there [pointing far left].

If we encode each arm of the balance scale into a near segment (Pegs 1 and 2) and a far segment (Pegs 3 and 4), then this protocol is easily interpreted. “They usually go down more if they’re on that side” means that if the far segment is occupied (“big distance”) then the scale will tip in that direction. “Two there and none there” means that the far segment on the right is occupied by two weights, whereas the far segment on the left is unoccupied. The second example comes from Problem 14 (0200/1300), just before the child gets feedback. She says: This side’s gonna g o down [pointing left] . . . Even though this one has four [pointing right] and this one only has two [pointing left]. . . . Even though this one has [pointing right] twice as much as this [pointing left], that means that because this one’s more [waves to far left] over, and that’s [pointing right] all on that side.

88

David Klahr and Robert S . Siegler

In this case, we garner support for the dichotomous distance encoding from the comment that the weights on the right arm of the scale are “all on that side.” “That side” of what? By our interpretation, they are on “that side” of the midpoint of the right arm, thus making distance “little,” rather than “big” on the right. In order to determine whether this interpretation of the protocols is valid, we need to construct a model that is consistent with Jan’s actual predictions on each trial, as well as her explanations. Based upon many such comments and our interpretations of them, we constructed the model whose predictions are shown in Table VI. In order to provide a clear overview of the model we will describe it first in terms of a binary decision tree, plus a few ad hoc mechanisms. Then we will present a running production system for a more complete model based on the same underlying logic. Figure 3 shows the binary decision tree representation for Model IIIA; Jan’s performance on the training sequence is shown in Table VI. The numbers under the terminal nodes correspond to the problems from Table VI that are sorted to those nodes. The first three tests are the same as those in Model I11 (Fig. l ) , and they account for balance, weight, and distance problems. If neither weight nor distance is “same,” then the model begins to test for “big” values. If either

Same ?

8,13.15 : I n i t i a l l y : Weight

Down 3.4

Down 7 , 9 , 1 2 , 1 6 (Weight) 5,14 (Distance1

W e i g h t : O n Any Single P e g , n Z 3

Weight -Distance

Or Distance -Weight

Fig. 3. Decision tree representalionfor Jan’sprediction model. D = distance; W = weight. The numbers under the terminal nodes correspond to the problems from Table VI that are sorted to those nodes.

weight or distance-but not both-is big, then the side with the big value determines the prediction. If both are big. then Model IIIA favors whichever one is currently its criterion value. The criterion value starts as weight, but whenever negative feedback is received the criterion switches from one value to the other. The state of the criterion value is indicated in the last column in Table V1. Note that it changes after any negative feedback, not just on conflict trials with negative feedback. (The terminal node labeled “?” in Fig. 3 is never reached by the set of problems in Table VI. Such a problem would be a conflict problem with neither weight nor distance “big.” We have no evidence upon which to base a prediction about what the subject would do with such a problem.) E.

PRODUCTION SYSTEM FOR MODEL IlIA

The production system for Jan is shown i n Fig. 4. The representation contains the actual computer listing (with a few inessential details not shown) for the production system, which is written in a special language called PSG (Newell & McDermott, 1975). Appendix B contains a trace of this model running on a sequence of four problems from Table VI; one of them-Problem 7-is also shown in Fig. 5. Before we embark on a detaiied description of the model, we will make a few comments about the properties of this rather complex representation of knowledge about the balance scale task. This model represents, in addition to the child’s knowledge about how the balance scale operates, her knowledge about the immediate experimental context in which she is functioning. The trial-by-trial cycle during the training phase comprises (1) observation of the static display, (2) prediction of the outcome, (3) observation of the outcome, (4) comparison of the outcome with the prediction, and (5) revision if necessary, of the criterion. The production systems shown previously (Fig. 2) represented knowledge sufficient to execute only the second of these five steps, while the present model (Fig. 4) explicitly represents all of this task-relevant knowledge in a homogeneous and integrated manner. This model utilizes, in one way or another, representations of knowledge about when and how to encode the environment, which side has more weight or distance, which side has a big weight or distance, what the current criterion value is, what the scale is expected to do, what the scale actually did, whether the prediction is yet to be made or has been made, and whether it is correct or incorrect. F . PRODUCTION SYSTEM INTERPRETATlON

Some general properties of production systems were described earlier. In this section we will add a few more details about how the model in Fig. 4 operates. Recall that the basic cycle for a production system is recognize-act. During a recognition cycle, all the productions compare their condition elements with an

90

David Klahr and Robert S . Siegler

:(CLASS weight distance) <side. I>:(CLASS left right both)

:(CLASS weight distance) <side,2>:(CLASS left right both)

:(CLASS u p down level)

Pl:((predict) (weight same) --> (made **) (expect both level) say.b) P2:((predict) (weight more <side.l>) --> (made **) (expect <side.l> down) say.d) P3:((prcdict) (weight same) (distance more <side.l>) --> (made **) (expect <side.l> down) say.d) P4:((predict) (weight morc)(distance more) -->

find.bie)

P5:((prcclict) (criterion )(big <side. I > ) ( big <side.2>) --> (made *t) (expect <side.l> d0wn)say.d) PG:((predict) (weight big <side.l>) --> (made **) (expect <side.l> down) say.d) P7:((prechct) (distance big <side.l>) --> (made **) (expect <side.l> down) say.d) P8:((predict)() abs --r ATTEND)

El:((expcct) --> look) E2:((cxpcct <side.l> )(see <side.l> ) --> (did **)(see ===> saw)(result correct)) E3:((expect <side.l> )(see <side.l> ) abs (see) --> (did **)(see =-=> saw) (result wrong))

weight)) SW1 :((result wrong)(criterion distance) --> (old **)(distance ==a SW2:((rcsult wrong)(criterion weight) --> (old *:)(weight

-==> distance))

SW3:((result corrcct)(criterion) --> (old * t ) )

find.big:(OPR CALL) ;returns (weightldistance big leftlright), one or two such. looh:(OPR CALL) ; looks for result of balance tipping; returns (see leftlright down) attend:(OPR CALL) : initial encotling. of same or difference on distance g! weight; returns (wcightldistance samclmoro Icftlright)

Fig. 4 . Production system ( P )for Jan. ABS = Absent. Written in a special language called PSG. See text for further explanation.

ordered list of elements in VlM.The trace in Fig. 5 shows the state of WM after each cycle. For example, at the beginning of the second cycle in Fig. 5 , we see that WM has four elements in it: (DISTANCE MORE RIGHT), (WEIGHT MORE LEFT), (PREDICT), and (CRITERION WEIGHT). An examination of the productions in Fig. 4 reveals that P1 is the only production whose condition elements are completely matched by working memory elements, so in this case, it fires. We can interpret

The Representation of Children’s Knowledge

91

(030310020) Cycle 1 WM: ((PREDICT) (CRITERION WEIGHT)) F i r e P8: ((PREDICT) () AES --> ATTEND) Output from ATTEND (input t o WM) ::> (weight more lett)(distance more right) Cycle 2 WM: ((DISTANCE MORE RIGHT) (WEIGHT MORE LEFT) (PREDICT) (CRITERION WEIGHT) ) CONFL1CT.SET: (P2 P4) F i r e P4: ((PREDICT) (WEIGHT MORE) (DISTANCE MORE) --> FIND.BIG) Output from FINl3.01G (input t o WM) ::> (distance big right)(weight big left) Cycle 3 WM: ((WEIGHT BIG LEFT) (DISTANCE BIG RIGHT) (PREDICT) (WEIGHT MORE LEFT) (DISTANCE MORE RIGHT) (CRITERION WEIGHT)) COIJFLICT.SET: (P2 P4 P5 P6 P7) CONFC1CT.SET: (P4 P5) AFTER SPECIAL.CASE.ORDER CONrLICT.SET: (P5) AFTER WM.ORDER F i r e P5: ((PREDICT) (CRITERION ~DIMENSION.l>)( BIG <SIDE.l>) ( BIG CSIDE.2.) --> (MADE 4 4 ) (EXPECT <SIDE.l> DOWN) SAY.0)

w#**w*** LEFT d o w n Cycle 4 WM: ((EXPECT LEFT DOWN) (MADE (PREDICT)) (CRITERION WEIGHT) (WEIGHT BIG LEFT) (DISTANCE BIG RIGHT) (WEIGHT MORE LEFT) (DISTANCE MORE RIGHT)) F i r e E l : ((EXPECT) --> LOOK) Output f r o m LOOK (input to WM) ::> (see right down) Cycle 5 WM: ((SEE RIGHT DOWN) (EXPECT LEFT DOWN) (MADE (PREDICT)) (CRITERION WEIGHT) (WEIGHT BIG LEFT) (DISTANCE BIG RIGHT) (WEIGHT MORE LEFT) (DISTANCE MORE RIGHT)) CONFCICT.SET: (E1,E3) F i r e E3: ((EXPECT -4IDE.1, ) (SEE <SIDE.l> (DID **) (SEE ===> SAW) (RESULT WRONG))

Cycle 6 WM: ((RESULT WRONG) (DID (EXPECT LEFT DOWN)) (SAW RIGHT DOWN) (MADE (PREDICT)) (CRITERION WEIGI-IT) (WEIGIdT BIG LEFT) (DISTANCE BIG RIGHT) (WtIGHT MORE LEFT) (DISTANCE MORE RIGHT)) F i r e SW2: ((RESULT WRONG) (CRITERION WEIGHT) --> (OLD **) (WEIGHT ===> DISTANCE)) Cycle 7 WM: ((OLD (RESULT WRONG)) (CRITCRION DISTANCE) (DID (EXPECT LEFT DOWN)) (SAW RIGHT DOWN) (MADE (PREDICT)) (WfIGI4T BIG LEFT) (DISTANCE BIG RIGHT) (WEIGHT MORE LEFT) (DISTANCE MORC R l i l i T ) )

Fig. 5 . Trace of Jan’s production system ( P ) running on a conflict-distance problem. ABS = Absent; WM = working memory, Written in a special language called PSG. See text f o r further explanation.

92

D n i d Klnhr. and Robert S. Siegler-

a production P:(ABC-+DE) as “If you know A and B and c (i.e., if they are currently in WM, in any order), then do actions D and E . ” There are two conflict resolution principles. The first one to be applied, special case order, has already been described. If, after applying special case order, there are still two or more productions in the conflict set, then a second resolution principle, WM order, is applied. This principle chooses the productions with the frontmost element in WM. New information always enters the “front” of WM, pushing all else down a “notch.” Furthermore, when a production fires, its evoking elements are moved to the front of WM (automatic rehearsal). Thus, the WM order conflict resolution principle says, in effect, “when in doubt, respond to the most recently important information.”3 There are several different types of actions: 1 . WM additions. These isimply add new elements to the front of WM. For example, if E3 fired, (result wrong) would be added to the front of WM. Other sources of new information are the encoding operators (described below). 2. WM modifications. Elements in WM can be altered directly. The action (A -+ B) changes symbol A to symbol B in the second element in WM. The action (x**) changes the first element in working memory from A to (X (A)), [e.g., (OLD**) would change (DOG) to ( O L D (DOG))]. 3 . Output. These actions are surrogates for action on the external environment. The only ones used here are say.b (say “balance”) and say.d (say “left [or right] down”).

1. Description of Model (Fig. 4 ) There are three major functional groups of productions.

a . Pn. These correspond to the major nodes in the decision tree representation. PI-P4 are essentially the same as Pl-P4 in Fig. 2; P5, P6, and P7 correspond to the tests for “big” values in Fig. 3. Some of the productions use variables that can be matched by specific values in WM elements. These variables are defined in the first three lines of Fig. 4 in terms of the members of the class on values that the variable can take on. Thus, and can take on the values “weight” or “distance”; <side. 1> and <side.2> can take on the values of “left,” “right,” or “both”; and cdirection> can take on the value “up,” “down,” or “level.” b. En. These control the model’s viewing of the balance scale after it tips, and compare what it expectled to see with what it actually sees.

c . SWn. These change the criterion whenever the system determines (via the E productions) that it has made an incorrect prediction. There are three ’In Fig. 5 and Appendix B , special case order is usually adequate to resolve conflicts. All instances in which WM order is also used are explicitly indicated in the trace.

The Representution of’Children’s Knowledge

93

encoding operators. None are modeled, but their conditions of evocation are explicit, as is the form of the encoding they produce.

d . Attend. “Attend” does initial encoding of weight and distance. This operator can detect sameness or difference of weight or distance and can indicate the side on which weight or distance is greater. Thus, it is only an encoding of relative quantity. The model assumes that i n the first instance this is all that is encoded. e . Find.big. “Find.big” encodes big weight or big distance and side on which they occur (if they occur).

f. Look.

“Look” encodes direction of tipping of scale.

2 . Dynamics of the Model The general procedure is as follows. First weight and distance differences, if any, are encoded. If there is no conflict, then a prediction is made, an expectation is formed, and the scale’s actual behavior is observed. If it is inconsistent with the prediction, then the criterion is changed. If initial encoding reveals no clear prediction, then a second encoding is effected, this time in terms of big distance or weight. Then the rest of the process follows exactly as in the case of a single encoding. Figure 5 contains a trace of the model working on one of the problems from Appendix B. The trace shows the state of working memory at the start of each cycle, as well as which production fired. Conflicts are shown when they occur, as are the results of the encoding operators. The system starts with an element in WM (PREDICT) indicating that it has a goal of making a prediction, and another element representing the current value of the criterion. Since there is no element representing weight or distance, the only production whose conditions are completely satisfied is P8, which tests for (PREDICT) and the absence (ABS) of a weight or distance element (DISTANCE. I ) . ATTEND, P8’s only action, is an encoding operator that is modeled only up to the point of its inpudoutput specifications. In this case the input is presumed to be the physical arrangement of disks on pegs in the configuration (0003/0020), and the outputs, as shown in the trace, are two comparative symbols indicating more weight on the left and more distance on the right. They are directly provided by the model builder. Thus, at the beginning of Cycle 2, WM contains four elements, and these elements satisfy both P4 and P2 (see Fig. 4). P4 is a special case of P2, so it fires. It recognizes that neither weights nor distances are equal, so it attempts a second encoding (FIND.BIG) to determine some absolute amounts of distance andor weight. Once again, an unmodeled encoding operator is assumed to produce two elements, indicating a big distance on the right and a big weight on the left. The results are shown at the start of the third cycle.

94

Datmid Klahr and Robert S . Siegler

Five productions are satisfied by the elements now in WM. P2 and P4 are still satisfied since none of the elements that satisfied them on the previous cycle have been changed. P5, P6, and P 7 are satisfied because they test for either big weight or big distance. Since P4 is a special case of P2, and P5 of P6 and P7, the special case order principle leaves P4 and P5 in the conflict set. But the elements that match P5 are newer than those that match P4, so WM order selects P5 to fire. P5 matches whatever the current value of the criterion is (in this case, it is weight) with the corresponding “big” element (in this case [WEIGHT BIG LEFT]) and then uses the value of the directional variable (LEFT) to form its expectation (EXCEPT LEFT DOWN) and to “say” its prediction. What the system knows at this particular momeilt is revealed by the contents of WM at the start of the fourth cycle. It knows that: It expects the left side to go clown (EXPECT LEFI DOWN);

It already made a prediction (MADE (PREDICT)); The current criterion is weight (CRITERION WEIGHT); And it knows the enccdings (WEIGHT BIG LEFT) (DISTANCE BIG LEFT), and (DISTANCE MORE RIGHT).

RIGHT), (WEIGHT MORE

The rest of the trace is straightforward. During Cycle 4, the system seeks an encoding of what the scale actually did, and it sees that the right side went down. On Cycle 5, it recognizes that what it saw is discrepant with what it expected (E3), so it knows that it got the problem wrong. Finally, on Cycle 6, it recognizes that it was wrong while using the weight criterion, so it changes it to distance. G . EVALUATION OF REPRESENTATIONS FOR JAN’S KNOWLEDGE

The decision tree in Fig. 3 and the production system in Fig. 4 are logically equivalent: Both account for all but the last of Jan’s predictions during the training series. As described above, they differ from the representations in Figs. 1 and 2 in that they model that subject’s response to feedback, and because they both represent idiosyncratic encodings of the stimulus. Thus, both models have certain advantages over the previous ones. However, the models are not equivalent in all respects, and the psychological properties of the production system-properties previously just alluded to-can now be clarified. The production system, since it embodies a general model of the human-information prccessing system, forces us to form very explicit hypotheses about things that the decision tree lets us finesse. There is no separation of control information from data in a production system. Every relevant piece of information is explicitly represented in WM, and all task-specific knowledge for acting on that information is represented by productions. As indicated by the final list of elements, we are postulating a sizable amount of

The Representation of Children’s Knowledge

95

material floating around in WM. It is clear that the size of WM is well beyond the estimated short-term memory capacity of from seven (Miller, 1956) to as little as three or four (Broadbent, 1975) items, or the “M-space” estimates (PascualLeone, 1970) in the same range. However, it is unclear how a system that did not have immediate access to all of this momentary knowledge could ever d o the task. Questions about the amount of control information sufficient to perform the task are not addressed by the decision tree representation. For all their emphasis on the importance of the outputs from the encoding operators, however, the production system models do not describe the encoding process itself. Neither do they indicate precisely what sort of encoding deficit might affect response to instruction. A remedy to the former limitations would take the form of a model of encoding, and we leave that for future investigation. The second issue, that of the nature and effect of encoding deficits, is directly related to Questions 4 and 5 of our initial set. The specific questions are (1) whether encoding deficits are in fact typical of the younger children, and (2) if such deficits exist, whether they account for the younger children’s inability to benefit from instruction on conflict problems. These questions were investigated in Experiment 3.

VIII. Experiment 3: Encoding Hypothesis Recall that the results of Experiment 2 indicated that older and younger children, equated for initial task-specific knowledge about the balance scale, responded quite differently to the training sequences. This finding motivated a shift in the representation and in the level, or grain, of our analysis of what was going on during training. Lisa’s protocol analysis revealed her difficulty in determining the appropriate encoding of the two relevant dimensions, and the analysis of Jan’s responses during training led to a production system which incorporated two levels of encoding-one relative, one absolute (bighot big)-for both dimensions. Analysis of other protocols revealed many such stimulus misencodings. This, together with the sizable literature on the development of attentional strategies (cf. Pick, Frankel, & Hess, 1975; Zeaman & House, 1963) suggested to us that differential encoding might be the cause of the differential responsiveness to instruction. Siegler (1976) described three steps that are necessary to test this hypothesis rigorously. ( 1) Assess encoding independently of predictive performance and establish the hypothesized encoding differences. (2) Show that the appropriate manipulation can eliminate or at least reduce encoding differences. (3) Demonelimistrate that when the difference on the explanatory variable-encoding-is nated, the initially observed difference on the to-be-explained variableresponsiveness to instruction-is also eliminated. In summary, then, the goal

David Klahr and Robert S. Siegler

96

was to show that i n a group of older and younger children who were all using Model I initially, there would be a consistent encoding deficit in the younger children, then to eliminate this deficit, and finally to expose both groups to the training sequence and to produce identical learning in both age groups. Attempting to do this at the fine-grained level of the preceding section would have led to a mass of detailed variation that would make it very difficult to verify the general properties of encoding differences; it also would have been prohibitively expensive in terms of time and effort. Therefore, in this section, we move back up to a more aggregated level of analyses. A.

RECONSTRUCTION PARADIGM

Chase and Simon (1973) utilized a reconstruction paradigm to study the differential ability of chess masters and nonmasters to extract meaningful information from briefly presented board configurations. This procedure suggested to us a means by which differences between older and younger children’s encoding of balance scale configurations could be assessed independently of their predictions about the effect of these configurations on the scale’s behavior. In the third experiment in this series, 5- and 8-year-old children were presented with various configurations of weights on a balance scale for a few seconds (the scale was not free to tip). Then the scale was removed from view, and the children were required to reconstruct the initial configuration as accurately as possible on an empty scale. Note that this procedure allowed independent assessment of encoding on both weight and distance dimensions. For example, when given an initial configuration (0300/0200) the child might reconstruct it, for example, as (0300/0200), or (0030/2000), or (0200/0100), or (0010/0003), revealing, respectively, no mi,sencoding, distance only misencoding, weight only misencoding, and both weight and distance misencodings. Our protocol analyses led us to expect that the older children would be accurate on both dimensions, while the younger children would do well on weight but poorly on distance.

1. Basic Procedure The same basic procedure was followed in all phases of Experiment 3, and the full details are given in Siegler (1976). Here we will only describe the major features. Overall, 40 kindergartners (“5-year-olds”) and 30 third-graders (“8year-olds”) from two public schools in Pittsburgh participated in Experiment 3. Two identical balance scales were used. They were slightly different from the one used previously, having seven rather than four pegs on each side of the fulcrum, and having a built-in lever rather than wood blocks to keep the scale from tipping until the experimenter released it. A large Styrofoam board was used to hide one of the balance scales during the reconstruction phases.

The Representiirron of Children’s Knowledge

97

The encoding test included 16 problems, on each of which there were from three to five weights on each side, all located on either the third, fourth, or fifth peg from the fulcrum. On any given problem, only one peg on each side was occupied. Children were tested individually in a vacant room in their school. Each child was presented the encoding test first, and then presented the same 24-item predictions task (without feedback) used in the Experiment 2 posttest. For the encoding test, the children were told: The idea of the first game is for you to look how the weights are set on the pegs on my balance scale and then make the same problem by putting the weights on the pegs on yours. First 1’11 put the weights on the pegs on my scale. You should watch closely to see how the weights are set on the pegs. Then I’ll put the Styrofoam board hack up so you can’t see my scale. You will then need to put the weights on the pegs on your scale in the same way that you saw them on my scale. Just put the weights on the pegs so it’s just like the problem you saw on my scale.

After the first trial, children were again told, “Remember, you should watch closely to see how the weights are on the pegs on my scale so that you can put the weights on your scale in the same way.” Children were allowed 10 seconds to observe the initial configurations, and then they were allowed to reconstruct the arrangement immediately on the other scale. There was no time limit for reconstruction, although children usually finished quickly. Following the last encoding trial, children were told that they were to play another game, and instructions similar to the previous predictions trials were given. The encoding and predictions tasks were given in a single session lasting about 25 minutes. There were several variations on this basic procedure. We will describe each variation and its results in sequence. The results from all phases are shown in three forms. Table VII shows the percentage of correct distance and weight encoding for both age groups. A strict criterion of perfect reconstruction of both sides of the scale was used for both weight and distance scoring. Table VIII shows the percentage correct predictions for each type of problem, and Table IX shows the classification by model type for each age group in each phase of Experiment 3.

2 . Experiment 3a Ten children from each age level participated in the experiment exactly as described above. As shown in Table V11, the results were consistent with the encoding hypothesis. The younger children showed a great disparity between their ability to reproduce weight and their ability to reproduce distance while the older children did not show a significant difference. This pattern held for individual subjects in each age group, and is not the result of aggregating over subjects (see Siegler, 1976, for more extensive statistical analyses of these re-

98

David Klahr and Robert S . Siegler

sults). Notice that these encoding differences between older and younger children were not accompanied by a corresponding difference in ability to predict how the balance scale would behave. A s shown in Tables VIIl and IX, there was virtually no difference in the percentage of different types of problems passed or in the distribution of children using each model.

3. Experiment 3b In this variant, 10 5-year-olds were given 15 rather than 10 seconds to view the initial configuration during the encoding tests. This was done to explore the possibility that the younger children were simply a bit slower than the older ones in encoding the configuration;s. If they were attempting to encode both dimensions, and had a preferred noticing order of weight first, then giving them more time would be expected to improve their distance scores. As shown in Table VII, this “insufficient time” explanation is unsupported by the results. 4 . Experiment 3c

Perhaps the younger children did not understand what was meant by “make the same problem.” In this variant, the children were told explicitly what to encode, and what constituted the experimenter’s criterion for the “same” problem. Ten children of each age level participated. The instructions for the encoding task were changed to the following: The idea of the first game is for you to look how the weights are set on the pegs on my balance scale and then to make the same problem by putting the weights on the pegs on yours. You want it to be the same problem in two ways. You want the same number of weights on each side of your scale a s 1 had on my scale, and you want the weights on each side of your scale to be the same distance from the center as they were on my scale.

Later in the instructions, children were again told that they should “watch closely to see how the weights. are set on pegs-how many there are on each side

TABLE VIl Percentage Correct Encodings-Experiment

3

5-year-olds

Experiment 3a 3b 3c 3d

8-year-olds

Weight encodings

Distance encodings

Weight encodings

Distance encodings

51 54

73

56

54

16 9 19

52

51

64 72

73 76

TABLE VIII Percentage Correct Predictions-Experiment

Experiment 3a

3h 3c 3d 3e

3

Age (years)

Balance

Weight

Distance

Conflict-weight

Conflict-distance

5 8

95 98

100 100

8 5

100 100

2 0

5 5

85 72 100

18 18 30

92 12 90

8

8

5 8 5 8

72

85 90 98 92 100 89 I00

22 22 12 94

86 100

100

92 100

89 61

12 20 17 0 33 50

Conflict-balance 0

0 2 15

0

6 0 0 0

Davi,d Klahr and Robert S . Siegler

I00

TABLE IX Number of Children Using Different Models-Experiment

3

Models

Experiment 3a 3b 3C

3d 3e

Age (years) 5 8 5

5 8 5 8 5 8

1

I1

111

Unclassifiable

9 8 7 7 6 6 6 1 0

0 1 0 0 2 0 1 3 3

0 0 0 0 1 0 0 4 7

I

I 3 3 I 4 3 2 0

and how far from the center the weights on each side are.” Finally, at the end of the instructions, children were asked to indicate the two ways their arrangements should be like the experimenter’s. This was to ensure that they understood what they had been told. The few children who did not understand were presented the instructions again and asked the identical question until they could answer appropriately. In all other ways, the procedure was the same as that used in Experiment 3b, with a 15-second viewing period. Once again, as shown in T,ables VII, VIII, and IX, the results differed hardly at all from those of Experiments 3a and 3b. Telling children what to encode did not reduce the discrepancy between their encoding of weight and distance, nor did it improve their performance on the predictions task.

5 . Experiment 3d This time, children were told not only what to encode, but also how to encode it. If the problem lay in the inability of the younger children to correctly encode distance, or to handle two dimensions simultaneously, then perhaps direct instruction might help them. Ten children of each age group were given the following additional instructions during the encoding trials: You do it like this. First you count the number of weights on this sid-ne, two, three, four. Then you count the number of pegs the weights are from the center-first, second, third. So you say to yourself “four weights on the third peg.” Then you would do the same for the other sid-ne, two, three, four, five weights on the first, second, third peg. So it would be five weights on the third peg. Then you would say “four weights on the third peg and five weights on third peg.” Then you would piut the right number of weights on the right pegs on each side. Let’s practice one.

The Represen!urron 01 Children's Knowledge

101

This was followed by seven practice trials on which the child received feedback on the correct counting of weights and distances. This procedure was expected to reduce or eliminate the weight-distance discrepancy for the younger children, but since the older children presumably already knew how and what to encode, it was not expected to affect their performance. No effect was expected on the predictions performance of either group. All of these expectations were confirmed. Table VII shows that the younger children performed equally well on weight and distance, and that the older children performed better overall, but with no weightdistance discrepancy. Tables VIII and IX show that the predictions performance of both groups was indistinguishable from previous results.

6 . Experiment 3e Having finally eliminated the younger children's encoding deficit, we next asked whether that deficit really was the cause of the initial differential responsiveness to instruction. In this final experiment, the same children who participated in Experiment 3d were given the conflict training sequence used in Experiment 2 a few days after they completed Experiment 3d. According to the encoding hypothesis, both older and younger children should now benefit from experience with conflict problems that previously had benefited only the older children. Following the training sequence, the predictions test (without feedback) was again given to the two groups. The results of this posttraining predictions test are shown in Row 3e of Tables VIII and IX. Note that Rows 3d and 3e are based on the same set of subjects at different times. The sequence of manipulations and their corresponding results were: ( 1 ) instructions about what and how to encode; ( 2 ) encoding task (Table VII, Row 3d); (3) predictions task (Tables VIIl and IX, Row 3d); (4) conflict training with feedback, a few days later; and (5) repeat of predictions task (Tables VIII and IX, Row 3e). Comparison of Rows 3d and 3e i n Tables VIII and IX shows that training now aided both age groups. Although there appears to be a slight advantage overall for the older children, there were no significant effects for either age alone, and no age-problem type interaction. It seems clear that the qualitative differences in responsiveness to training were eliminated by prior training in encoding. Although the younger children did not benefit as much as the older, it should be remembered that their encoding performance also did not reach the level of the older children. B.

SUMMARY

The results of Experiment 3 provide strong support for the encoding hypothesis: Younger children clearly do not tend to encode the distance dimension in this task. Without such encoding, they can derive little benefit from the instruction series. However, if given careful and explicit instruction on encoding,

102

David Klahr and Robert S . Siegler

they do begin to do it correctly, and such improvement subsequently enables them to benefit spontaneously from a training sequence.

IX. Discussion: Some Answers and Some Further Questions In this final section we will briefly summarize the work reported thus far, and give some indication of possible future efforts. A.

WHERE WE HAVE BEEN

Let us summarize where this series of investigations has taken us in our attempt to address the five questions posed at the outset. Question 1: What are the differences in knowledge that underlie different levels of task performance? The results of Experiment 1 indicated that the four models, in either decision tree or production system fomiulation, could accurately represent different kinds of knowledge that underlie distinct behavior patterns. Question 2: What are the alternative strategies that might result in any given level of task performance? Our analysis defined levels in terms of four modal forms of rule systems. Thus, neither representation could account for alternative means by which a subject might be generating the pattern of responses that led to his classification according to the models. However, the comparison of predictions and explanations for the Model I11 children suggested that such variations were indeed occurring. Detailed analysis of Lisa’s protocol further indicated the need to account for individual variations, and with the construction of Model IIIA for Jan we began to demonstrate how these representations could account for highly idiosyncratic processes underlying Model I11 response patterns. As we argued earlier, the advantage of the production system representation for this individual level lay in its explicit set of psychological assumptions, assumptions that are consistent with a developing general view of some of the properties of the human information-processing system. Question 3: For a given level of performance, what is the optimal level of difficulty for an instructional sequence? The results of Experiment 2 provided a specific example of the general view that “near” training is better than “far” training. Although the particular definitions of “near” and “far’” were clearly derived from the underlying representation for knowledge on the task, we did not do enough parametric variation

The Representution of Children’s Knowledge

I03

to be able to claim true optimality. However, the dimensions along which an investigation of such variation could take place are very clear. Question 4: What are the critical features of an instructional sequence that enable it to have any effect? The results of Experiment 2 demonstrated that experience with particular types of problems is critically important for improving subsequent performance. This sensitivity was derivable from the initial modal forms of the models. In addition, the role of negative feedback was specified in Jan’s model, which is responsive to a mismatch between expectations and the actual outcome of each trial. Question 5 : When and why will two learners at the same initial performance level learn differently from the same instructional sequence? Most of the development of the protocol analysis and the final production system model for Jan was stimulated by our attempt to answer this question. In Experiment 2 we detected the differential response to experience; Lisa’s protocol suggested the encoding hypothesis; Jan’s production system indicated the potential complexity and importance of encoding operators and suggested the operational form of the encoding hypothesis, for which the results of Experiment 3 provided strong support. B.

WHERE WE ARE GOING

These investigations have suggested further explorations in two interacting domains: conceptualization of models and experimental studies. 1 . Types of Knowledge in the Human Information-Processing System

Our exploration of the issues surrounding the evaluation of different representations for knowledge has revealed that it is possible to distinguish between several different types of knowledge. The suitability of a representation depends upon the particular type of knowledge in which we are interested. In this section, we will briefly indicate what appear to us to be distinctly different kinds of knowledge. The order in which they are described corresponds roughly to their degree of permanence and stability in the human information-processing system. a. K1: Knowledge About the Momentary State of Affairs. This is the knowledge represented by the elements in WM in a production system, or in the more general concept of “active memory” in other cognitive theories. In a production system, all the productions are continually attempting to recognize familiar elements of K1, and to act upon it through modification. K1 represents what is “going on” from one moment to the next. It contains information about the environment-information that has been produced by encoding operations

104

David Klahr and Robert S. Siegler

and by the actions of satisfied productions. It constitutes a record of the system's immediate past. 6 . K2: Knowledge About h'ow to Do a Task or Solve a Problem. This type of knowledge is represented by decision trees of the sort used in Figs. 1 and 3 , or by the productions in a production system. The knowledge in K2 typically consists of tests for the type of knowledge represented by K1. A production system provides a convenient
The Repre.wnrtirion of Children’s Knowledge

I05

capacity refer to changes in the amount and type of K I (Chi, 1976; Huttenlocher & Burke, 1976), while studies of “metacognition” (Flavell, 1976) would appear to be primarily addressed to the development of the interaction between K2 and K3. Questions about “readiness” refer to the K2-K4 interaction and issues related to the competence-performance distinction seem to involve the interplay between K5 and K3, that is, between the “deep” interpretive capacity of the system and particular task-specific knowledge.

2. Experimental Extensiotis The results of our experiments suggest a number of directions for further research. One would be to examine problem isomorphs-tasks similar to the balance scale in formal properties but differing in specific characteristics. Efforts in this direction have already been made. Decision tree models have been formulated and tested on Inhelder and Piaget’s (1958) projection of shadows task, Bruner and Kenney’s (1966) fullness of a water jar task, and Chapman’s (1975) probability learning task. In each case the models have been found to accurately represent children’s predictive perfonnance. The experiments have also revealed a rich variety of reactions to feedback and encoding strategies (Siegler & Vago, in press). Another approach would be to construct and test more detailed production system models of exactly how children encode balance scale and other problems: Detailed analyses of reaction times might provide the appropriate test for such models. Finally, a host of instructional issues might be examined such as: When should tasks be taught directly and when should appropriate encoding strategies be taught first? Can procedures be devised to teach effective encoding on a variety of problems, or must instruction in encoding proceed on a task by task basis? Do differences i n encoding account for individual differences among children of a given age, as well as developmental differences? Pursuing these problems will almost certainly lead to new insights about the five questions posed at the outset of this paper, and also to new questions.

X.

Conclusion

Representation of children’s knowledge requires that we make testable assertions about both the basic encoding of the environment and the processes that operate on those encodings. Cognitive development and instructional procedures involve changes in both the encoding operators and the rule systems. Instruction will tend to be ineffective if the instructional situation is encoded by the learner in a manner that is unexpected by the instructor. In a limited domain, we have demonstrated that such misencoding was indeed occurring, that we could locate the point of difficulty, eliminate it, and have instruction proceed as we expected it to.

David Klahr and Robert S . Siegler

106

From a broader viewpoint, we have tried to show that the appropriate representation for knowledge depends upon the goals of the scientific endeavor. Different kinds of knowledge are best represented by different formalisms, and are best investigated by different empirical procedures. This pluralistic view of knowledge representation ma,y facilitate our understanding of, and influence upon, what it is that children know.

Appendix A Protocol Excerpts from Lisa, a 5-Year-Old, on Training Sequence 00400 TI

0001/2000

4:46

00500

00600 E. Okay, Let’s put these two here, and this one here. 00700 S . This side will go down (points left). 00800 E. Which side? Touch the side that will go down. 00900 S . (Touches left side) 01000 E. Okay. Let’s see if you were right. (Removes blocks. Scale tips left-up; right-down.) 0 1 100 Were you right? 01200 S . (Nods no) 01300 E. Which side went down? 01400 S . (Points right) 01500 E. Okay. Why do you think that was? Why did you think before this side would go down? Oi600 01700 S . ’Cause that one (points left) didn’t have as much as that one (points right). 01800 01900 E. Uh-huh. But what actuallj happened? 02000 S . This side went down because that one’s heavier (points right). 02100 02200 T3 0100/1OOO 6:lO 02300 02400 E. Okay. What do you think will happen this time? 02500 S. They will both stay up. 02600 E. Why do you think that? 02700 S. ’Cause they are both the same. 02800 E. Let’s see if you are right. (Removes blocks. Scale tips left-down.) Were you right? 02900 03000 S. (Nods yes) 03100 E. You were? Look. Do they both. . , Are they balanced? Is it like it was before? 03200 03300 S . Well, why are they both the same thing and one’s up and one’s down? 03400 E. Why do you think that is? 03500 S . I don’t h o w . 03600 03700 T4 0020/0020 7:00 03800 03900 E. Okay. What do you think will happen this time?

The Representation of Children’s Knowledge

107

04OOO S . The same again. 04100 E. They will stay the same again. why do you think that? 04200 S. ’Cause. Wait a minute. It won’t. 04300 E. It won’t? 04400 S. ’Cause this one (points left) is closer to this one (points to fulcrum). And this one (points right) is closer to this one (points to fulcrum). 04500 04600 E. So what will happen? 04700 S. This side (points right) will go up. 04800 E. This side will go up? 04900 S. Uh-huh. 05000 E. Okay. What do you mean by “up”? Point which way it will go. 05100 S. I think.. . 05200 E. Which? 05300 S. . . . it will go down. 05400 E. This side will go down (points left)? 05500 S. Uh-huh (nods yes). 05600 E. And this side. , . and so it will be like this? (Tilts balance manually left-down; right-up.) 05700 05800 S. Uh-huh. 05900 E. Is that right? 06OOO S. Uh-huh. 06100 E. Okay, let’s see if you are right. (Removes blocks, Scale tips right-down.) Were you right? 06200 06300 S. (Nodsno) 06400 E. What happened? 06500 S. This went down (points right). 06600 E. Why do you think that is? 06700 S. I don’t know! 06800 E. Well, think about it. 06900 S . Ummmm. 07000 E. Okay. 07100 S . I just don’t know why. 07200 E. You just don’t know why. 07300 S. Uh-huh. 07400 E. Well, we’ll keep on working and maybe you’ll figure it out. 07500 07600 07700 0102/2010 943 07800 T6 07900 08000 E. Okay, what will happen on this one? 08100 S. Yes. This side (points right) will both stay the same (points to both sides). 08200 08300 E. Let’s see i f . . . Why do you think that? 08400 S. Because they both look the same. One is empty in the middle and one is empty in the 08500 middle. 08600 E. Okay (scale balances). . . , Were you right‘? 08700 S. (Nodsyes) 08800 E. Uh-huh. You were. That’s right. 08900 S. 1 was right!

David Klahr and Robert S. Siegler

I08

09OOO 09100 09200 09300 09400 09500 09600 09700 09800 09900

E. That’s right.

T7

0200/2000

10:25

E. All righty. Now, what will happen this time? S. This side (points left) is farther away from this (indicates fulcrum), and this side (points right) is closer to this (indicates fulcrum). So, I think this side (points left) will go up. And this side will go down (points right).

IOOOO E. Let’s see if you are right. (Removes blocks left-down.) Were you right?

10100 S. Uh-uh (no). 10200 E. What happened? 10300 S. (Points left, center, right, and up) 10400 E. What? 10500 S. This went up (points right:) and this went down (points left). I thought this would go up 10600 (points left). 10700 E. Okay. Try to figure out what’s happening. 10800 S. Hm. 1 don’t know why. 10900 E. Well you just keep on trying to figure out. 1 1000 11100 11200 11300 E7 OOO3/0004 26:46 11400 11500 S. I’m ready. Ha, wait a minute, 1 forgot. 1 did it wrong. I gotta think. . . (can’t hear). This is four. This is three. This one will go down (points right). I1600 11700 I1800 E. What do you think would happen if we put one more here? What do you think would I1900 happen? [0004/0004] 12000 S. Both stay the same. This lone is crooked a little bit (adjusts right weights). 12100 12200 E. What do you think’ll happen? 12300 S. Stay the same. 12400 E. Yeah? Let’s see if you’re right. (Removes blocks. Scale tips rightdown, with sharp rap as it hits table.) Did they? 12500 12600 S. No! 12700 E. No? No, they didn’t. Did they? 12800 S. Plunk. Plunk. 12900 E. Plunk! Why do you think that was? 13000 S. I don’t know. They both ‘had four. See, one, two (counts left): one, two, three, four; 13100 (counts right): one, two, three, four. 13200 E. They both have four. Is that what made this side go down so much and this side go up SO 13300 much? 13400 S. No. 13500 E. What do you think it was? 13600 S. I don’t know. 13700 E. Think about it. What could it be? 13800 S. I just don’t know.

The Represvnrtirron of Children’s Knowledge

I09

13900 E. Just don’t know. Look at it for a moment and b y to figure out what it could be. Real 14000 carefully. 14100 S. This one is far away (points right) and this one is close (points left). 14200 14300 E. Okay. Have any other ideas? 14400 S. Uh-uh (no). 14500 14600 14700 E8. 0003/3000 28:40 14800 14900 E. Okay. Now, want to make up another problem? 15000 S. Uh-huh. 15100 E. Okay. 15200 S. This one is gonna be a good one. Stay the same. 15300 E. You think so? 15400 S. Uh-huh. 15500 E. Okay, let’s see if you’re right. (Removes blocks. Scale balances.) Were YOU right? I5600 15700 S. Uh-huh! 15800 E. Yeah, you were. 15900 S. I’m being right and right and right, but one time I was wrong. 16000 16100 16200 16300 E9 0003/0003 2926 16400 16500 16600 E. (Requests information on how scale works, and about what would happen on this trial.) 16700 16800 S . There’s three, and this side (points right) would go down, I guess. 16900 E. That side would go down? 17000 S. And this side would go up (points left). 17100 E. Why? 17200 S. Because this is far away (points right) and this is close (points left). So 1 think it would. 17300 17400 E. Think so? 17500 S. Uh-huh. 17600 E. Let’s see if you’re right. 17700 S. Ohh! Right! 17800 17900 E. What would you do to make it balance, now? (S. starts to move scale manually to balance 18000 position.) No. , . . I mean by moving the little, . . little circles around. What could you do 18100 to make that balance? 18200 18300 S. This three here (points to right, first peg), and this three stay here (points left). 18400 [0003/30001 18500 E. Let’s see if that’s right, what you do. 18600 S. I have to hold this up. (Lifts right side and moves weights.) [0003/3000] 18700 E. Are you right?

David Klahr and Robert S. Siegler

110

18800 S. 18900 19OOO E. 19100 S. 19200 E. 19300 S. 19400 19500 E. 19600 19700 S .

(Nods yes) What would you do to make the other side go down? Whoops. [3000/3000]That side will go up (points right). Whoops, there. Why does that happen? Because that one’s far away (points left) and that one’s close (points right).

I see. But if both had them far away. [3000/0003]Both sides would go down (giggle). They balance.

Appendix B Trace of production system for Model IIIA on four problems from Table VI. Assume that at start of Problem 5, criterion is set to distance, and retain final criterion value when moving on to next problem. Start with Problem 5. (Notation: Production system written in a special language called PSG. Numbers shown are counts of actions taken since start of each problem. Abbreviations are used for terms such as weight, distance, (etc.) Problem 5

(8288/8488)

0. WM: ((PRED) (CRITERION DST) ) Fire P8: ((PRED) (DI) ABS --* ATTEND) (wgt more right) (dst more left) ATTENDING-INPUT NEXT STIMULUS 1. WM: ((DST MORE LEFT) (WGT MORE RIGHT) (PRED) (CRITERION DST) ) CONFL1CT.SET: (P2 P4) Fire P4: ((PRED) (WGT MORE) (DST MORE) FIND.BIG) ATTENDING-INPUT NEXT STIMULUS 3 (wgt big right) (dst big left) 2 . WM: ((DST BIG LEFT) (WGT BIG RIGHT) (PRED) (WGT MORE RIGHT) (DST MORE LEFT) (CRITERION DST) ) CONFLICT.SET: (P2 P4 P5 P6 P7) CONFLICTSET: (P4 P5) AFTER SPECIAL.CASE.ORDER CONFLICT.SET: (P5) AFTER WM.ORDER (MADE **) (EXPECT XI Fire P5: ((PRED) (CRITERION DI) (D1 BIG XI) (D2 BIG X2)-+ DOWN) SAY.D)

+

-

6 . WM: ((EXPECT LEFT DOWN) (MADE (PRED)) (CRITERION DST) (DST BIG LEFT) (WGT BIG RIGHT) (WGT MORE RIGHT) (DST MORE LEFT) ) Fire E l : ((EXPECT) - 4 LOOK) ATTENDING-INPUT NEXT STIMULUS = (see right down)

Ill

The Representation of Children’s Knowledge

7. WM: ((SEE RIGHT DOWN) (EXPECT LEFT DOWN) (MADE PRED)) (CRITERION DST) (DST BIG LEFT) (WGT BIG RIGHT) (WGT MORE RIGHT) (DST MORE LEFT) ) CONFLICT.SET: (El E3) CONFLICT. SET: (E3) AFTER SPECIAL .CASE. ORDER Fire E3: ((EXPECT XI X2) (SEE X1 X2) ABS (SEE) --+ (DID **) (SEE SAW) (RESULT WRONG))

==+

10. WM: ((RESULT WRONG) (DID EXPECT LEFT DOWN (SAW RIGHT DOWN) (MADE (PRED)) (CRITERION DST) (DST BIG LEFT) (WGT BIG RIGHT) (WGT MORE RIGHT) (DST MORE LEFT)) Fire SW1: ((RESULT WRONG) (CRITERION DST) --+ (OLD **) (DST WGT))

==+

12. WM: ((OLD (RESULT WRONG)) (CRITERION WGT) (DID (EXPECT LEFT DOWN)) (SAW RIGHT DOWN) (MADE (PRED)) (DST BIG LEFT) (WGT BIG RIGHT) (WGT MORE RIGHT) (DST MORE LEFT) )

Now do Problem 6 . Keep criterion.

TE: ((010212010))

8. WM: ((PRED) (CRITERION WGT) ) CONFLICT.SET: (P8) Fire P8: ((PRED) (Dl) ABS -+ ATTEND) ATTENDING-INPUT NEXT STIMULUS j (dst same) (wgt same) 1. WM: (WGT SAME) (DST SAME) (PRED) (CRITERION WGT)) Fire P1: ((PRED) (WGT SAME) -+ (MADE **) (EXPECT BALANCE EVEN) SAY.B)

5. WM: (EXPECT BALANCE EVEN) (MADE (PRED)) (WGT SAME) (DST SAME) (CRITERION WGT) ) Fire E l : ((EXPECT) -+ LOOK) ATTENDING-INPUT NEXT STIMULUS j(see balance even) 6 . WM: ((SEE BALANCE EVEN) (EXPECT BALANCE EVEN) (MADE (PRED)) (WGT SAME) (DST SAME) (CRITERION WGT) ) CONFLICT.SET: (El E2) CONFLICT.SET: (E2) AFTER SPECIAL.CASE.ORDER SAW) (RESULT Fire E2: ((EXPECT X1 X2) (SEE XI X2) -+ (DID **) (SEE ==j CORRECT)) 9. WM: ((RESULT CORRECT) (DID (EXPECT BALANCE EVEN)) (SAW BALANCE EVEN) (MADE (PRED)) (WGT SAME) (DST SAME) (CRITERION WGT))

10. Wh4: ((OLD (RESULT CORRECT)) (CRITERION WGT) (DID (EXPECT BALANCE EVEN)) (SAW BALANCE EVEN) (MADE (PRED)) (WGT SAME) (DST SAME) )

I12

David Klahr and Robert S. Siegler

Problem 7 (888318820)

0. WM: ((PRED) (CRITERION WGT) ) ATTENDING-INPUT

NEXT STIMULUS j (wgt more left) (dst more right)

1. WM: ((DST MORE RIGHT) (WGT MORE LEFT) (PRED) (CRITERION WGT) )

C0NFLICT.SET: (P2 P4) C0NFLICT.SET: (P4) AFTER SPECIAL.CASE. ORDER Fire P4: ((PRED) (WGT MORE) (DST MORE) -+ FIND.BIG) AlTENDING-INPUT NEXT STIMULUS 3 (dst big right) (wgt big left) 2. WM: ((WGT BIG LEFT) (DST 131G RIGHT) (PRED) (WGT MORE LEFT) (DST MORE RIGHT) (CRITERION WGT) ) CONFLICTSET: (P2 P4 P5 P6 P7) C0NFLICT.SET: (P4 P5) AFTEiR SPECIAL.CASE.ORDER C0NFLICT.SET: (P5) AFTER WM.ORDER Fire P5: ((PRED) (CRITERION DI) (D1 BIG X1) (D2 BIG X2) -+ (MADE **) (EXPECT X1 DOWN) SAY.D)

6. W M ((EXPECT LEFT DOWN) (MADE (PRED)) (CRITERION WGT) (WGT BIG LEFT) (DST BIG RIGHT) (WGT MORE LEFT) (DST MORE RIGHT) ) CONFLI(JT.SET: (El) Fire El: ((EXPECT) -+ LOOK) A'ITENDING-INPUT NFXT STIMULUS j (see right down) 7. WM: ((SEE RIGHT DOWN) (EXPECT LEFT DOWN) (MADE (PRED)) (CRITERION WGT) (WGT BIG LEFI') (DST BIG RIGHT) (WGT MORE LEFT) (DST MORE RIGHT ) ) CONFLICTSET: (El E3) CONFLICT.SET: (E3) AFTER SPECIAL.CASE.ORDER AW Fire E3: ((EXPECT XI X2) (SEE XI X2) ABS (SEE) - + (DID **) (SEE = =Sj (RESULT WRONG))

18. W M ((RESULT WRONG) (DID (EXPECT LEFT DOWN)) (SAW RIGHT DOWN) (MADE PRED)) (CRITERION WGT) (WGT BIG L E W (DST BIG RIGHT) (WGT MORE LEFT) (DST MORE RIGHT) ) Fire SW2: ((RESULT WRONG) (1CRITERION WGT) -+ (OLD **) (WGT ==3 DST)) 12. W M ((OLD (RESULT WRONG)) (CRITERION DST) (DID (EXPECT LEFT DOWN)) (SAW RIGHT DOWN) (MADE (PRED)) (WGT BIG LEFT) (DST BIG RIGHT) (WGT MORE LEFT) (DST MORE RIGHT) ) Problem 8 (0188/0288) [notice, this has big dst, but not big wgt]

0. W M ((PRED) (CRITERION DST) ) Fire P8: ((PRED) ( D l ) ABS -+ A'ITEND) ATTENDING-INPUT NEXT STIMULUS j (wgt more right) (dst more left) TE:((0188/0200))

The Repre.wnratron (JJ Children's Knowledge

I13

1. WM: ((DST MORE LEFT) (WGT MORE RIGHT) (PRED) (CRITERION DST) ) CONFLICTSET: (P2 P4) CONFLICT.SET: (P4) AFTER SPECIAL.CASE.ORDER Fire P4: ((PRED) (WGT MORE) (DST MORE) -+ FIND.BIG) ATTENDING-INPUT NEXT STIMULUS 3 (dst big left)

2. WM: ((DST BIG LEFT) (PRED) (WGT MORE RIGHT) (DST MORE LEFT) (CRITERION DST)) CONFLICT.Sm: (P2 P4 P7) CONFLICT.SET: (P4 P7) AFTER SPECIAL.CASE.0RDER CONFLICT.SET: (P7) AFTER WM.ORDER Fire P7: ((PRED) (DST BIG X1) -+ (MADE **) (EXPECT X1 DOWN) SAY.D)

********** LEFT down 6. WM: ((EXPECT LEFT DOWN) (MADE PRED)) (DST BIG LEFT) (WGT MORE RIGHT) (DST MORE LEFT) (CRITERION DST) ) Fire El: ((EXPECT) -+ LOOK) ATTENDING-INPUT NEXT STIMULUS 3 (see right down) 7. WM: ((SEE RIGHT DOWN) (EXPECT LEFT DOWN) (MADE (PRED)) (CRITERION WGT) (WGT MORE RIGHT) (DST MORE LEFT) (CRITERION DST) ) CONFLICT.SET: (El E3) CONFLICT .SET: (E3) AFTER SPECIAL .CASE. ORDER FireE3:((EXPECTXlX2)(SEEXlX2)ABS(SEE)(DID**)(DID**)(SEE 11 11SAW) (RESULT WRONG)) 18. WM: ((RESULT WRONG) (DID (EXPECT LEFT DOWN)) (SAW RIGHT DOWN) (MADE PRED)) (DST BIG LEFT) (WGT MORE RIGHT) (DST MORE LEFT) (CRITERION DST) ) Fire SWI: ((RESULT WRONG) (CRITERION EST)--+ (OLD **) ( D S T = = + WGT))

12. WM: ((OLD (RESULT WRONG)) (CRITERION WGT) (DID (EXPECT LEFT DOWN)) (SAW RIGHT DOWN) (MADE (PRED)) (DST BIG LEFQ (WGT MORE RIGHT) (DST MORE LEFQ )

ACKNOWLEDGMENTS This study was supported by grants from The Spencer Foundation, by Public Health Service Grant MH-07722 from the National Institute of Mental Health, and by a grant from the Sloan Foundation. Thanks to S. Famham-Diggory, C. Glenn, and A. Newell for comments on an earlier version.

REFERENCES Baylor, G. W., & Gascon, J. An information processing theory of aspects of the development of weight seriation in children. Cognitive Psychology, 1974, 6 , 1-40. Becker, J . D. Reflections on the formal description df behavior. In L. Bobrow & A. Collins (Eds.), Representation and.learning. New York: Academic Press, 1975.

114

David Klahr and Robert S . Siegler

Blatt, M. The effects of classroom discussion upon children’s level of moral judgment. In L . Kohlberg & E. Tunel (Eds.), Recent research in moral development. New York: Holt, 1971. Bobrow, D. G. Dimensions of representation. In D. G . Bobrow & A. Collins (Eds.),Representation and understanding. New York: Academic Press, 1975. Bobrow, D. G . , & Collins, A. (Eds.), Representation and understanding. New York: Academic Press, 1975. Bower, G.H. Cognitive psychology: An introduction. In W. K. Estes (Ed.), Handbook of learning and cognitive processes (Vol. I). Hillsdale, N.J.: Lawrence Erlbaum Associates, 1975. Braine, M. D. S. Development of a grasp of transitivity of length: A reply to Smedslund. Child Development, 1964, 35, 799-8110. Broadbent, D. E. The magic number s(:ven after fifteen years. In R. A. Kennedy & A . Wilkes (Eds.), Studies in long term memory. N I : ~York Wiley, 1975. Bruner, J. S., & Kenney, H. On relational concepts, In J. S. Bruner, R. R. Olver, & P. M. Greenfield (Eds.), Studies in cognitive growth. New York Wiley, 1966. Chapman, R . H. The development of children’s understanding of proportions. Child Development, 1975, 46, 141-148. Chase, W. G., & Simon, H. A. The mind’s eye in chess. In W. G . Chase (Ed.), Visual information processing. New York: Academic Press, 1973. Pp. 215-281. Chi, M. T. C. Short term memory limitations in children: Capacity or processing deficits? Memory and Cognition, 1976, 4, 559-512. Flavell, J . H. Metacognitive aspects of problem-solving. In L. B. Resnick (Ed.), The nature of intelligence. Hillsdale, N . J . : Lawrence Erlbaum Associates, 1976. Gelman, R. Logical capacity of very young children: Number invariance rules. Child Development, 1972, 43, 75-90. (a) Gelman, R. The nature and development of early number concepts. In H. Reese (Ed.), Advances in child development (Vol. 7). New York: Academic Press, 1972. (b) Glaser, R. Evaluation of instruction and changing educational models (Center for the Study of Evaluation of Instructional Programs, Occasional Report No. 13). Los Angeles, Calif.: University of California, 1968. Greeno, J. G . Cognitive objectives of instruction: Theory of knowledge for solving problems and answering questions. In D. Klahr (Ed.), Cognition and insrruction. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1976. Hunt, E. What kind of computer is Man? Cognitive Psychology, 1971, 2, 57-98. Huttenlocher, J . , & Burke, D. Why does memory span increase with age? Cognitive Psychology, 1976, 8 , 1-31. lnhelder, B., & Piaget, J. The growth of logical thinking from childhood to adolescence. New York: Basic Books, 1958. Jackson, S. The growth of logical thinking in normal and subnormal children. British Journal of Educational Psychology, 1965, 35, 255-258. Klahr, D. Designing a learner: Some questions. In D. Klahr (Ed.), Cognition and ihtruction. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1976. (a) Klahr, D. Steps toward the simulation of intellectual development. In L. B. Resnick (Ed.), The nature of intelligence. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1976. (b) Klahr, D., & Wallace, J. G . Development of serial completion strategies: An information processing analysis. British Journal of Psychology, 1970, 61, 243-257. Klahr, D., & Wallace, J. G. The role of quantification operators i n the development of conservation of quantity. Cognitive Psycholo,qy, 1973, 4, 301-327. Klahr, D., & Wallace, J. G. Cogni,tive development: A n informarion-processing view. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1976.

The Representation of Children’s Knowledge

115

Kuhn, D. Mechacisms of change in the development of cognitive structures. Child Development, 1972, 43, 833-844. Lee, L . C. The concomitant development of cognitive and moral modes of thought: A test of selected deductions from Piaget’s theory. Genetic Psychology Monographs, 1971,85, 93-146. Lovell, K . A follow-up study of lnhelder and Piaget’s “The growth of logical thinking.” British Journal of Psychology, I96 I , 52, I43 - 153. Mager, R. F. Preparing instructionul objectives. Palo Alto, Calif.: Fearon Publishers, 1962. McDermott, J. & Forgy, C. Production system conflict resolution strategies. In D. A. Waterman and F. Hayes-Roth (Eds.), Pattern-directedinference systems. New York: Academic Press (in press). Miller, G . A. The magical number seven, plus or minus two: Some limits on our capacity for piocessing information. Psychological Review, 1956, 63, 81-97. Moore, J . , & Newell, A. How can MERLIN understand? In L. W. Gregg (Ed.), Knowledge and cognition. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1974. Newell, A . A note on process-structure distinctions in developmental psychology. In S . FarnhamDiggory (Ed.), Information processing in children. New York: Academic Press, 1972. (a) Newell, A. A theoretical exploration of mechanisms for coding the stimulus. In A. W . Melton & E. Martin (Eds.), Coding processes in human memory. Washington, D.C.: Winston, 1972. (h) Newell, A. Production systems: Mcdels of control structures. In W. G. Chase (Ed.), Visual information processing. New York: Academic Press, 1973. Newell, A,, & McDermott, J . PSG Manuul. Pittsburgh: Carnegie-Mellon Univetsity, Department of Computer Science, 1975. Newell, A,, & Simon, H. A. Human problem solving. Englewood Cliffs, N.J.: Prentice-Hall, 1972. Pascual-Leone, J. A mathematical model for the transition rule in Piaget’s developmental stages. Acta Psychologica, 1970, 32, 301 -345. Piaget, J . The child‘s conception of time. New York: Ballantine, 1971. Pick, A. D., Frankel, D. G., & Hess, V . L. Children’s attention: The development of selectivity. Chicago: University of Chicago Press, 1975. Reddy, R., & Newell, A. Knowledge and its representation in a speech understanding system. In L. W. Gregg (Ed.), Knowledge and cognition. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1974. Resnick, L. B. Task analysis in instructional designs: Some cases from mathematics. In D. Klahr (Ed.), Cognition and instruction. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1976. Rychener, M. D. Introduction to Psnlst. Pittsburgh: Camegie-Mellon University, Department of Computer Science, 1976. Siegler, R. S . Defining the locus of developmental differences in children’s causal reasoning. Journal of Experimental Child Psychology, 1975, 20, 5 12-525. Siegler, R . S. Three aspects of cognitive development. Cognitive Psychology, 1976, 8, 481 -520. Skgler, R. S . (Ed.). Children’s Thinking: Whaf Develops? Hillsdale, N.J.: Lawrence Erlbaum Associates, 1978. Siegler, R. S . , & Liebert, R. M. Effects of contiguity, regularity and age on children’s causal inferences. Developmental Psychology, 1974, 10, 574-579. Siegler, R. S., & Liebert, R. M. Acquisition of formal scientific reasoning by 10- and 13-year-olds: Designing a factorial experiment. Developmental Psychology, 1975, 11, 401-402. Siegler, R . d., & Vago, S . The development of proportionality concepts: Judging relative fullness. Journal of Experimental Child Psychology (in press, 1978). Smedslund, J. The development of transitivity of length: A comment on Braine’s reply. Child Deveio.oment, 1965, 36, 577-580. Turiel, E. An experimental test of the sequentiality of developmental stages in the child’s moral judgment. Journal of Personality and Social Psychology, 1966, 3 , 61 1-618.

I I6

David Klahr and Robert S.Siegler

Waterman, D. A. Adaptive production systems (CIP Working Paper No. 285). Pittsburgh: CarnegieMellon University, Department of Psychology, 1974. Waugh, N. C., & Norman, D. A . Primary memory. Psychological Review, 1965, 72, 89-104. Young, R . M. Children’s seriation behavior: A production-system analysis. Unpublished doctoral dissertation, Carnegie-Mellon University, 1973. Zeaman, D., & House, B . J . The role of attention in retardate discrimination learning. In N. R. Ellis (Ed.), Handbook of mental deficiency. New York: McGraw-Hill, 1963. Pp. 159-223.

CHROMATIC VISION IN INFANCY

Marc H . Bornstein PRINCETON UNIVERSITY

I. INTRODUCTION.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. COLOR VISION AND PATTERN VISION, . . . . . . . . . . . . . . . . . . . . . . . . . . . . B , EARLY INTEREST IN THE ONTOGENY OF COLOR PERCEPTION. . . . . . C . MODERN VIEWS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

117 117 119 122

11. FUNDAMENTAL DATA OF COLOR VISION: ONTOGENY . . . . . . . . . . . . . . . . . A. BRIGHTNESS AND SPECTRAL SENSITIVITY . . . . . . . . . . . . . . . . . . . . . . . . B. CHROMATIC VISION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123

111. SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

167

APPENDIX: TRANSLATION OF THE ADDENDUM TO DARWIN’S “BIOGRAPHICAL SKETCH OF A YOUNG CHILD” ......................

169

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

169

125

132

I. Introduction A.

COLOR VISION AND PAlTERN VISION

Students of sensation and perception agree that vision predominates among the senses (e.g., Geldard, 1972; Marks, 1974). Specific examples from several different perspectives abound. The largest proportion of sensory neurons in the central nervous system is devoted to visual system function (e.g., 540 million, as compared to 100 million for the auditory system, according to Sinsheimer, 1971). Careful studies of human sensory dominance (e.g., Colavita, 1974; Pick, Warren, & Hay, 1969) and chronometric analyses of sensory information processing (e.g., Posner, Nissen, & Klein, 1976) agree that vision is prepotent. Vision is certainly the richest of the sense departments; the highly complex 117

1 I8

Marc H.Bornstein

structure of the eye differentiates and develops earliest in ontogeny; and vision mediates our most elaborate perceptual experiences. In the same way that vision dominates the sensory hierarchy in adulthood, so it does early in development. Literature evaluations and experience attest that vision is a principal factor in perceptual and cognitive development (Goodnow, 197 1; Gottesman, 1976), exploratory behavior (Nunnally & Lemond, 1973), and the development of social responsiveness (Walters & Parke, 1965) early in life. Tronick and Clanton (1971) and Bruner and Koslowski (1972) have shown the human infant to be “visually preadapted” for the “exploration and extraction of information from the environment” (Tronick & Clanton, 1971, p. 1483). “Babies grasp the world first with their eyes and then with their hands. Vision is therefore a prime constituent in the development of the total child.” The significance of visual perception in general in human infancy cannot be overstated, even by Gesell (1950, p. 3). For example, i n a direct test of Berkeley’s (1709/ 1901) view that touch teaches vision, Bower, Broughton, and Moore (1970) found that in infancy visual information predominates over tactile; likewise E. J. Gibson (1969) found that older infants favor visual cues over tactile on the visual cliff, and this pattern is continued in even older children who use visual information at the expense of other cues in intermodal matches (Abravanel, 1972; Bryant, 1974). For reasons like these vision has received extraordinary theoretical and experimental attention (e.g., C. H. Graham, 1965; Kaufman, 1974; Kling & Riggs, 1971; Osgood, 1953; Wyszecki & Stiles, 1967). Two aspects of visual information processing have been emphasized in research, pattern vision and color vision, and as a consequence of that research much is now known about these two visual functions in adults. Pattern vision obviously subserve! many basic visual differentiations, including those of form, shape, anld orientation, and therefore the question is sometimes asked, “Why color vision?” Whatever its original purpose or evolutionary significance (Polyak, 1957; Walls, 1942), color today fills our lives, and 96% of the human population possess normal color vision. For most, color has physical, chemical, physiological, psychological, aesthetic, commercial, natural, or agricultural interest. In perception and cognition, for example, colors enhance photocontrasts and thereby potentiate discrimination, identification, search, and recognition. Colors attract and maintain attention differentially. Moreover, hues ubiquitously symbolize and signify. Though the experimental sldlls of numerous developmental investigators have shown that human infants can see, process, remember, and act on patterned information within the environment (Cohen & Salapatek, 1975), relatively much less is known about the ontogeny of chromatic information processing. Yet the question of whether and how infants and young children see color and utilize chromatic information is of interest for perceptual, theoretical, and normative

Chrimcrrrc Vision in Infane?

I19

developmental reasons. First, color obviously functions as a salient perceptual cue: Children attend to, and hence order, environmental changes i n a hierarchy with color change more prominent than changes of size, number, or rotation (Dodd &Lewis, 1969; Lewis & Baumel, 1970). Second, Bornstein, Kessen, and Weiskopf (1976a) looked at the organization of hue in infancy, before language acquisition, to assess the relative primacy of language and perception, and Pollack (1970) tested hue- against lightness-contrast illusion figures in an experimental examination of Piagetian cognitive theory. Empirically, Baldwin ( 1895) saw “determination of the order of rise of the child’s perceptions of the different qualities of colour” (p. 39) to be exemplary of more general problems of experimenting with children. Finally, examination of the literature on chromatic development also continues a tradition of programmatic research in the development of sensory system capacities, while advances in techniques to assess visual status early have obvious clinical value. This chapter is intended primarily to review studies on the early ontogeny of chromatic vision in man for their several historical, theoretical, descriptive, and empirical values. Color serves obvious attentional, perceptual, cognitive, and aesthetic functions in vision. Each of these will be examined from a developmental perspective. B.

EARLY INTEREST IN THE ONTOGENY OF COLOR PERCEPTION

Interest in the development of color vision was prominent in the first observations of early baby biographers. Darwin’s (1877b) “A biographical sketch of an infant” began with observations on the early sensory development of his son, Doddy. Darwin reported that although Doddy fixated the light of a candle at nine days, he did not attend to color until his forty-ninth day.’ Preyer (1890), the nineteenth century developmental neurobiologist, introduced his two volume work on The Mind of the Child with notable discussions of his child’s developing “sensibility to light” and “discrimination of colors.” Though he observed that sensitivity to light was present at birth, Preyer was unable to demonstrate to his own satisfaction that his son could discriminate among basic colors until much later. Satisfactory demonstrations relied, in Preyer’s view, on successfully naming and recognizing colors, accomplishments that were not attained until his son’s second birthday. Early visual processes constituted very important subjects for Preyer and others in the nineteenth century since mental development was then believed to depend upon the activity of the senses. ‘Darwin’s benchmark “Biographical sketch of a young child” was published simultaneously in English in Mind (1877b) and in German in Kosnios (1877a). Intriguingly, the German edition contains an addendum, not included in the English version, that specifically concerns the early development of chromatic vision. An original translation of his addendum is given in the Appendix.

I20

Marc H . Bornstein

At the beginning of the twentieth century, observation turned to experimentation. In order to examine color and other perceptions in infancy, Baldwin (1893, 1895) invented his “dynamogenic” method. According to Baldwin, more dynamogenic stimuli should elicit more motor responses from infants than less dynamogenic stimuli. At 9 months Baldwin’s daughter, H., showed equal attraction to (i.e., grasping at) red and blue, somewhat less to green, and least to brown. Differential dynamogenic responsivity indicated to Baldwin that H. discriminated among these colors. Many of Baldwin’s quite sophisticated notions concerning this aspect of early behavior persist in contemporary infant research. Baldwin’s lead was immediately followed by several researchers. Holden and Bosse (1900) failed to find any differential reactions to colors in infants younger than 6 months; however, Marsden (l903a, 1903b), using preferential grasping and looking to indicate discrimination, found that his son attended to yellow and then red at approximately 4 months of age. On the basis of similar preference criteria, McDougall (1908) rnaintained that infants 5 months of age can discriminate among blue, green, and red. Myers (1908) concluded that his 6-month-old child “distinctly preferred” reds and yellows to blues, and Wooley (1909), using a paired-comparisons procedure, found that differential interest in colors definitely developed in the sixth month of life. Valentine (1913-1914), later author of books on The Psychology of Early Childhood (Valentine, 1942) and The Experimental Psychology cf Beauty (Valentine, 1962), studied looking at 4 months and grasping at 8 months in his own son. Using a paired-comparisons procedure, he found preference for yellow which was longitudinally stable. Baldwin’s idea that differential preference indicates discrimination failed to convince all developmental researchers (e.g., Nagel, 1906; Schallenberger, 1896). In a dissertation tbat included meticulous observations of one child’s reactions to colored flowers and book jackets, Shinn (1909) rejected her own observations of the infant’s differential attention to colors and concluded on the basis of language acquisition data that her niece, Ruth, passed through three distinct stages of chromatic development. The first, from birth to 15 months, was marked by sensitivity to light, but blindness to color. The second was a period of partial color vision as evidenced by Ruth’s errors in color naming, association, and recognition. The third, beginning at 3 years, characterized the child as adultlike in her color perceptions. Semantic development was the measure many other investigators depended upon. Before Shinn, Garbirii (1894) had suggested that because younger children are unable to name or match colors, total color deficiency must extend for the first 18 months of infancy. and partial color deficiency until the fourth year of life. Thus, color deficiency was widely believed to be “the normal condition of the newborn child” (Tracy & Stimpfl, 1909, p. 11). According to Major (1915), Baldwin’s tests tapped preferences, not discrimination, and R., Major’s son, truly discriminated black from red only at 3 years when he could name them

Chrornutic Vision

in

Infancy

121

correctly. This view persisted with inevitable consequences. Thus, Stern ( 1924) found 3-year-old children to be confused in their color naming and attributed the “legendary color blindness” of childhood to the late appearance of accurate color naming. Like views are with us still (Vernon, 1962). At least one vision researcher i n this period, Nagel (1906), himself a deuteranope, observed the color sense of a child, in this case his own 2-year-old son, Gerhard. Intrigued by contemporary opinions he termed “contradictory,” Nagel was interested in whether in fact children were color deficient. He deduced from Gerhard’s ability to learn the names of colors that by at least their second year children possessed normal color vision. The confusion yielded by the different methods, nonverbal and verbal, are manifest in Dearborn’s (1910) chronology of the development of his child, L. Paraphrased excerpts from his diary on the subject demonstrate what Dearborn himself called “the color-question in infant psychology.” Red stimulates the child’s attention. Distinct red-green discrimination. Visual preferences clearly based on hue rather than brightness. Failure to abstract redness indicates child has no idea of color qua color. Conduct gives evidence of discrimination, color naming evidence of confu. sion. 16-17 months: No appreciation of redness by name. 18.5 months: Distinguishes basic color names. Color discrimination “very certain and complete.” 24 months: Child orders a spectrum of eight Bradley sample-colors. 26 months:

3.5 months: 6 months: 7.5 months: 15 months: 15.5 months:

The development of L . from apparent preverbal hue discrimination to verbal confusion led Dearborn to wonder whether the child’s errors were perceptual or semantic in nature; L.’s early behavior notwithstanding, Dearborn conservatively accepted only the child’s later accurate color naming as satisfactory evidence of the presence of color vision. Historically, two questions have dominated the research interests of investigators in this area: When does color perception begin? and, Do the colors have an order of appearance? Clearly, the principal question, about when color vision begins, had not then been satisfactorily answered. In addition to the failure to establish a clear criterion, the confounding of hue discrimination with other perceptual variables, particularly with brightness, was responsible for a lack of progress. Natural stimuli in the world almost inevitably vary in brightness as well as hue. Quite early, Darwin and Preyer, among others, remarked on the infant’s keen sensitivity to light and changes in brightness. Many authorities, for example Nagel (1906), recognized the problem of stimulus control. Thus Baldwin (1895, p. 56) asked about such experiments “should not the colours be chosen to be equal in purity, intensity, lustre, illumination, etc.?” Because of their failure to eliminate these variations. or others such as movement

122

Marc H . Bornsfein

and contour, among chromatic stimuli, careful investigators hesitated to conclude when color vision began. The same problem has persisted in contemporary research. When Shirley (1933) tested color preference in babies 33 and 46 weeks of age and found a consistent preferential grabbing at yellow, she would conclude little since, by her own admission, yellow was the brightest color among her experimental stimuli. Previously, Valentine (1913-1914) had noted consistent yellow preferences when yellow was his brightest test color. Arlitt (1946) reported methodological difficulties of another kind. Her student, Luken, found children at 2 years unable to match colors to sample; because children at age 4 could, Luken concluded that 2-year-olds possess inadequate color perception. However, failure to demonstrate an ability, here color vision, does not necessarily imply its absence. Rather, inadequate test instruments, poor methods, or inappropriate psychological state may prevent or obscure a positive result. Researchers have, in general, been too willing to accept the null hypothesis in studies designed to demonstrate the existence of a phenomenon like color vision. The second research question focused on the order in which colors become discriminable. Most opinions accorded with one of two color theories. Peiper (1961/1963) summarizes diverse opinion on the ontogenetic sequence of color appearance. Myers ( 1908), for example, predicted from the Young-Helmholz theory that color receptors-and color perception-would mature in the order red, green, and blue. Garbini (1894) modeled his view of ontogenetic development on parallel views of phylogenetic development. According to Garbini, blue and yellow distinctions differentiate after red and green. Ladd-Franklin (1929) also proposed a stagelike development. She believed, however, that the differentiation of retinal elements would first permit only intensity discriminations, and later long (yellow) versus short (blue) hue discriminations. Finally, yellow would further evolve into red and green. Note that the speculative orders of chromatic development proposed by Garbini and Ladd-Franklin are mutually exclusive. Koffka (1927) thought that a “warm” color separates into red and yellow at the same time that a “cold” color separates into blue and green, and H. C. Smith (1943) supported this view. Despite the confusions which prevailed, the relative impoPance of color vision and color naming was not lost on nineteenth and early twentieth century thinkers. Indeed Binet’s (1890; Biner & Simon, 1908/1916) original IQ test included asking young children to name a color pointed to. In order to obtain any credit on this item, 8-year-olds were expected to identify correctly the four “fundamental” colors: blue, green, yellow, and red. C.

MODERN VIEWS

Since the second quarter of the twentieth century, experimental interest in the study of early color visual processes has gone unabated. Nevertheless,

Chromatic Vision in Infancy

123

psychophysical considerations such as stimulus control, methodological weaknesses such as reliance on preference measures where the presence or absence of preference indicates iittle about the quality or type of discrimination, and problems of interpretation such as acceptance of the null hypothesis, have severely limited scientific understanding of the young infant’s visual world and have colluded to insure persistent skepticism about infant capabilities. These and other shortcomings have been raised ir! early (Staples, 1931) as well as recent reviews of this literature (Kessen, Haith, & Salapatek, 1970; Kaye, cited in Reese & Lipsitt, 1970). Perhaps for these reasons, most contemporary developmental and perceptual discussions are practically devoid of reference to the topic of early color vision. Major developmental texts often omit the subject. Liebert, Poulos, and Strauss (1974), Mussen, Conger, and Kagan (1974), Hurlock (1972), and Jersild, Telford, and Sawrey (1975) all discuss perceptual development, including form, pat:ern, depth, the constancies, and attention, but omit color. Even more specialized works on early perceptual development do not cover color vision. Of 1 1 chapters in Cohen and Salapatek’s ( 1975) two-volume Infant Perception: From Sensation to Cognition, nine are devoted to studies of visual development. None concerns early color vision or its development. Likewise, the term “color” does not even appear in the index of Bower’s ( 1975) perceptual Development in Infancy or in the index of Rosinski’s ( 1977) The Development of Visual Perception. Nor is the developmental dimension usually included in more specialized essays on sensory or perceptual psychology. When it is, however, authors often lament a paucity of knowledge. Indeed, “Experts in color vision are frequently embarrassed to have to admit that there is little hard scientific fact about the nature of color vision in the young of our own species” (Jacobs, 1976, p. 81). In comparison, the twentieth century has witnessed remarkable growth in our understanding of the psychophysiology and the psychophysics of color vision in adult humans and various infrahuman species. In the following review of issues and advances in the ontogeny of chromatic vision, some knowledge of adult and comparative data as well as of the basic concepts of visual science is assumed; for further elucidation and elaboration, the reader is referred to Geldard (1972), C. H. Graham (1965), Kaufman (1974), Kling and Riggs (1971), and Wyszecki and Stiles (1967).

11. Fundamental Data of Color Vision: Ontogeny Light, the physical source of information in vision, varies in two principal ways, in wavelength composition and in intensity. Both influence the three psychological dimensions of color vision. Spectral wavelengths effective for human color vision vary between approximately 400 nm and 700 nm. The psychological correlate of wavelength and the first principal dimension of color vision is hue.

124

Marc H . Bornstein

Intensity, the second way light varies, is specified in radiance (the energy radiated by a light source) or in luminance (the energy which is effective for vision). The psychological correlate of intensity and the second principal dimension of color vision is brightness. Colors also vary in purity or chromaticness, the degree to which a light approaches monochromatic. The approximate psychological correlate of purity and the third principal dimension of color vision is saturation. Relationships among these three dimensions of color vision are best represented diagrammatically in a color solid (Fig. 1) that encompasses all possible chromatic experiences. Brightness is shown on the major axis from black to white. Hue is ordered as a circular prray that includes spectral and extraspectral colors. (Extraspectral colors are mixtures of spectral wavelengths.) Saturation is represented on the radii and shows the amount of color in a mixture. Nickerson and Newhall (1943) estimated that the number of discernible colors that fill this color space approximates 7,500,000. Color vision allows organisms to distinguish among stimuli that may or may not be the same size and shape. Judd (1951a) defined colors as those “characteristics of light other than spatial and temporal inhomogeneities” (p. 862). Color is not, however, a physical property of light; light itself has no color. Color is a sensation. The colors of objects in a natural scene depend, therefore, upon several factors, including (1) characteristics of the source of illumination on the scene, (2) the reflectance characteristics of the objects and their surround in the scene, and (3) the attenuation and sensitivity characteristics of the visual system to wavelength. Color vision thus exemplifies potential relationships among the physics of the natural world, the anatomy, neurochemistry , psychophysiology of the human body, and the psychology of the human mind. Several quantitative functions describe perception along the three color axes.z First, the absolute and differential thresholds for brightness perception are important to understanding the limits of color vision. The chromatic threshold as well as other suprathreshold measures of photopic spectral sensitivity define the overall sensitivity of the visual system to light, and the wavelength discrimination function describes the differential sensitivity of the visual system to hue. Hue categorization functions delimit the subjective organization of the spectrum and thus the extent and number- of qualitatively dissimilar hues experienced by an organism. Wavelength preference and color memory are equally important to a comprehensive description of color vision. Finally, different criteria define the *Although the physical dimensions associated with changes in brightness, hue, and saturation are independent, the psychological coirelates themselves are interdependent. Thus, the luminance of a wavelength must reach photopic levels for that wavelength to yield visible color. When intensity is increased, the hues of all but unique hues change, a phenomenon known as the Bezold-Briicke hue shift (Boynton & Gordon, 1965). Lightness and saturation are also integrally related (Davidoff, 1974), and changes in hue elicit concomitant changes in saturation (C. H. Graham, 1965). These natural interactions observed, we shall continue to discuss the three principal dimensions of color vision as though they were essentially separable continua of experience.

Chromatic Vision in Infancy

\

I25

\\

\

\

i Fig. I . The psychological color solid

quality of color vision that an organism possesses; among them hue categories and the spectral quality of light are prominent. The ontogeny of each of these standard functions in infancy, and the concepts associated with them receive further elaboration below. Treatments of each, however, incorporate one quite important assumption, namely, that adaptation, surround, retinal locus, and other state and physical variables surely influence the perception of color. The ensuing discussions therefore assume the following viewing conditions: foveal stimulation, neutral adaptation, absence of surround induction, and photopic levels of illumination. Under these viewing restrictions, each of the basic visual functions characterizes normal adult performance, that of an “ideal” or “standard” observer. In the ensuing considerations of developmental data, the results from studies of the very young are often compared against the performance of such a standard adult observer. A.

BRIGHTNESS AND SPECTRAL SENSITIVITY

I . Brightness The perception of brightness, from dark to light, varies with the luminance of a visual stimulus. Photopic luminance, associated with cone function and hence

126

Marc H . Bornstein

color vision, varies between chromatic threshold (- 10 mL) and the point at which light may cause visual damage (- lo9 mL). Darwin (1877b) first observed Doddy’s pronounced visual attention to candle light, and Champneys (1881:i as well remarked that neonates are sensitive to and experience pleasure from visual stimulation. Early experimental observations of infants’ sensitivity to brightness were indexed by preferential looking. Hershenson (1964) examined the newborn’s visual choice for pairs of grays of different lightnesses. Although Hershenson interpreted his results to indicate that infants prefer intermediate intensities to bright, and bright to dim, Thomas (1973) reanalyzed Hershenson’s data using Coomb’s (1964) unfolding model of choice preference and found that Hershenson’s babies actually preferred bright to intermediate to dim levels of illumination. Independent of the disputed monotonicity of these results, the infants’ differential visual preference strongly indicated that Hershenson’s infants discriminated among different levels of illumination, The newborn’s sensitivity to light has also been investigated by the measurement of reflexes and changes in steady-state activity. Preyer (1890), for example, began to document the newborn’s sensitivity to light quantitatively by studying the pupillary reflex: Preyer reported that in bright daylight his son’s pupils constricted to less than 2 mm. Likewise, Sherman, Sherman, and Flory (1936) used the pupillary reflex as a dependent measure of newborn infants’ sensitivity to light intensity. Peiper ( 1961/ 1963) investigated brightness sensitivity in infants by measuring the eye-neck reflex. When light is shined into an infant’s eyes, the child throws back his head; threshold for the reflex depends on the brightness of the stimulus. Peiper (1961/1963) reported a one-hundredfold increase in sensitivity following dark adaptation. Other investigators have looked at changes in the infant’s gross motility as indicative of his sensitivity to light. Hall ( I 891), for example, noted blinking and manifest movement to changes in sunlight on the first day of life. Watson (1925) discovered visual orienting in infants to light presented in the dark. Still others, notably contributors to the Iowa Studies of Infant Behavior (e.g., Irwin & Weiss, 1934; Redfield, 1939; J . M. Smith, 1936), typically measured gross movement with a stabilimeter. Under these conditions, Irwin and Weiss (1934) and Redfield (1939) found that movements during the first week of life could be elicited in infants with even a very dim stimulus, and Redfield (1939) and Irwin (1941) found that infant movements decreased with simple increases in photopic lighting conditions. More recently, discriminations along the intensity dimension have been examined in infants, and they have been shown to be surprisingly good. Two studies of simultaneous brightness discrimination have investigated the difference threshold (hl) among 2-month-olds. Doris and Cooper (1966) and Doris, Casper, and Poresky (1967) measured optokinetic nystagmus in young babies and thereby derived a difference threshold for brightness. They found that brightness sensitivity at a luminance of approximately 3.4 c d m 2 increased with

Chrotiiutic Vision it1 Infancy

127

age: the Weber fraction for a group of neonates was .50 and that for a group of 3-month-olds was .26. Peeples and Teller (1975) measured brightness discrimination in 2-month-old human infants by utilizing the infants’ visual preference for pattern over homogeneity (Bond, 1972). They showed infants two fields (luminance = 4.0 cd/m2) and observed that the infants preferred the field that contained a bar which differed from the background by only .02 log units. These data show that by 2 months infants are able to discriminate small differences in photopic luminance (Weber fraction = .05), and that infants do not differ greatly from adults. Under similar stimulus conditions, adults can discriminate a 1% luminance difference (Steinhardt, 1936). The detection of simultaneous contrast is more acute in the infant (as it is in the adult) than is the detection of luminance difference without contrast (successive discrimination). Kessen and Bomstein (unpublished observations) studied infants’ discrimination of lights presented in temporal succession. Their results show that 4-month-olds are insensitive to a doubling or halving of luminance. This amount is already considerably larger than the comparable adult changes, 8-10% approximately, but not unexpectedly so since the successive discrimination of brightness necessarily relies both on the infant’s memory and, in the habituation task used, on the novelty value of the luminance change. In this study narrow-band stimuli (A = 490 nm) were used in successive discrimination of luminance; among adults, differential sensitivity functions for spectral lights typically parallel that established for white light (J. L. Brown & Mueller, 1965), and there is little reason to believe that an infant comparison would be different. Studies of brightness perception therefore indicate that even young children can detect small differences in luminance. Certainly, they find brightness a ready cue in discrimination (Clifford & Calvin, 1958). In fact, as we shall see, the infant’s keen sensitivity to small changes in energy has presented the single most frequent stumbling block to the study of chromatic vision proper.

2 . Spectral Sensitivity Spectral sensitivity denotes the basic relation between the eye and visible radiation. Different wavelengths of light vary in the degree to which they stimulate the human eye, and spectral sensitivity is defined as the inverse of the energy necessary at different wavelengths to elicit a constant visual response. It therefore represents a measure of the apparent brightness of the natural spectrum as a function of wavelength, and the term spectral sensitivity is sometimes used interchangeably with luminosity or efficiency. Spectral sensitivities reflect photoreceptor function, and in the duplex retina like man’s, where two classes of photoreceptors (rods and cones) operate, at least two spectral sensitivity functions are prominent. One, associated with the rods, reflects sensitivity at very low light levels and is termed scoropic. Another, associated with the cones, reflects sensitivity at higher light levels and is termed photopic. Maximal scotopic sensitivity occurs in the neighborhood of 505 nm; maximal photopic

128

Marc

H.Bornstein

sensitivity near 555 nm; and a series of curves distributes itself in between at intermediate or mesopic light levels. The change from rod to cone vision that reflects the change from scotopic sensitivity to photopic sensitivity is called the Purkinje shift (Boynton, 1971; Wald, 1945). Knowledge of spectral sensitivity is prerequisite to the study of color vision on two grounds. First, as a basic psychophysical datum, spectral sensitivity defines the limits of vision across the photic spectrum and indexes maturation and function of the primary subsystem of vision. Second, the analysis of color vision per se depends on a knowledge of spectral sensitivity: To isolate hue (or saturation) as an object of experimental study, it is necessary to eliminate or minimize other visual cues, especially brightness differences based on spectral sensitivity. Several attempts have been made to measure spectral sensitivity in human infants, but early research in this area generated substantial controversy. Peiper (1927) obtained the first indication of a Purkinje shift in light- and dark-adapted preterm infants by measuring intensity thresholds for elicitation of the eye-neck reflex. Peiper found that in infants sensitivity values for long visible wavelengths are greater following light adaptation than following dark adaptation. To Peiper these data indicated that a duplex retina is functional before term, and they suggested to him that the newborn is capable of photopic vision. Moreover, Peiper maintained that perceived photopic brightnesses for different colors in infancy and adulthood were similar; in his own words, “brightness sensitivity in man does not develop after birth but. . . the ability to see in daylight and at dawn is present in the premature infant in the same way as is characteristic for adults” (Peiper, 1961/1963, p. 73). Trincker and Trincker (1955/1967) used Peiper’s startle method to perform more elaborate investigations on the characteristics of preterm and term spectral sensitivity. They tested 38 infants 1-10 weeks of age with five colors and found, like Peiper, that threshold sensitivities of the lightadapted newborn were similar to photopic threshold judgments of adults. Trincker and Trincker (1955/1967) also observed that to some extent scotopic sensitivity developed ontogenetically. Pratt, Nelson, and Sun (1930) recorded reactions of infants 1 - 1 1 days old to white and “colored lights” (red, green, yellow, and blue). They counted the number of specific movements per child per stimulation and the average number of millimeters children disturbed a stabilimeter on stimulation. From these measures., Pratt et al. concluded that neonatal sensitivity to light and color was well developed at birth and during the first 2 weeks. Likewise J . M. Smith (1936) measured the effectiveness of three wavelengths to inhibit activity among 20 newborns. J . M. Smith (1936) found that blue suppressed 50% of (baseline) movements, green 30%, and red 23% on the average, and she interpreted these results to iindicate no Purkinje shift. (In fact, J. M. Smith concluded from these “spectral sensitivity” data that infants were to one degree or another color deficient, and boys were more so than girls.) These studies have been criticized in several places (Banks & Salapatek, 1975; Bornstein, Kessen, & Weiskopf, 1976b; Dobson, 1976; Munn, 1955; Peeples &

Chromatic Vision in Infancy

I29

Teller, 1975; Peiper, 1937, 1961/1963; J . M. Smith, 1936; Wooten, 1975). Peiper ( 1927) and J . M. Smith (1936) used a small sample size, and they and the Trinckers (1955/1967) used only a small number of stimuli whose physical characteristics were not well specified. Two failed to report reliabilities of observing an otherwise quite subjective dependent measure. In addition, the comparison adult photopic function that the Trinckers derived in their own apparatus does not match Standard Observer luminosity as defined by the International Committee on Illumination. Smith’s measures, stabilimeter deflections and crying, were very gross and her data so unstable as to question their validity as measures of sensitivity (Munn, 1955; Peiper, 1937). Nevertheless, subjecting her data to reanalysis, Trincker and Trincker (1955/1961) suggested that J. M. Smith’s infants indeed displayed the Purkinje effect. More recent investigators (Dobson, 1976; Peeples & Teller, 1978) have measured photopic spectral sensitivity in infants with greater success. Both looked at 2-month-olds. Peeples and Teller (1978) used a constant behavioral criterion patterned on Fantz’s finding that infants will selectively fixate heterogeneous, in preference to homogeneous, stimulation. To measure white-adapted spectral sensitivity, they determined the radiance necessary to elicit tracking of a moving monochromatic grating on a white background. Dobson (1976) used a constant electrophysiological criterion, implicit time of the visually evoked cortical potential (VEP) (Wooten, 1972), to derive a measure of photopic spectral sensitivity. Data from Dobson (1976) appear in Fig. 2. Neither of these modem experiments is immune to criticism, however. Peeples and Teller’s (1978) observations were, like Dobson’s, derived using only a small number of wavelengths and only a small number of infants, and Dobson’s ( 1976) technique did not account for stray light, to which the VEP is sensitive, or for the infant’s attentional state. Both showed large individual differences. Nevertheless, their mutual concordance and psychophysical sophistication lend them a certain validity. Each of the experiments generated data from infants and adults in the same apparatus and under identical testing procedures. The two studies were, however, conducted under different adaptation conditions. They are, nevertheless, in fairly close agreement. When anchored with adult spectral sensitivity at 550 nm, both show a m.atch between relative adult and infant spectral sensitivities at wavelengths longer than about 550 nm, but the infants’ sensitivity is increasingly elevated at wavelengths shorter than 550 nm. Figure 2 shows photopic spectral sensitivities at three different ages and that of the average tritanopic (blue-deficient) observer (W. D. Wright, 1952). Bomstein (1977a) has termed the apparent ontogenetic change in short-wavelength visual sensitivity “developmental pseudocyananopsia.” As can be seen, with age spectral sensitivity approaches the function characteristic of a formal color defect. The developmental deficiency has a different etiology, however. Before light reaches the retina and becomes an effective stimulus for vision, it passes through several optical structures: the sclera, cornea, lens, aqueous and virtreous

130

Marc H . Bornstein

1

I

I

I

1

I

I

100

f

t

t

;

10

fn

z W

v)

TRITANOPE 1

400

500 WAVELENGTH

600 INM)

700

F i g . 2 . Photopic spectral sensitivities near birth (2 months) (Dobson, 1976), at midlife (26-30 years) (Ruddock, 1965b), and near the end of life ( 6 1 4 5 years) (Ruddock, 19656) in comparison with that of the average tritanope (CV. D . Wright, 1952). (From Bornstein, 1977a. Copyright The American Academy of Optometry. Vsed by permission.)

humors, and the macula lutetz. Because each of these structures is dense, each attenuates incoming light by scattering and absorption. Classic (Ludvigh & McCarthy, 1938) and modem (Norren & Vos, 1974) studies of the attenuation characteristics of the eye agree that intraocular structures combine selectively and cumulatively to absorb and scatter incoming short-wavelength (blue) radiation. In infancy, these optical media are transparent (Marg, Freeman, Peltzman, & Goldstein, 1976). During ontogeny, however, the intraocular structures thicken, increase in pigment density, or both: The sclera becomes denser and yellows (Vannas & Tier, 1960), the cornea yellows (Boettner & Wolter, 1962), and the vitreous (Balazs, 1961), the lens (Coren & Girgus, 1972; Walls, 1942; Weale, 1963), and the macular pigment all thicken (Chevallereau & Polack, 1907; Duke-Elder & Cook, 1963; Weale, 1963). Investigators who have examined these influences of age upon visual sensitivity have generally been satisfied to ascribe resultant differences in the perception of blue to optical pigmentation (e.g., Weale, 1963). Thus, Dobson (1976) suggested that infants’ shortwavelength hypersensitivity reflects primarily their lack of macular pigmentation, and Ruddock ( l965b) ascribed the elderly’s hyposensitivity principally to their excessive lenticular pigmentation. The two conclusions appear to be optically reasonable especially in the light of the foregoing physical and anatomical data

Chromutic Vision in Infancy

131

and Ruddock’s (1965a) studies which rule out age changes in the retinal receptor system. In all probability, however, these two (and other) sources of attenuation combine (Bornstein, 1977a). Future infant researchers may continue to invent behavioral or other techniques to measure spectral sensitivity, such as methods based on a flicker discrimination criterion (Bornstein & Marks, 1972), but this will be substantially difficult with newborns. Two passive procedures based on the electroretinogram (ERG) and pupillary reflex may, however, provide alternative means to calculate spectral sensitivity from neonates. Armington (1979, among others, has demonstrated that photopic spectral sensitivity may be derived from ERGS. However, following a successful, if difficult, experimental demonstration of the photopic ERG in the neonate (Barnet, Lodge. & Armington, 1965), no investigator has yet come forth with a derivation of neonatal spectral sensitivity based on the ERG. A second technique might be based on the pupillary reflex of the eye to light. Iris sphincter constriction is reflexively related to the brightness of a light, and some time ago Baldwin (1895, p. 43, n. 1 ) reported Ladd-Franklin’s (1894, cited in Baldwin, 1895), suggestion, based on Sach, that “reflex changes in the width of the pupil when certain colours are looked at might be used to test the colour sensations of very young children.” For Baldwin, the pupillary reflex represented a variation on the reaching theme. Peiper (1927; see Peiper, 1961/1963, p. 75) referred obliquely to experiments along these lines by De Rudder, but not until Mcnsinger and Banks (1974) was pupillometry tried as “an objective measure of visual sensitivity for infants” (p. 677). Testing over 4 log units of energy with three narrow-band wavelengths (486, 577, and 671 nm), these investigators showed that pupillometry was reliable and closely related to perceived brightness in 3-year-olds and in adults. One-year-olds in their study showed equal responsivity to 486 and 671 nm but more to 577 nm. In a follow-up study these investigators actually derived a photopic spectral sensitivity function pupillometrically from a 4-year-old (Banks & Munsinger, 1974). That subject showed a consistent decrement in long-wavelength (A > 546 nm) sensitivity and a small increment in short-wavelength (A < 500 nm) sensitivity relative to adults. As suggested, studies of infants’ spectral sensitivity are important and informative for two principal reasons. First, they provide a kind of psychoanatomic diagnosis. Photopic sensitivities indicate mature, color-normal functioning of retinal elements, and when measured by criteria such as VEP they suggest that cortical functions in the infant have developed. Indeed, behavioral measures of spectral sensitivity in monkeys indicate that only the scotopic system functions in the destriate preparation (Lepore, Cardu, Rasmussen, & Malmo, 1975). The behavioral and electrophysiological data presently available indicate that photopic spectral sensitivity operates normally by the second month of life. The same cannot yet be said of the neonatal period. Second, some measure of spectral sensitivity is critical to the accurate evaluation of other visual functions in chromatic perception. In many color studies, an

132

Marc

H.Bornstein

average measure of adult spectral sensitivity is used to equate stimuli for brightness for infants. In view of such widespread application, several caveats are in order. First, measures of spectral sensitivity are influenced by several parameters, including the criterion level, procedure, adaptation, state of the organism, and location on the retina. In addition, spectral sensitivities represent only single individuals or at best the central tendency of a small group of observers. Consequently, they do not apply equally well to each observer in a given study, nor do they apply to all observers, nor to every study. Finally, spectral sensitivities are not brightness matches pe:r se even if the two sets of data are closely related. Since “brightness matches” are usually suprathreshold and depend at minimum upon good experimenter-observer communication; matches are not likely to be derived from young infants. In practice it is unlikely that researchers will be able to derive individual spectral sensitivity functions from younger observers preliminary to color studies. The utility and hazards of applying mean spectral sensitivity measurements may, then, be obvious.

3. Summary Determination of brightness idiscrimination and spectral sensitivity describe two basic visual functions related to color perception. Studies reviewed in this section indicate that under conditions of simultaneous contrast, 2-month-olds, like adults, discriminate small brightness differences, but that they have not been shown to approximate adult behavior in situations that involve successive discrimination. Relative spectral sensitivity among 2-month-olds deviates from that of the adult only at the shortest visible wavelengths. The infant’s remarkable brightness sensitivity and his departures from adult sensitivity suggest that simultaneous discrimination of chromatic stimuli purportedly equated for luminance may not preclude responses to residual brightness differences instead of color differences. Data on spectral sensitivity are prerequisite to the study of color vision though they do not yield any information about the existence of color vision per se. Monochromatic, dichromatic, and trichromatic eyes all evidence roughly similar spectral sensitivity curves. Thus, establishment of spectral sensitivity in early infancy does not afford evidence that the newborn actually sees color. The next section first reviews several key studies which were specifically designed to demonstrate chromatic vision in infancy, that is discrimination between two stimuli differing in wavelength composition on the basis of that difference alone. A review of recent experiments of basic color visual functions then follows. B. CHROMATIC VISION

1. Attempts to Demonstrate Chromatic Vision in Infants Not all researchers have ignored the brightness problem. Typically, two strategies to eliminate the brightness confound have prevailed. In one, chromatic stimuli are equated for perceived brightness by adult spectral sensitivity coeffi-

Chromcrrrr Vision in Infancy

I33

cients. (At this writing, no researcher had adopted infant coefficients.) In the second, brightness controls are not based on prior assumptions about the infant's spectral sensitivity, but brightnesses of test stimuli are varied randomly and orthogonally to chromatic cues. Some researchers used visual discrimination paradigms-preferential paired comparisons and instrumental learning-that were intended to go beyond simple preference; others used electrophysiological measures. Notwithstanding these advances, the thrust of this set of research papers has been to demonstrate simple chromatic discrimination in infants. The main questions have been: Do infants see color'? and, When do they see it? In general, the scope of such an approach is limited since i t tells us little about the type, quality, or functions of such vision. Moreover, each of the following studies is, to a certain extent, inconclusive. Chase (1937) correctly intuited that infant tracking of a chromatic target against equally bright achromatic or chromatic backgrounds would evidence color vision. He therefore projected mobile chromatic discs (7.2" X 10.8') on a ground (53.1" X 40.1") and had two observers judge eye movements of 24 infants 15-70 days of age. Using a variety of color combinations, Chase found that disc-field color contrasts yielded pursuit on 90-100% of the trials. When he paired different intensities of the same hue i n a control condition, however, infants failed to pursue, and supposedly to differentiate, the target. Chase might have increased brightness in the control condition until tracking occured to provide proof that the brightness did not control pursuit in the hue-contrast condition; since his colors were equated for brightness, however, Chase's study stands as the first experimental indication that babies might distinguish hues. Using optokinetic nystagmus as an index, Trincker and Trincker (1955/1961) tested the hypothesis that infants would not discriminate hue. Over 100 babies between 1 and 11 weeks of age were exposed to a moving field of alternating colored and gray stripes. Babies 1 or 2 weeks old failed to show a response to colored and gray stripes of equal photopic brightness; by the third week blue was discriminated, then red, yellow, and green in succession. As part of a larger study of early visual preference, Spears (1964) measured the relative fixation times of 60 4-month-olds to pairs of equally bright Munsell Hues (22.4" squares of blue, yellow, red, and gray). Blue, red, and yellow were equally preferred, but blue and red were preferred to gray. Since his colors were matched for brightness, Spears concluded that 4-month-olds' preferences indicated discriminations among color patches not based on distinctions of brightness, but based on distinctions of hue. Fagan (1974) studied 124 infants 13-25 weeks of age. His stimuli were pairs of checkerboard patterns that subtended 24' of visual angle. One member of each pair alternated two Munsell Hues of equal Value and Chroma (4.8" square), and the other member was a single-hue pattern. Fagan's thesis, that hue discrimina-

134

Marc H . Bornstein

tion would be evident in infants’ preference for double-hued checkerboards, was based on previous demonstrations that infants prefer patterned to unpatterned stimuli, and indeed infants i n Fagan’s study looked longer at the two-hue patterns. Further, Fagan’s data suggested that fixation time was directly related to the degree of difference among hues; for example, infants looked longest at a red-green checkerboard. Fagan (1974, p. 974) concluded, therefore, that “infants are capable of discriminating on the basis of hues by 4 to 6 months.” Studying heart rate activity in newborns ( 1-4 days of age), Adkinson and Berg (1976) used two equally luminant colors, a blue (A = 472 nm) and a blue-green (A = 496 nm) as discriminative stimuli. They found in the older newborns that six exposures to one color resulted in reliable habituation of onset decelerations, and then two exposures to the other color elicited reliable dishabituation of onset declerations. These data provide some evidence of simple discrimination of color in the newborn infant. Brightness controls in these studies have proved problematical. Chase, the Trinckers, Spears, Fagan, and Adkinson and Berg all assumed that special sensitivity is the same in infants as in adults. Although, as we have seen, infant sensitivity is a good approxirnation of adult sensitivity (particularly at middle and long wavelengths), Peeples and Teller (1975) have shown that in simultaneous discrimination situations, such as color or brightness contrast, infants are so exquisitely sensitive that even small differences in brightness may provide sufficient discriminative cues. In many of these studies, using adult sensitivity to match brightnesses may not provide adequate stimulus control. In addition, Wooten (1975) argued that Fagan (1974) used inappropriate illumination on Munsell papers, which may have provided extraneous brightness cues even if relative spectral sensitivities of infants and adults were equivalent. Wooten then showed how Fagan’s results could be accounted for in terms of brightness differences among the ~timulli.~ Fagan (1975) countered Wooten’s criticism by pointing out that partial coirelations of Fagan’s original data substantiated the predominance of hue over brightness in determining preference, and Fagan showed further that hue preferences were maintained in follow-up tests in which colors were presented under appropriate (C) illumination. Beside the facts that Spears assumed infant-adult parity with regard to spectral sensitivity and did not specify the illuminant of his Munsell chips, his preference technique quite severely limited the conclus~ionsin his study. Although the pattern-preference paradigm provides an extraordinarily valuable index of discrimination, a lack of preference among chromatic stimuli [as, e.g., in Spears (1964) or Wickelgren (1967)l is consistent with but does not confirm a lack of discriminative capacity. 3The Wooten-Fagan exchange hauntingly echoes Eldridge-Green’s criticism of Marsden ( 1 903a, 1903b) some 70 years earlier. Martiden’s studies of infant color perception were based on Baldwin’s paired-preference method; Eldridg,e-Green suggested that the young child‘s “photoaesthetic and visive perceptions” of color were actually based on perceived brightness differences.

Cttronrcitic Vision in Infumy

I35

Peeples and Teller (1975) also based their assessment of infant color vision on the visual-choice paradigm. They modified their brightness discrimination approach and showed two 2-month-old girls a white screen on either side of which a red bar (A = 633 nm) of variable luminance could appear. The infants consistently discriminated the presence of the red bar at different luminances, one of which presumptively matched the white ground in brightness. Oster (1975) selected a camouflage strategy. She adopted von Frisch’s (1964) technique (as recommended by Peiper, 1961/1963, p. 71) of systematically varying the brightness context of a target stimulus and found that 10-week-olds will discriminate chromatic from achromatic patterns. Several conditioning procedures have involved hue discrimination in infants. Peiper cited Russian investigators, including Krasnogorski (1913, cited in Peiper, 1961/1963) and Rahlmann (1903, cited in Peiper, 1961/1963), who successfully conditioned sucking to colored milk bottles. Brackbill (1962) reported Zonova’s comparative studies of sucking and eye-blink conditioning to “red, green, and blue lights.” Fewer than 50% of the subjects discriminated at 3 months when the index response was sucking, while virtually 100% reached an analogous eye-blink criterion at 55 days. In 1935, Kasatkin and Levikova successfully conditioned a differential visual discrimination between red or yellow and green in six infants approximately 3.5 months of age. Although they were not interested in color discrimination per se, Kasatkin and Levikova concluded from this demonstration that infants can discriminate color. In a recent and worthy study, Schaller ( 1975) operantly conditioned differential eye movements in six of eight 3-month-olds to red and green stimuli that varied irregularly over a wide range of brightnesses. The quality and quantity of stimuli used in many of these infant studies have provoked consistent criticism. Most investigators used only two or three stimuli. The Russians used unspecified color bulbs, and Schaller (1975) apparently selected complex, wide band, bipeaked color filters. Like Chase, Schaller might also have varied the brightness of his stimuli until he was certain brightness was not a factor in conditioning. Finally, Milewski and Siqueland’s (1975) recent conditioning study is subject to the same brightness criticism. They instrumentally conditioned high-amplitude sucking in 10 1month-old infants with red (or blue) patterns and found that following habituation infants would dishabituate to (discriminate) a change to blue (or to red). However, Milewski and Siqueland’s red and blue slides differed by .5 log unit in luminance and therefore allowed discrimination to be based on brightness. Some investigators have used electrophysiological measures to investigate chromatic vision in infants. Such measures, which reflect the relationship between stimulus processes and electrochemical changes in nervous tissue, include the ERG and the VEP. Neither the ERG nor the VEP is particularly suited to studies that involve variation in wavelength. However, the wavelength parameter has been investigated in adults and in infants and, independent of other difficulties associated with gross electrophysiology , the results are illuminating. In

I36

Marc H . Bornsrein

general, wavelength affects the ERG only modestly but in response to longwavelength light, the adult ERG displays an early positive x-wave (Adrian, 1945). Bamet et a l . (1965) and Lodge, Armington, Barnet, Shanks, and Newcomb (1969) recorded this characteristic x-wave in ERGS from newborn infants exposed to “orange” light (A > 580 nm). They deduced from the existence of the x-wave that the retina in the human newborn is capable of photopic responses and that newborn humans possess chromatic vision. To these retinal data, Lodge et al. (1969) added neonatal VEPs recorded over occipital cortex. In their study. newborns, like adults, showed higher amplitude VEPs to orange than to white light, a result characteristic of photopic activity. Since Armington ( 1975) had demonstrated that the occipital potential in the adult primarily reflects foveal activity, from the Lodge data one may further deduce that the central retina functions at or soon after birth. Infantile evoked potentials to white as well as red, green, and blue stimuli were measured by Fichsel(l969). Preterm infants showed essentially no difference in waveform or latency to the different stimuli, but term newborns and older children evidenced systematic distinctions. Finally, Polikanina (1968) found different patterns of EEG in 2-week-olds to red and green Ilight. Thus, data from Lodge et al. (1969), from Fichsel (1969), and from Polikanina (1968) provide evidence that visual cortex functions soon after birth, and they implicate it in some higher order, central analysis of chromatic stimulation. These and behavioral data on color vision reviewed below specifically contradict Bronson ( 1974), whose view is that geniculostriate pathways are nonfunctional during the first month of life. In summary, although electrophysiological data are suggestive of chromatic vision in infancy, they do not prove behavioral function. Many of the behavioral studies of color vision in infants are, as we have seen, plagued by criticisms of stimulus control. Several of these studies are individually questionable; together, however, the better studies (e.g., Peeples & Teller, 1975; Schaller, 1975) provide strongly suggestive evidence of color vision. None of the studies cited above provides information about the quality, extent, or type of color perception in infancy. As a consequence, even the best investigations have been limited to claims that infants “must have some form of color vision” that is “at least dichromatic” (Peeples & Teller, 1975, p. 1102). More extensive data are necessary to delimit the type or qualiity of color vision that an organism possesses. The next sections review such data.

2 . The Neutral Zone Research with adults indicates that color-normal trichromats, possessing at least three cone types, see all wav’elengths in the spectrum as hued, saturated, and discriminable from white, but that color-deficient dichromats, possessing only two cone types, see at least one point in the spectrum in chromatic balance and therefore as achromatic or gray. Deficiencies in chromatic vision are typically

Chromatic Vision in Infancy

I37

characterized, indeed they are diagnosed, by the presence of this achromatic “neutral zone.” According to Hsia and Graham (1965, p. 396), the neutral zone, “is an important characteristic of dichromatic vision.. . . A true dichromat has no difficulty in selecting a spectral color to match white, a performance that is impossible for the normal trichromat.” Recently, Bornstein (1976a, p. 429) surveyed empirical and theoretical derivations of neutral points for the two major subclasses of dichromasy, protanopia and deuteranopia. For protans, estimates of neutral points extended from 489.0 to 497.0 nm, and for deutans they extended from 495.0 to 500.5 nm. For protan and deutan alike, however, a somewhat larger spectral region is desaturated and whitish in appearance. “The neutral point obtained in actual measurements is not invariably a sharp point but usually covers a range of several millimicrons [nanometers]” (Hsia & Graham, 1965, p. 396). Meyer (1932) estimated that the neutral zone for dichromats is 5.4-7.0 nm in width, and Kalmus (1972) obtained widths of 4-8 nm. The achromatic appearance of this region to a dichromat was quite strikingly evidenced in the color-naming behavior of the unilateral dichromat (C. H . Graham, Sperling, Hsia, & Coulson. 1961). This observer named 490 and 500 nm “blue-green” and “green, blue-green’’ respectively when viewed with her normal eye, but she named them “mostly grey, a little blue” and “yellowish grey” when viewed with her dichromatic eye. For these reasons, no doubt, several investigators of sensory processes have labeled this region the dichromats’ “neutral band” (Walls & Matthews, 1952), “grey band” (Geldard, 1972), or “neutral zone” (Kalmus, 1972). The existence of a neutral region is a diagnostic sign of dysfunction of chromatic vision, and its absence is diagnostic of trichromatic vision. It has been used to assess chromatic vision behaviorally in various infrahuman primates (Boothe, Teller, & Sackett, 1975; De Valois, Morgan, Polson, Mean, & Hull, 1974; Grether, 1939) and in human infants. Teller and her associates (Peeples, 1976; Peeples & Teller, 1975; Teller, Peeples, & Sekel, 1978) used a visual-choice paradigm, and Bornstein (1976a) used an attention-habituation paradigm, to discern whether or not infants could discriminate different spectral wavelengths from white. Bornstein’s habituation test was designed specifically to assess whether young infants would discriminate wavelengths in the region 490-500 nm from white of the same mean luminance, that is, to test whether infants would differentiate achromatic stimulation from chromatic stimulation selected from the spectral region which is perceived as achromatic in dichromatic vision. Infants 3 months old were first exposed on 12 familiarization trials to varying monochromatic lights selected randomly from the 490- to 500-nm wavelength range. During familiarization, stimulus luminance also varied randomly over nearly a log unit. White light (5100”K), whose luminance was held constant at the mean value for the spectral luminances, then appeared on two consecutive trials. The sequence

Marc H . Bornstein

138

ended with two trials on which spectral lights chosen under the same constraints as the first 12 appeared. Since the varying luminance of the familiarization stimuli had as a mean the luminance of the white stimulus, dishabituation to white could not be attributed to a change in brightness, as opposed to chromatic change. Analyses of the infants’ visual attention in terms of total looking time per trial and duration of first fixations yielded parallel patterns of responding. Figure 3 shows habituation (the decrement in looking) to repeated presentation of wavelengths in the 490- to 500-nm range, statistically reliable dishabituation (recovery of looking) to white, and rehabituation to wavelengths in the 490- to 500-nm range. The data indicate that even in the absence of brightness cues 3-month-olds can discriminate: white light from spectral light selected from the protanopic and deuteranopic neutral zone. These results strongly suggest that infants of this age are not classical dichromats, but may possess ‘trichromatic vision. Teller and Peeples (Peeples., 1976; Peeples & Teller, 1975; Teller et al., 1978) used infants’ visual preference for patterned stimulation to study whether or not they saw a neutral zone. Essentially, these investigators created a two-alternative forced-choice situation where a colored bar could appear embedded in a white screen in either the infant’s left or right visual field. They then varied the luminance of the bar systematically .4 log units around the luminancc of the background screen. For differlent infants they varied the wavelength composition of the bar. Their observations showed that 2-month-olds looked at blue, bluegreen, orange, red, and nonspectral purple bars in preference to the white screen

1-2

5-6

9-10

TRIALS

13-14

(w)

F i g . 3. Mean durations of infants’ total a d j r s t j x a t i o m on serial presentations of spectral (490 nm < A 5 500 nm) and white (W)light. The data are averaged over pairs of consecutive presentations. (From Bornstein, I976a. Copyright Academic Press, Inc. Used by permission.)

Chromatic Vision in Infancy

I39

with no pattern. Most important among these is the blue-green ( A = 496 nm) since, were human infants classical protanopes or deuteranopes, they would not be able to distinguish a blue-green bar against a white ground at every level of l ~ m i n a n c e Curiously, .~ some infants could not or would not pick out green or yellow-green bars in preference to white. On the basis of their results with this method, Peeples and Teller tentatively concluded that infants 2 months of age see a neutral zone in the green to green-yellow region of the spectrum. No classical dichromatic type can be characterized by such an achromatic placement; however, irregular neutral zones are not unknown. Color vision defects acquired in diseases of the optic nerve, for example, show a variety of neutral zones across the spectrum (Griitzner, 1972). The zones in acquired deficiencies tend to be quite wide, though wider than infants in the present study displayed. Two alternative explanations for this infantile behavior might be entertained. First, infants younger than 3 months may not be dichromatic but may be otherwise color weak; in specific they may be deuteranomalous trichromats, since insensitivity to the middle (green) region of the spectrum is endemic to deutan deficiencies. The trichromat’s specbum is more saturated overall than that of the dichromat or anolmalous trichromat (Bornstein, 1976a; Chapanis, 1944, Table IV). Consequently, young deuteranomalous trichromats might confuse midspectral greens with white. In support of this view, morphological studies show small differences between immature and mature retinal cones (Duke-Elder & Cook, 1963). Alternatively, infants may not like to look at these two colors. Indeed, a general result of color preference studies suggests that infants specifically avoid green and green-yellow (e.g., Bornstein, 1975d) (see Section 11, B, 6 ) . Thus; infants’ lack of preference for midspectral hues may have influenced their responding in the Peeples and Teller situation. In 2-month-olds, however, it is not yet possible to distinguish between these alternatives: Since the infants will look only very little at a greenish light, they give no indication that they can discriminate it from white which they also regard only very little. In summary, results of these two studies of the neutral zone suggest that human infants by 2-3 months of age are not classical dichromats (i.e., protanopes or deuteranopes) though they may be color weak. The results further suggest though that human infants may be color-normal at 3 months. Certainly additional research is desirable to clarify the nature of chromatic vision near birth. “Boothe el al. (1975) applied an operant discrimination paradigm to light- and dark-reared infant monkeys. They reported that, with brightness systematically varied and attention strategies controlled, macaques younger than 2 months of age discriminated all wavelengths from white, and that one monkey did so even though it had been light deprived between the second week and the third month of its life. These results provide an ontogenetic baseline for De Valois and De Valois’s (1975) research with adult macaques, which exhibit trichromatic vision like adult humans. Additionally, the Boothe et al. results provide an animal model for future work in the ontogeny of primate color vision.

140

M a r c H . Bornstein

Of course not all adults are trichromats. As infants, therefore, dichromats should give evidence of a neutral zone in their spectrum. Since dichromasy can be predicted on the basis of genetic linkage-for example, half the sons born to mothers whose fathers were dichromats should be dichromats themselves (Hsia & Graham, 1965)-it should be possible to correlate the results of behavioral approaches with genetic determinations. As part of the studies just described, Bornstein (1976a, n. 4) and Peeples (1976) tested infants who were identified as potential dichromats on the basis of family history. Although the results are themselves quite tentative because of diagnostic inadequacies and the failure to conduct the studies in an experimentally blind manner, some presumptively color-deficient children in both studies failed to discriminate white from wavelengths in the range 490-500 nm. This result gives some indication, then, that like methods could be useful (1) in further studies of the earliest ontogeny of chromatic vision and ( 2 ) in the identification and early diagnosis of genetically indicated dichromasy .

3 . Wavelength Discrimination When the spectral sensitivity of an organism is known so that brightness differences among wavelengths can be controlled, it is possible to ascertain how many different wavelengths the organism can discriminate. (The absolute number will actually depend upon a threshold criterion and upon stimulus conditions.) Derivation of a wavelength-discrimination (AX) function has threefold importance. First, the wavelength-discrimination function reflects differential discriminability of available visible radiation and thus defines the limits of visibility. In man, wavelength discrimination is best (AA is small) near 490, 565, and 610 nm. Wavelengths shorter than 490 nm and longer than 610 nm are increasingly i ndi sti ngui sh able . Second, wavelength-di scrimi nati on capacities must reflect the ontogenetic maturation and function of the visual system since lateral geniculate cell sensitivities have been directly implicated in discrimination capacity (De Valois, Abramov, & Mead, 1967; De Valois & De Valois, 1975). Finally, wavelength discrimination and hue perception are strongly related: Wavelength regions at which adults show their best discrimination correlate strongly with wavelength regions that represent the changes between hues, and regions where adults show their poorest discrimination reflect regions where hue perception is unchanging (e.g., B. V. Graham, Turner, & Hurst, 1973; D. P. Smith, 1971; W. D. Wright, 1947). The wavelength-discrimination characteristic of adults has been known for quite some time (Siegel, 1962, 1963, 1964; Siegel & Dimmick, 1962; W. D. Wright, 1947), and studies of several infrahuman organisms including the bee (von Helversen, 1972), the pigeon (A. A. Wright, 1972), and the monkey (De Valois & De Valois, 1975) have been reported. But no complete studies have been published of discrimination capability of the developing visual system in humans.

Chromatic Vision in Infancy

141

In general, investigators interested in ascertaining discriminability along the dimension of wavelength have utilized simultaneous-discrimination paradigms. Chase's ( 1937) stimulus-superposition, movement-detection paradigm tested simultaneous wavelength discrimination, though he made no attempt to measure minimal or constant discriminability. Teller and Peeples' ( 1974) patternpreference paradigm could as well be adapted to a simultaneous wavelengthdiscrimination task. Additionally, the attention-habituation technique lends itself to such a test. In a preliminary investigation, Bornstein (unpublished observations) used the attention-habituation paradigm to study simultaneous wavelength discrimination in one 5-month-old infant. First, the child habituated to one wavelength that appeared simultaneously in two juxtaposed fields (each 10") on 15 15-second trials. Then, on several subsequent test trials light in one field was varied at 5, 10, or 15 nm longer than light in the second field, which was held constant at the habituated wavelength. All wavelengths had been matched for luminance by Judd's ( 195lb) modification of Standard Observer spectral sensitivity, and the test wavelengths appeared an equal number of times in the left and right fields to control for effects of any position preference. The spectral loci of habituated wavelengths were varied on different days. Five spectral regions were explored; they correspond to wavelengths on the adult wavelength-discrimination function that represented two minima and three maxima. The infant in this study was observed generally to look more at wavelengths increasingly different from an habituated wavelength that corresponded to a point of good discriminability, and he was observed generally to distribute his attention equally among the series of test wavelengths that fell in regions of poor adult discriminability. That is, the infant in this preliminary study demonstrated wavelength discrimination similar to that of an adult trichromat. Other studies of simultaneous hue discrimination with children complement the infant research. Gaines ( 1972) tested hue discrimination in 47 kindergarteners (mean age 5.7 years) in a simultaneous oddity task with narrow ranges of Munsell Hues. She assessed subjects' errors and decision latencies when Value and Chroma were varied. Gaines determined that children found it easiest to discriminate among yellows and among oranges (which had 15 and 22% error rates), harder among purples and among blues, and most difficult among reds and among greens (with 42 and 43% error rates, respectively). Discrimination latencies followed the same pattern: Yellow and orange took 4.1 and 4.2 seconds, while red and green took 6.3 seconds each. A clear explanation of these hue differences lies in native discriminative capacity: Wavelength discrimination is superior in the yellow and orange and less acute in the central green and red regions of the spectrum. Although the children in this first study led Gaines to believe that the young are quite skillful at hue discrimination, i n a cross-sectional follow-up study Gaines and Little (1975) unearthed a developmental trend. There, fifth-graders

142

Marc

H . Bornstein

( 10.6 years), high-school sophomores (15.8 years), the kindergartener’s parents

(39.2 years), and a matched adult group of “master artists” (44 years) showed linearly increasing hue discrimination overall, although the error ranking by hue within each group was the same as had been found among the kindergarteners. Thus, the quality of hue discrimination may increase somewhat with age on account of learning, attention, or perceptual development, but discriminability continues to reflect biologically given wavelength-discrimination capacities. Other developmental studies of simultaneous wavelength or hue discrimination (e.g., Gilbert, 1957; Lakowski, 1958-1959; Verriest, 1963) have uniformly demonstrated a decline in discriminability, particularly in the blue, in old age. The results of these studies reflect the phenomenon of developmental pseudocyananopsia discussed above (Bornstein, 1977a). The transposition of discriminative judgment across space or time will certainly result in degraded discrimination performance, but results with successive procedures have been found to parallel the basic form of simultaneous wavelength discrimination. Bornstein ( 1976a) tested eight 3-month-olds for their ability to resolve differences imong three fixed wavelengths, 560, 570, and 580 nm. The three stimuli in this study were matched for luminance (at 3.4 cd/m*) by Standard Observer sensitivity; as noted above, Dobson (1976) and Teller and Peeples (1974) determined that the 2-month-olds’ relative sensitivity in this spectral region matched that of adults and was typical of the Standard Observer. In this successive-discrimination design, infants were shown a wavelength, 570 nm, serially on 15 trials during a familiarization phase, which was immediately followed by a test phase of nine additional trials. The test trials were divided into three triplets in which the thrce wavelengths, 560, 570, and 580 nm, appeared in one random order, then in a second, and finally in a third. Each infant saw different random orders of the test stimuli, and the infant’s attention was redirected with an auditory prompt to the position of the test stimulus prior to each stimulus onset. All trials lasted 10 seconds, and the intervals between stimuli averaged 5 seconds. The habituation criterion was looking on Trials 13-15 less than 80% of the maximum looking in the preceding three-trial block. Figure 4 shows the results of both the habituation and test phases. As can be seen, infant looking showed rapid habituation over the last three familiarization trials. Looking during the nine test trials has been reduced to three-trial blocks to correspond to the three test wavelengths. The mean times shown in Fig. 4 indicate a statistically reliable dishabituation to 560 nm relative to 570 and 580 nm. The results of this study thus suggest that 3-month-old infants habituated to 570 nm will, in a successive test, discriminate 560 from 570 nm, but not 580 from 570 nm. Although simultaneous wavelength discrimination among adults is quite good throughout this region of the spectrum, successive discrimination is somewhat poorer and there is a phenomenal hue difference among the three wavelengths. For the adult trichromat, 560 nm appears greenish, while 570 and

570 nrn W I I-

560 nm

5

I I I

I 1-3

4-6

I

10-12

7-9

I

I T

13-15

TRIALS

Fig. 4 . Mean durations of infants' fotul fixcltions on serial presentations of an habituation wavelength (570 nmj and three test (Tj wai,elengrhs (560, 570, and 580 nm). (From Bornstein, 1 9 7 6 ~Copyright . Academic Press, Inc. Used bv permission.)

580 nm are yellowish (Beare, 1963; Boynton & Gordon, 1965). In summary, after habituation to a predominantly yellow light (570 nm), infants looked significantly less at yellows (570 and 580 nm) than at a predominantly green light (560 nm). In his successive discrimination capacity, the infant qualitatively resembles the adult. In summary, developmental studies of simultaneous and successive discrimination of hue are few, but they tend to suggest that the adult form of the wavelength-discrimination function is present early in infancy. This would be expected on the view that wavelength discriminability actually reflects native visual system function, rather than learning or experience (Kopp & Lane, 1968). Studies with older children and adults also suggest that discriminability improves somewhat until old age when it may show decrements selective to short wavelengths. The foregoing data on wavelength discrimination help to define further the state of chromatic vision 3 months after birth. The protanopic or deuteranopic eye matches all visible wavelengths longer than about 500 nm (500-700 nm) with a wavelength seen in the normal eye lying near 570 nm (yellowish) and all visible wavelengths shorter than about 490 nm (400-490 nm) with a wavelength seen in the normal eye lying near 470 nm (blue) (C. H. Graham & Hsia, 1958; Hurvich, 1972). These classical dichromats, therefore, do not differentiate green from yellow the way trichromats do; that is, they do not see and respond to wavelengths just shorter than 565 nm as similar and yet as qualitatively different from wavelengths just longer than 565 nm. The 565-nm boundary is exclusive to the trichromat's green and yellow hue distinctions. The young infants' discrimination of

144

Marc H . Bornstein

560 from 570 and 580 nm (Bornstein, 1976a) tends therefore to corroborate by an independent converging procedure results from neutral-zone studies, and it confirms that by 3 months postpartum, infants are probably trichromatic observers. 4 . The Organization of Hue The psychological correlate of stimulus wavelength is perceived hue. Spectral hues derive from the prismatic dispersion of white sunlight, and discriminable spectral hues have been estimated to number approximately 128. Wavelength discriminability varies nonmonotonically along the wavelength continuum, and discriminability correlates strongly with the qualitative variations of hue. To quote W. D. Wright: The [wavelength discrimination] curves as a whole have the characteristics that might be anticipated from a qualitative examination of the spectrum. The part of the spectrum where a minimum exists must obviously occur where there is a rapid change of hue; thus in the yellow where the color turns redder on one side and greener on the other, in the blue-green where it turns bluer on one side and greener on the other, in the violet where it becomes redder or bluer, minimum steps would be expected. But beyond [610 nm] the colour changes steadily to a deeper and deeper red, and in the green where there is only a gradual change to either a blue-green or a yellow-green, the discrimination is poorer and the step consequently greater. (W. D. Wright, 1947, pp. 117-1 18)

W. D. Wright’s observation restates the inverse relationship between generalization (or categorization) and discrimination (Lashley & Wade, 1946; A. A. Wright, 1972). Although wavelength generalization and discriminability do not coexist in a perfectly correlated way, the relationship seems logical and, in the opinion of many investigators, empirically valid (e.g., Ekman, 1954; B. V. Graham et a l . , 1973; Jacobs 1Pr Gaylord, 1967; D. P. Smith, 1971). Because spectral hues are salient, interest in their differentiation has been sustained from ancient through modem times: Aristotle discussed the distinct color qualities of the rainbow (Loveday & Forster, 1913), and Newton (16711672) marvelled at perceptual discontinuities in the nature of light. Hue has served for some time as a focus of epistemological inquiry (Bomstein, 1975b), but only recently has the organization of hue come under direct psychophysical investigation in identification and discrimination tasks (Bornstein, 1975b; C . H. Graham, 1965). The results of several methods as diverse as color naming (Beare, 1963), color scaling (Boynton & Gordon, 1965), and the estimation of chromatic similarity (Ekman, 1963) confirm that the relationship between hue and wavelength is regular. When centrally viewed, at moderate durations and luminances, and under conditions of neutral adaptation, the spectrum is perceived to display a discontinuous, though orderly, array of hues. From the short to the long wave end, four principal qualities are commonly identified: blue, green, yellow, and red.

Chromatic. Vision in Infancy

145

Wavelengths of light differ from one another equally in physical terms; psychologically, however, some are unique (Berlin & Kay, 1969; Bornstein, 1973; Dimmick & Hubbard, 1939a, 1939b; Sternheim & Boynton, 1966). Unique hues represent psychologically pure and unmixed sensations; other colors in and outside the spectrum represent combinations of one or more unique hues green, and and may be so analyzed visually. Three “unique” hues-blue, yellow-are spectral; their dominant wavelengths fall approximately at 475,5 15, and 582 nm, respectively (Boynton, 1975; Dimmick & Hubbard, 1939a). All long wavelengths look yellowish; therefore, unique red is extraspectral, that is, a mixture of long and short waves, which cancel yellow (Dimmick & Hubbard, 1939b). Bornstein (1973, 1975b) has reviewed several lines of evidence that suggest that blue, green, yellow, and red are biologically, psychologically, and linguistically unique and meaningful. Although the three dimensions of the color solid generate more than 7,500,000 colors (Nickerson & Newhall, 1943), only four hue terms-blue, green, yellow, and red-are required to provide a comprehensive yet parsimonious qualitative description of the spectrum (Beare, 1963; Boynton & Gordon, 1965; Ekman, 1963; see also Gothlin, 1943, 1944; Sternheim & Boynton, 1966). In a color-naming task, when only basic terms such as red or yellow are allowed to describe a given wavelength, wide regions of the spectrum yield plateau-shaped naming functions (Beare, 1963). Narrower ranges i n between are marked by competing, complementary, and symmetrical use of two terms. When a combination of two basic terms is allowed, peaked generalization-type gradients result (Boynton & Gordon, 1965). Boynton and Gordon’s more sensitive color-naming technique demonstrates important facts about hue perception. First, hue “category” effects depend on method; increasing the number of allowable terms reduces the absoluteness of the category. This effect indicates, second, that adults can discriminate among wavelengths within hue categories, a fact not evident in Beare’s identification data but clearly manifest in wavelength-discrimination data. “Categories” of hue seem to exist, nevertheless, since color naming reflects the fact that discrimination among wavelengths within a hue is less acute than discrimination of wavelengths between hues. One of the common characteristics of color normalcy, then, is the adult’s predisposition to partition the spectrum into four regular and basic hue categories. Regardless of method, color-normal adult observers place hue boundaries near wavelength-discrimination minima. Since it has been suggested that infants 3 months of age also possess trichromatic vision, studies of the successive discriminati on of wavelengths should reveal that they also partition the physical spectrum into qualitative categories like hues. One indication that infants might distinguish less well within a hue than between hues came from Fagan (1974). Infants in his study looked longest at (discriminated best) checkerboards constructed of different hues, such as red and

I46

Marc H . Borrisrein

green; they looked less at checkerboards constructed of similar hues, such as red and yellow-red; and they looked least at checkerboards constructed of two samples of the same hue, such as two different reds. Bornstein et al. (1976b) used the attention-habituation paradigm with monochromatic spectral lights to assess the organization of hue in young infants. Fourteen different groups each with 10 4-month-olds looked at 10” spectral stimuli repeatedly until their visual attention waned. These habituation stimuli were selected from the four basic adult hue categories-blue, green, yellow, and red. After 15 habituation trials with the one wavelength, infants were shown a series of 9 test wavelengths that included 3 presentations each of the original light and two new lights. Several groups saw the original, a second wavelength selected from the same adult hue category, and a third wavelength selected from a different aduh. hue category. Other groups saw the original and two other lights selected from the same hue category as the orginal. For both types, the new wavelengths presented in the test phase were carefully selected to be equal physical distances (in nanometers) from the original habituation stimulus, in order that the infants’ differential disinhibition of attention to new lights could be attributed only to the psychological dissimilarity of the new and original hues. The original habituation hue appeared randomly in the test series, and statistical contrasts were made among all stimuli in the test phase. These two provisions contributed stringent intrainfant controls against alternative interpretations of dishabituation that implicated simple fluctuations of attention. Still another control group looked at the same light for 24 consecutive trials; these infants showed no selective disinhibition. The results of this study essentially represent swcessive hue discriminations, and they followed a course similar to those discussed previously (Bornstein, 1976a). As an example, the results of two groups are shown in Fig. 5. Those children who originally saw a blue of 480 nm and were tested with that blue, another of 450 nm, and a green of 5 10 nm paid most attention to the green. They treated the blue which they had not seen before (450 nm) as they did the blue which they had (480 nm). A complementary group of children who saw a green of 5 10 nm first paid the most attention to a blue of 480 nm during a test series that included the original green habituation stimulus ( 5 10 nm), another green (540 nm), and the blue of 480 nml. Together these two groups show that infants see 450 nm as similar to480 nm, 510 nm as similar to 540 nm, but 480 nm as distinct from 510 nm. Adults make similar distinctions and call them “blue” versus “green.” Three boundaries, at 490, 565, and 610 nm, were tested in this way with pairs of complementary infant groups: For one group, the habituation stimulus was selected from the category on the long-wavelength side of the boundary, and for the other group the habituation stimulus was selected from the category on the short-wavelength side of the boundary. In addition, related groups of infants tested the extents of individual hues between 430 and 660 nm.

I47

Chromtaic Vision in Infancy

450nm-

-

480nm

Group I 1

I

1.3 4.6

,

I

1

I

-

\

-

540nm

-

Group 2

I l l

I

1-3 4 6

7-9 10.12 13-15 T

510nm

I

I

I

I

I

I

7.9 1012 13-15 r

TRIALS

Fig. 5 . Same as Fig. 4 , but for two groups of infants who habituated and were tested with the wavelengths indicated.

Brightness was eliminated as a factor in these investigations . 5 Summary analyses of the infants’ visual attention to the different colors that were presented during testing are shown in Fig. 6 and compared with adult color-naming data (Boynton & Gordon, 1965). The results indicate that, on the whole, infants categorize wavelengths by their perceptual similarity very much the way adults do. That is, infants perceive four qualitatively distinct spectral categories or hues: blue (approximately 4 3 0 4 8 0 nm), green (approximately 5 10-560 nm), yellow (approximately 570-600 nm), and red (approximately 620-680 nm). As with adults, interhue boundaries for infants appeared around 490, 565, and 610 nm. Note that violet is not distinguished from blue. Nagel’s (1906) son Gerhard called purple, violet, and blue all “lilla,” and Ladd-Franklin (1901) also observed that children see colors in the short wave range as perceptually similar. Having noticed purples for the first time, Ladd-Franklin’s own daughter exclaimed: “B’u!-Wed!-Wed!-B’u!” “It was impossible to doubt that she, at least, saw in purple the blue and red of which it’s composed” (p. 398). ’Adult standard spectral sensitivity coefficients, in specific Judd’s ( 1951b) modification of Standard Observer sensitivity, were used to balance wavelengths (luminance = 3.4 cd/m2). Most groups in this study saw wavelengths longer than 550 nm, where relative infant and adult sensitivities closely match (Dobson, 1976; Teller & Peeples, 1974). For the groups who saw wavelengths shorter than 550 nm, Judd’s coefficients, which revised the too low weighting of the Standard Observer at shorter visible wavelengths, more closely approximate those shown for the infant. By design, independent groups experienced sets of wavelengths which spanned relatively short ranges, over which brightness differences are reduced, and the technique used in these s t u d i e d e t e c t i o n of hue boundaries-is one which minimizes the potential influence of brightness differences. Boynton and Gordon (1965). for example, have shown that although the absolute luminance level affects the apparent hue of some wavelengths (Bezold-Briicke shift), the wavelength boundaries between hues remain approximately constant over wide ranges of retinal illuminance or luminance. As Boynton (1971) has observed, color naming among color-normal observers shows much less dependence on luminance than on the wavelength of the stimulus.

Marc H . Bornstein

I48 I

9-

rn

5

-+

:

4 -

Z

5 6-

-

4 a

1;1

I00

20

50

1

1

1

I

I

I

I

I

I

I

I

l

l

I

+ -

-

H b -

ltl H

C

+ t l

13 -

INFANT SUMMARY

1

1

2 3 -

0

4

-

,~ H

+---I

H

-

H

-

BLU’E

-

I-

\

z W u

,“

I

I - 4

1 -

:

2

I

0-

I

e I

I

,

I

,

I

I

I

I

,

I

1

1

,

I

I

In a follow-up to this experiment, rate of habituation was used to index the organization of hue in comparable groups of young infants (Bornstein et al., 1976b). It is known that habituation to re-presentation of the same stimulus is faster than it is to varied stimulation (Cornell, 1974; Fantz, 1964; Friedman, 1972; McCall & Kagan, 1967). In this study three groups of 10 4-month-old infants were seen under stimulus conditions used previously: One group was shown the same wavelength (480 nm) repeatedly; a second group was shown, in quasi-random alternation, two wavelengths (480 and 450 nm) which had been selected from the same infant hue category as determined in Bornstein et al. (1976b); and a third group was shown serially two wavelengths (480 and 5 10 nm) which had been selected from different infant hue categories. The one stimulus which was common to all three groups, 480 nm, appeared to each group in the same sequential positions. Consequently, experimental treatments for all three groups differed only by the context in which 480 nm appeared. Again, as in the

Chromatic Vision in Infancy

149

original category study, wavelengths of 450 and 510 nm were selected so that physical differences between each and 480 nm were equal but the psychological distances varied; this choice controls for potential response variation ascribable to simple physical (wavelength) change rather than psychological (hue) change. The results of this study, plotted in Fig. 7, showed that (1) visual attention to spectral light waned with stimulus repetition in each group, ( 2 ) the courses of habituation for Group 1 (same-wavelength) and Group 2 (same-hue) were indistinguishable, but (3) the two-hue group (Group 3) exhibited more overall looking than Groups 1 and 2 and a slower rate of habituation. Figure 7 shows the course of habituation only on the 480-nm trials for the three groups. In summary, the intergroup comparisons indicate that rates of habituation to physically identical and to psychologically similar stimuli are the same, but that rate of habituation is reduced when psychologically dissimilar stimuli are intermixed. Since order effects and group heterogeneity were not factors, these data also can be taken to indicate corroboratively that infants perceive and respond to 480 nm as though it were the same as or highly similar to 450 nm but different or highly dissimilar from 5 10 nm. These results therefore provide independent converging evidence for the existence of an organization of hue in infancy. That preverbal infants organize the visible spectrum into perceptual categories like adult hues may be somewhat surprising. Commonly, psychologists (Beare, 1963; C. H. Graham, 1965), psycholinguists (Lenneberg, 1967), and an-

c

.~

o--o

GROUP I

4

10

W

E

I-

W

zY 0

sz

5

4

W

I

-

0 2 5 6 7 11 12 14 17 18 19 23 24 TRIAL NUMBER

Fig. 7. The course of habituation of visual attention in three groups over the 12 trials (of 24) on which the common wavelength, 480 nm, was presented. (From Bornstein et al., 19766. Copyright American Psychological Association. Used by permission.)

150

Marc H . Bornstein

thropologists (Carroll, 1964; Ray, 1952, 1953; Whorf, 1952) have maintained that categories of hue reflect arbitrary discriminations, and they have previously associated regularities in hue categorization (and even wavelength discrimination, see Kopp & Lane, 1968) with common cultural training. The fact that preverbal infants organize wavelengths into perceptual categories is, however, not unexpected from the points of view of biology or ethology. The neural processes through which visual information about color is analyzed begin at the retina and follow the main geniculostriate pathways in the brain. Different lines of evidence suggest that specific cells in the lateral geniculate nucleus decode the hue information in the color signal. They are selectively sensitive to the four principal hues of the spectrum, and De Valois and De Valois (1975) have suggested that these cells account for wavelength discrimination in man and other primates. The relationships between wavelength discrimination and hue categorization, and between, underlying lateral geniculate cell sensitivities and wavelength discrimination (De Valois & De Valois, 1975), indicate that to a large degree hue categorization reflects indigenous visual function. Significantly, hue categorization (related to wavelength discrimination) has been found in three norhuman species that possess color vision. Von Frisch (1964) demonstrated the organization of hues in the European honey bee (Apis mellifera), A. A. Wright (1972; A. A. Wright & Cumming, 1971) in the pigeon (Columba liviu), and Grether (1939) and Essock (1977) in the chimpanzee (Pan). This ethological perspective makes the assumption of a biological basis for qualitative chromatic distinctions more plausible than alternative assumptions about learning and language. In summary, color-naming or other hue-categorization studies of centrally and photopically presented monochromatic lights yield normative and verifiable data about the ways in which humans and infrahuman species perceive and organize color in their world. As it is for other organisms that possess color vision, the color world is organized into basic hues in infancy, that is before language acquisition. These basic hues number four: blue, green, yellow, and red. Hue naming or categorization has come to play an important role in the assessment of color vision and related perceptual phenomena in recent years. Using color naming, C. H. Graham and Hsia (1958) and Scheibner and Boynton (1968), for example, discovered the important fact that neither protans nor deutans typically perceive boundaries between basic hues at 565 and 610 nm. The data on hue organization in infancy argue that insofar as infants distinguish four hues across the spectrum, a color-normal or trichromatic “adult” view prevails quite early in life. It is a curious fact, in this light, that accurate color naming is so late in developing (see Section I, B, above). Goodenough and Anderson (1931), for example, reported that according to Merrill-Palmer test norms, only 4- to 5-year-olds will name a wide range of colors with some accuracy. Norsworthy and Whitley (1920) gave 5 years as the age when children know all four colors in

Chromtrtic Vision in Itfuncj

151

the Binet test, although Binet and Simon themselves reported a norm older than 7 years on the average (G. A. Miller, 1974). Naming seems not to stabilize until approximately the fourth year (Dale, 1969; Yendovitskaya, Zinchenko, & Ruzskaya, 1964/1971), and it remains problematic even late into development: witness psychophysical data from Boynton and Gordon (1965). One of three subjects in that study was M. B., Boynton’s 12-year-old son. Despite the fact that blue and yellow are mutually exclusive visual perceptions (Hering, 1878/1964), he named a few of the same wavelengths blue, yellow, or some combinations of blue and yellow. Late color naming is striking; recall that i n the absence of other information early investigators took the paucity of color vocabulary to indicate a lack of development of color vision in infants and young children (e.g., Shinn, 1909; Stern, 1924). Bornstein (1977d) evaluated four hypotheses to account for this discrepancy. First, younger children may simply have a problem with or have a lack of experience in associating color terminology with color percepts. Second, the slow acquisition of color vocabulary may reflect a bipartite process, conceptual on the one hand and linguistic on the other. Conjoining these processes requires time. Third, semantic development may proceed in such a way that names are attached to certain classes of words, such as entities, before others, such as attributes. Last, the color categorizationxolor naming schism could reflect physiological immaturity leading to the absence of percept or organization of color independent of color discrimination. The existence of different cell types functionally segregated by a cortical region (De Valois & De Valois, 1975) raises the possibility that i n ontogeny color differentiations could occur in the absence of the accurate utilization of other color information. The infant’s perception of color may serve several functions toward his eventual understanding of an otherwise visually complex world. The fact that early in life colors are organized and segregated by hue must increase their cognitive utility and value. At a sensory level, colors provide for more varied stimulation about the environment. At a perceptual level, color, in addition to luminance, subserves visual contrast and promotes spatial detection (Boynton & Greenspon, 1972; McFarland & M u m , 1975; Walls, 1942). Cognitively, colors provide a unique type of information about the environment (Conklin, 1973), and color similarities provide a basis for structuring the perceptual world. Aesthetically, colors please and differentially influence attention. Thus, hue perception may facilitate different mental operations and aid memory. It is compelling, for example, that the preverbal child must utilize purely sensory information, like color, in coding and remembering objects. The infant’s capacity to remember colors is the subject of the next section.

5 . Color and Memory Development Historically, memory has been thought to reflect verbal reprocessing of physical stimulus information. Memory researchers since Ebbinghaus ( 1885/1964) have used linguistic or pseudolinguistic stimuli, and even when nonlinguistic stimuli

I52

Murc H . Bornstein

have been used, both adults (e.g., Glanzer & Clark, 1962) and older children (e.g., Perlmutter & Myers, 1975) have often been found to use verbal codes to remember. More recently, ‘however, studies of adult memory have evidenced retention of nonverbal or “physical” information about a stimulus (e.g.. Melton & Martin, 1972). As Posner (1973) has observed, memory cannot be characterized as solely verbal since our ability to recognize is vastly superior to our ability to categorize and label. This observation is significant from a developmental viewpoint since memory processes for stimulus features are important, especially in infancy. They are integral to the growth of the child’s knowledge about the physical world. The early development of object recognition, for example, must depend wholly on the child’s encoding, processing, and retention of physical properties of the object. Nonverbal encoding must assume ontogenetic priority to verbal encoding. Like other dimensions of sensation, adult memory for color has been thought to reflect verbal skills (e.g., R. W. Brown & Lenneberg, 1954; Lenneberg, 1967; Siegel & Siegel, 1976). Recent research (e.g., Berlin & Kay, 1969; Bornstein, 1973, 1974; Rosch, 1973) has suggested, however, that color coding and color memory may in many instances operate independently of language labels; that is, there may be universal, sensory-based codes for color. Color therefore qualifies as one of the physical characteristics especially available to early memory encoding. That human infants about 4 months of age already categorize surface colors supports this view. Reasonably, then, infants’ perceptions of various surface qualities, such as hue, shoulld not only function in early perceptual differentiation but should also be subject to useful recollection. Theories of habituation (Cohen & Gelber, 1975) usually imply the constmction of some internal representation of a stimulus that can be compared with a test stimulus later. In this sense.,habituation tests tap the infant’s recognition memory as well as his discriminative abilities. If the temporal interval between the familiarization and test phases in the attention-habituation paradigm is expanded, longer term recognition memory is emphasized. Several investigators have studied infants’ recognition of color subsequent to habituation. In most experiments, however, color variations have been confounded with variations in form. For example, Martin (1975) showed that older infants can retain a color-form pattern up to 24 hours, and De Loache (1976) showed that a change in two or four of four color-form elements used in habituation will reexcite infants’ discriminative fixation. Because color and form were not varied orthogonally in these studies, they provide little information about color retention per se. In other investigations color and form were vaned orthogonally. Between 3 and 4 months, 4.5 minutes of familiarization are sufficient to elicit preferential attention to a change in both color and form, but insufficient for a change in color alone or form alone (Saayman, Ames, & Moffett, 1964; Welch, 1974). Cohen (1973) has suggested that after approximately 4 months,

Chrotiiiiric

Visiorr rn Iiifiincy

153

infants store information about color and form as separate components independent of their compound, since when tested with a novel form, novel color, or novel color-form configuration, they exhibit dishabituation to changes Lither in the color or in the form and exhibit the most dishabituation to a pattern novel in both color and form (Cohen, Gelber, & Lazar, 197 I ) . The critical age suggested by Cohen can be questioned, however, because Milewski and Siqueland (1975), using an operant measure, habituation of high-amplitude sucking, showed that I-month-olds can discriminate red from a blue with which they have been familiarized. Other studies have demonstrated color recognitior! at 6, 8 , and 12 months (e.g., Miranda & Fantz, 1974; Schaffer & Parry, 1969, 1970). Schaffer and Parry (1969) observed a sharp rise in visual attention among 6- and 12-montholds at the introduction of a test object identical in size and shape to a familiarization object but different from it i n color. These objects, whose presentation was counterbalanced across subjects, were predominately red or predominately blue. In a follow-up study, similar results were reported with yellow and blue objects equated for size, shape, and contour (Schaffer & Parry, 1970). In none of the above studies was there any attempt to control stimulus qualities such as bandwidth or brightness. I n fact, complex, unspecified colors continue to be used in developmental studies. Yet without such stimulus control it is possible, though not likely, that young infants excerpt brightness or some other dimension of the stimulus from the “color” and base their memory for the stimulus on it. The use of brightness-equated hues in the category studies with infants (e.g., Bornstein et a l . , 1976b) suggests that babies can use hue alone in immediate memory. Immediate-recognition data in Bornstein et al. (1976b) and Bornstein (1976a) evidence an important aspect of infant color memory. In the serial habituationtest paradigm, exact stimulus (wavelength) information is apparently lost, while gross category (hue) information is retained. Theories of habituation (e.g., Sokolov, 1958/1963) suggest that habituation reflects the construction of an internal representation or neuronal model of stimulus properties based on repeated experience with those properties. The existence along the visual pathway of neural tissue specifically sensitive to different regions of the spectrum suggests that Sokolov’s theory should be modified. Not all cell assemblies or neuronal models need be constructed experientially. Rather, the neurological equipment for some properties, like hue, may exist i n the absence of specific experience. Moreover, the child’s categorical response to color, that is his failure to recognize one wavelength following habituation to another wavelength of the same hue, fleshes out the nature of these cell assemblies for hue. Hue-sensitive cell types may respond to a variety of wavelengths that fall within a given hue class. Presentation of a wavelength from any one hue class may mean excitation of the hue assembly in toto, so that only a change to a wavelength in another hue class

I54

Marc H.Bornstein

will excite a different cell assembly and activate attention. The corollary of this “category effect” is that, for color at least, neural coding in habituation is not physically exact. If exact neural coding subserved habituation, then infants should dishabituate to a new wavelength which represents a familiarized hue. However, the data from immediate- and delayed-recognition studies indicate that infants process chromatic information categorically using qualitative “hue codes” rather than exact “wavelength codes.” Data from adult studies of color memory suggest that color is recoded verbally for long-term retention (Bornstein, 1976c; Hochberg, 197 l ) , and analogous evidence suggests that color memory in infancy may be lost in several minutes. After habituating two groups of 4-month-olds to a color-form geometric configuration, for example a green circle, Pancratz and Cohen (1970) tested them with the original pattern and also novel color-form configurations, for example a blue triangle. The group tested immediately after habituation remembered, the group tested 5 minutes after habituation did not. Infants in Bornstein’s (1976a) study gave a pattern of recognition responses which suggested that by 3 months of age immediate visual memory lor pure hue alone has begun to develop; infants recognized 570 nm immediately following habituation to it. A follow-up study that employed the same stimuli and procedures (Bornstein, 1976b) showed that 4-month-old infants’ recognition of the habituation hce resisted 3 minutes of retroactive interference from the same visual dimension (different hue) and from different visual-auditory dimensions (mother’s face and voice), but was not so resistant to 5 minutes of interference. Infants therefore show good immediate recognition of hue, and this result further substantiates reports of nonverbal coding of (chromatic) information in memory. Bryant (1974) has argued that the coding in juvenile or infantile memory is relative in nature and that successive discriminations depend upon a constant frame of reference rather than an absolute internal code of a stimulus. Results from infantile color memory suggest that young children can and do code wavelengths absolutely by apparent hue; that is, infants successfully discriminate hues in the absence of alternatives other than inferred internal categories of information (Bornstein, 1 9 7 5 ~ ) . In 1924, Stern questimed whether young children actually use color codes in memory at all, bct at the time children were generally thought to be “color blind.” Certainly, young children classify objects by color (Brian & Goodenough, 1929; Suchman & Trabasso, 1966), aild the studies outlined above show that absolute c d o r codes for this purpose are available to them from infancy. Not unexpectedly, therefore, researchers who have carefully manipulated chromatic discriminative cues have demonstrated that color can facilitate memory. Daehler, Bukatko, Benson, and Myers (1976), for example, found that the addition of color cues f,acilitated 18- to 36-month-olds’ performance on a delayed hidden-object task, and having established the perceptual salience of

color over number and position among 5 % - , 8%-, and 11%-year-old children (Odom & Guzman, 1972), Odom (1972) found that this salience was directly related to recall accuracy in problem solving. Normally, the number of colors recalled increases linearly with age, between 8 years and adulthood, regardless of reterdion interval or the nature of interference (Belmont, 1972). Belmont (1972), however, found retardates (MA = 56-83) deficient in short-term (0-12 seconds) memory for color relative to normal children (MA = 96-104), and he attributed this difference to increases in information-processing capacity with age and mental ability. Sinson and Wetherick (1972, 1973, 1975) also observed a specific deficit in short-term retention of color information in Down’s syndrome children relative to a group of MA- and CA-matched controls: Down’s syndrome children consistently averaged nine times the number of successive color matching errors as normal children, though these same groups performed comparably in successive form matching and in simultaneous color matching. Sinson and Wetherick ascribed this color-memory deficiency to unidentified genetic origins; intriguingly, Miranda and Fantz (1974), who tested discrimination and recognition memory for color and form in Down’s syqdrome and normal infants, found a similar deficit pattern: In their habituation study, Down’s syndrome infants failed to recognize a change in color at 3 0 4 0 weeks of age, when normal infants did. The extent to which Miranda and Fantz’s (1974) results are ascribable to abnormal memory or to abnormal development of color vision is unknown. From other results in their study it is clear that the development of recognition memory in Down’s syndrome infants lags behind that i n normal infants generally; also the distribution of color deficiency i n retarded populations does not differ from that i n the normal population (Courtney & Heath, 1971). The stimuli used in the study were, however, not equated for brightness or saturation. In addition, the two colors used were red and blue, two hues which infants prefer nearly equally, and the influence of color preference in developmental studies of color vision may be significant. In summary, infants remember colors i n “hue codes” rather than as exact physical information. Early in life, memory for color is comparatively tenuous. The fact that infants do recognize colors, though, strongly suggests that color memory is not language dependent or necessarily relational and that infants may use the chromatic information about a stimulus toward its future identification. 6. Color Preference Experimental observations of the ontogeny of color vision began with studies of color preference (Baldwin, 1895), and interest in the development of chromatic vision has been sustained by continued investigation in this area (e.g., Bornstein, 1975d; Dashiell, 1917; Gesell, Ilg, & Bullis, 1949; Myers, 1908; Spears, 1964; Staples, 1932). In consonance with a belief that children were color deficient,

I56

Mort H . Bornstein

early researchers concluded that no color preference was discernible before the fourth year of life (e.g., Dashiell, 1917; Major, 1915). Later studies with babies, however, confirmed preferences for colored objects. Staples ( 1932), for example, studied preferences for the Atlas series of Munsell colors in different groups of children. The youngest (imean age, 3 months) looked more at primary colors (red, yellow, green, and blue) than at a neutral gray. Older infants (mean age, 13.3 months) generally reached or pointed more at red than at other colors. Spears (1964) paired Munsell colors in front of 4-month-olds and found blue reliably preferred to (regard’edlonger than) gray, but blue, red, and yellow were equally preferred. Indeed, reliable color preference of a general sort has received clear experimental support in adults and infants (Bornstein, 197%; Guilford, 1940) and even in neonates (Jones-Molfese, 1977). Preference methodologies can yield data with multiple interpretations; frequently they have been applied to issues i n attention and discrimination, infrequently in aesthetics. Often and unfortunately, investigators interested in color preference have not been schooled in psychophysics or color science, and investigators with psychophysical credentials have not been interested in the study of preference. Therefore, historically, stimulus control has been inadequate in preference studies, and this fact, perhaps more than any other, has contributed to variable results. I n ontogenetic studies of color preference, as in other color psychophysics, the control of brightness has been a perennial problem, since preferences for brightness vary in infancy (cf. Hershenson, 1964; Thomas, 1973). One experimental course has been to assume infant-adult parity in matching colors for brightness. Notwithstanding the questions associated with this assumption (see Section 11, A , 2 ) , hue is a more influential factor than brightness in adult judgments of color preference (Guilford, 19341, and therefore the results of these studies merit consideration. Two principal findings have emerged from color-preference tests with infants and adults: ( 1 ) primary or focal colors, that is category centers, tend to be preferred to mixed or boundary hues, and (2) the spectrally extreme hues tend to be preferred to midspectral hues. One study (Bornstein, 1975d) exemplifies both these trends. Bornstein ( l975d) monitored 4-month-olds’ looking at eight equally bright monochromatic spectral lights: violet (430 nm), blue (460 nm), blue-green (490 nm), green (520 nm), green-yellow (560 nm), yellow (580 nm), yellow-red (600 nm), and red (630 nm). These stimuli subtended lo” of visual angle and were presented at a luminance of 55 cd/m2. The eight colors span the visible spectrum and represent “centers” of and “boundaries” between infants’ color categories. Infants in paired-comparisons groups saw all possible pairs of the eight wavelengths twice; infants in a comparable single-stimulus group saw two random presentations of each of the eight. The babies in these two conditions

157

-37.

-z,

-

w

r

23

-11)

z

0 k-

- 1 4

a

X - 9

LL

Fig. 8 . Mertn c/urritions ofit$itnts’ totiilfiwtiom and meat1 adult pleasantness rutings as a function ofwavelength. Letters B (blue), G (green). Y (wllo w), iinci R ( r e d ) . indicate the upparent hue of the wavelength: see te..rtfbrfirrther e.xpl(ttuttron /‘he two orilinute scales have been udjjusted by a double linerrr-regression terhniytie (Bornsrein. I075d. p p . 41 I 4 1 2 ) so as to he equrvulent. (From Bornstein, 1975d. Copyright Acudemic Pr.e.s.3, I n c . U . w I by permission.)

provided separate replications of the same basic relationship between fixation time and wavelength: rank order correlation ( p ) = +.92. The pairedcomparisons data are displayed in Fig. 8. Although not all color-category centers were looked at more than all interhue boundaries, the attention infants in both conditions paid to category centers as a group reliably exceeded that paid to boundaries as a group.‘ Other investigators, including Jastrow ( 1897), Guilford (1934), and Eysenck (1941), have previously found center hues preferred to boundary or mixed hues. Second, the Bornstein study showed that blue and red were most preferred. Although space limitations prevent a complete review of the plethora of studies of color preference, it is worthwhile to look at a recurrent trend among some of the better studies that shows consistent observer preferences for the spectral extremes red and blue, relative to midspectral green and yellow. Baldwin’s infant daughter, H . , showed an equal number of “dynamogenic” responses (see Section I , B) to red and blue and fewer to white, green, and brown (Baldwin, 1893), ‘Has this infant perference an influeiice beyond immediate attention, say to habituation and memory’? C. J. Brown (1974) found that X-week-olds preferred an 8 X 8 checkerboard to a more “complex” 24 X 24 pattern; her infants also showed less habituation to the 8 X 8 than the 24 X 24. Habituation was a linear function of prefercnce for complexity. Intriguingly, habituation data from the hue-category study (Bornstein et i i l . , 1976b) show the opposite result: Slopes of habituation for more preferred, purer blue (450 nm) and red (630-660 nm) stimuli averaged - 1.84 and - I .63, respectively, while habituation slopes for wavelengths near less preferred, mixed hue boundaries averaged - 1 .18.

I58

Marc

H. Bortlstern

and Holden and Bosse (1900) found blue and red preferred to yellow and green. The same trend emerged in infant studies by Staples (1932), Stirnimann (1944), and Spears (1964). A preference for blue over yellow also appeared in Schaffer and Parry’s (1970) study of infant discrimination. In tests with older children, Dashiell (1917) found that kindergartners ranked blue and red as more pleasant than yellow and green, as did the college sophomores he asked. Garth and Porter (1934) found that 1032 children 1-7 years of age generally preferred red and blue Milton Bradley colors to green, yellow, or white. Likewise, Bradbury and Nelson (1974) found the 7-, 9-, and 1 I-year-olds preferred red to blue to green. Adults display similar consistent patterns of color preference. Bornstein (1975d) asked 24 adults (mean age, 24 years) to judge the pleasantness of spectral lights his infants had seen. In general, adults’ ratings of the pleasantness of the colors paralleled infants’ differential attention to them ( p = .80; see Fig. 8). Among the most careful and extensive adult studies of color preference are those of Guilford, who used Milton Bradley colored papers (Guilford, 1934, 1940; Walton, Guilford, & Guilford, 1933). The rank-order correlation coefficient between the adult data in Bornstein’s (1975d) study and the averages for men and women (n = 10) tested by Guilford (1934, Table 11) was calculated at .66, which is statistically significant. The data for Bornstein’s adults also replicated Waltonetal. ’s (1933) extensive (n = 1279) paired-comparisons data (p = + .74). These studies confirm that adults prefer centers to boundaries and that they prefer the spectral extremes. Jastrow (1897) and Eysenck (1941) also found that the red-blue-green preference hierarchy was a general result in adults. Although color preference has often been viewed as susceptible to fluctuating influences, the resilience clf these preference findings to diachronic change, development, laboratory method, and investigator bespeaks their robustness. Often with such data in hand, the psychologist will begin to think in terms of biological influences. Certainly the infant data support this view. Two additional sources of information are (crucial to tendering of such an argument. One is a cross-cultural confirmation. For the red-blue color preference Garth (193 1) and Eysenck (1941) have provided just the appropriate data. Garth conducted tests among 1000 Caucasian, Negro, Filipino, American Indian, Japanese, and Mexican children. He found similar color preferences among the young of all six populations: Red and blue were favored over other colors. Eysenck’s (1941) review provides overwhelming confirmation of the effect. The second source is a plausible biological basis for these two preference phenomena. Perhaps preferences among infants and adults are stable because infants process color like adults or because color perception at all ages is founded on a similar neurological basis. The degree of infant attention may relate quite directly, then, to the activity level of color-sensitive neural tissue. Each of the eight stimuli used by Bornstein (1975d) was a hue-category center or boundary, and as suggested infants looked reliably longer at category centers

+

+

Chrornritic. Visiorr in Infancy

IS9

than at boundaries just as adults rated centers more pleasant than boundaries. The lateral geniculate cells that analyze chromatic information and subserve wavelength discrimination have spectral sensitivities that are relatively narrow, and they fire at maximum frequency in four select regions of the visible spectrum: the blue, green, yellow, and red. De Valois, Abramov, and Jacobs (1966) have argued that as “hue is signalled by differences in response magnitudes among the various cell types, then those spectral regions which appear red should be those producing maximal responses in [red-sensitive] cells, and so on” (p. 976). Additionally, correlational analysis (Bornstein, 1973) has shown a remarkable correspondence between ( 1 ) the wavelength regions of maximal response for the four spectrally opponent geniculate cell types and (2) the wavelength regions which mark category centers or universal “foci” of “blue,” “green,” “yellow,” and “red” color designations from cross-cultural studies (Berlin & Kay, 1969). The idea that attractiveness or attention relates to the level of excitation or firing rate of stimulus-sensitive neurons is not new. Karmel (Karmel & Maisel, 1975) has formally proposed that infants’ attention to patterned stimulation (checkerboards) varies as a function of the density of contour in the stimulus. Moreover, Karmel, Hoffman, and Fegy (1974) recorded VEPs from surface electrodes on the scalps of infants who looked at differently patterned checkerboards. These investigators found that contour densities that elicited the longest epochs of infants’ looking were the same as those that evoked the greatest amplitude of gross electrical excitation in the cortex. In their analysis infant fixation time and the amplitude of the cortical electrical potential represent interchangeable ordinates against the independent variable, contour density. Perhaps, then, both contours and colors (here hue centers) that produce more central stimulation also tend to hold attention longer. Among the different colors, infants and adults display distinct and reliable preference for the spectral extremes, blue and red. Does this preference also have a basis i n neurological function? Perhaps so, as it is related to saturation perception. Even though hue is the principal dimension of color preference (Child, Hansen, & Hornbeck, 1968; Guilford, 1934), it is ceferisparibus the saturation of a color that is striking. Saturation detines the degree of chromaticity of a color, or the percentage of monochromatic light in a mixture with white. The spectrum appears unevenly colored to the eye, and the spectral extremes (red and blue) naturally appear more saturateddifferent from white-than does yellow (C. H. Graham, 1965). Shrnimann (1944) suggested that infants stared more at more saturated colors, and Bornstein (1975d) found that the pattern of infant attention to spectral lights paralleled the way adults discriminate the saturation of different wavelengths (C. H. Graham, 1965). Vernon (1962) believed that children are attracted to more saturated colors; indeed, to Beebe-Center (1932, p. 306) early studies of infant

160

Marc H . Bornstein

and juvenile color preference seemed to imply that “in the first half-year, preference depends only upon saturation and brilliance.” Recently, Bornstein (1978) investigated infant discrimination and preference for various levels of saturation, Chroma in Munsell terms. Bornstein showed 20-month-olds two Munsell Chroma levels (/2 and /lo) of each of three equally bright hues, blue, green, and red. Each baby saw each color once in a random order and once in reverse order; no babies saw the same order. The results showed that infants distinctly and reliably prefer the higher saturation level of each hue. The few preference investigations with older children in which saturation was varied in a controlled way, such as Child et al. (1968), have consistently revealed preferences for high levels of saturation, and these results agree with major adult color-preference studies, such as Eysenck (1941), Granger (1955), Guilford and Smith (1959), and Helson and Lansford (1967), in that i n each “preference is a positive function of chroma” (Child et al. 1968, p. 244). Why should the saturation of a color so influence attention? How is saturation preference related to preference for the spectral extremes? Perhaps reasons exist similar to those explaining preference for exemplars of color categories. The same model that Karniel (Karmel & Maisel, 1975) relied on to show that infant looking time and the: amplitude of the VEP in infants systematically depend upon contour density may be invoked to explain the infants’ attention to saturation. Since infants attend longer to more saturated colors, one might ask, Do variations in saturation produce systematic changes in the amount of electrical activity in the visual system? The opponent cells that De Valois and his associates (De Valois & De Valois, 1975; De Valois et al., 1966) have identified in the geniculate bodies of the macaque have been shown by De Valois and Marrocco (1973) to increase monotonically in frequency of cell output with the chromaticity of wavelength input. A physiological theory of attention would predict, therefore, that as infants are exposed to a variety of saturation levels, they will attend increasingly to purer (more saturated) colors. This is what Bornstein (1978) found. This outcome generalizes the neural model of infant attention to color and further elucidates infants’ selective preference for the spectral extremes, since red and blue are naturally more saturated hues than are midspectral green and yellow. Insofar as information in the brain is coded as the frequency of nerve action potentials (Stubbs, 1975), the response characteristics of specialized receptors may provide a biological basis for attention. Such results imply that saturation acts as a salient stimulus dimension involved in the mediation of visual attention to color. Saturation is also a key dimension in memory coding of color. For example, Bartleson (1960) found strong evidence that adults increase the saturation of colors in memory, as Gilbert (1894) had found that children 6-17 years typically code colors into higher levels of saturation in memory. In summary, preferences are usually not thought to bespeak an exquisite amount of consistency, and it is widely believed that they must be subject to fad,

Chromcitic Visiori in

Inhincv

161

fancy, and (cultural) context. Consistent results of color-preference tests in infancy and the developmental stability of such results help to refute such beliefs; they show that simple or pure hues are preferred to mixed hues and that the spectral extremes-blue and red-are preferred to midspectral hues. Further they suggest that more saturated hues are preferred to less saturated hues. Each of these findings relating attention to color is predictable from a neurologically based model of central processing of chromatic information.

7 . Coloratura The range of subject areas related to the development of chromatic vision is not limited to those discussed above. Other diverse issues are of interest either for the advancement of our understanding of color vision and visual development, or because the unique characteristics of color offer theory testing possibilities. a. Chromatic Visual Acuity. The ability to resolve a visual target depends upon both its luminous intensity and color (Bornstein, 1975b; C. H. Graham, 1965); acuity is worse in dimmer illumination and, because of the chromatic aberration of the eye, worse in general under short-wavelength and longwavelength illumination. Pollack and his associates (Holmes, Kelton, & Pollack, 1973; Kelton & Pollack, 1974; Skoff & Pollack, 1969) have reconfirmed in children 7-14 years of age the effects of hue contrasts on visual acuity. Pollack also uncovered an age trend toward better acuity. This trend reflects in part ontogenetic changes that occur in optical pigmentation (see Section 11, A). Because intraocular pigments selectively absorb, and thereby reduce or eliminate out-of-focus short-wavelength visible light, acuity would tend to increase. Since Fantz, Ordy, and Udelf's (1962) studies of the early ontogeny of visual acuity, several other investigators have focused on the early practical and empirical importance of visual acuity (e.g., Marg et al., 1976; Salapatek, 1975). Yet no investigators have extended the study of visual acuity in infants to chromatic contrasts. Given that such contrasts are important early to visual resolution and to detection, such an extension would appear to be a logical development. b. Peripheral Color Vision. We are moved from one part of the visual environment to another by stimulation in the peripheral visual field. Studies of infantile looking patterns (e.g., Tronick & Clanton, 1971) clearly demonstrate that the visual system is adapted to environmental exploration and information extraction through continuous attention deployment to new loci. Still open to investigation are the processes and parameters that govern the young child's detection of and orientation to peripheral chromatic stimulation. How do central and peripheral chromatic stimuli compare in their attention-getting and featureextraction characteristics? L. G. Williams ( 1967), for example, found that adults select objects in the extrafoveal field much better on the basis of chromatic cues than on the basis of achromatic cues, such as size or shape. Chromatic vision depends initially upon the existence and operation of &he

162

Marc H . Bornstein

retinal cones. In the adult, the cones are distributed unequally across the retinal surface; they are concentrated centrally, particularly in the macula. Important questions related to physiological maturation concern the extent of the effective visual field in newborns and young infants, whether the child’s effective visual field grows, and whether that growth varies for different hues. Results from Tronick (1972), Harris and MacFarlane (1974), and A s h and Salapatek (1975) suggest that between 2 days and 2 weeks the child’s visual world is narrowly central, but over the coursc: of the first 2 months of life it begins to grow peripherally. None of these investigators used chromatic stimulation. Earlier, however, Luckey (1893) documented visual field growth in color perception. Although observer decision processes almost certainly influenced his data, Luckey reported that, like adults.,children (7-13 years of age) saw peripheral stimuli as exclusively blue or yellow. Trichromasy, the inclusion of red and green with yellow and blue discriminations, was reserved to the central visual field. Characteristics of the young child’s attention to chromatic information outside his momentary focus may thus provide additional information about physiological maturation, stimulus control, and environmental influence on perception. The study of peripheral color vision has secondary methodological implications, since both central-stimulus and paired-comparisons techniques call for some degree of spatial distribution of visual stimuli. Results of infant studies that are interpreted in terms of higher perceptual or cognitive functions might alternatively be explained by sensory limitations on peripheral vision. c. Color Constancy. Although the appearance of an isolated surface depends upon the manner in which it reflects light, in everyday perception surface colors are actually not influenced very much by variation in lighting. This subjective phenomenon is known as color constancy. In color constancy, brightness and chroma seem to become stable properties of the object. Two explanations of such constancies have been cognitive and structuralist. Helmholtz (1866/1962), for the cognitive approach, maintained that relative brightnesses are invariant because sensation is always followed by a judgment based on past experience in the environment. Hering (1878/1964), for the structuralist view, maintained that brightness and color constancies reflect retinal responsivity and adaptation. Since these alternative explanations disagree about whether experience is a prerequisite for constancy, a test of color constancy in infants would help to resolve a long-standing theoretical question in perception. Because color constancy represents a visual phenomenon with enormous adaptive significance, one might expect to find it present in infants near birth.

d . Perceptual Salience and Color. It is unknown how the three dimensions of chromatic experienc-hue, saturation, and brightness-zompete in perception for dominance. Chase (1937) observed that hue constrasts were more interesting than brightness contrasts, that is, they elicited longer looking in young

Chromatic Vision in Infancy

163

infants. However, Clifford and Calvin (1958) suggested that attention to brightness relative to chromatic aspects of a stimulus eased color discriminations and, similarly, Corah and Gross (1967) found that brightness was a most important variable in children’s color perception. H . C. Smith (1943) and Gaines (1972) have both observed that all three dimensions contribute to the chromatic experience and to the salience of color. It is certain that some dimensions of color covary (e.g., see Davidoff, 1974) and are perceptually integral (Garner, 1974). Others are perceptually separable. Garner has argued that integrality of dimensions affects not only discrimination, but also classification and memory. Whether the integrality-separability distinction is learned or innate is presently open to question and may find a solution in infant color discrimination experiments. It is also certain that solutions to many discrimination problems are influenced by attention to one dimension or to the compound of two or more dimensions (Zeaman & House, 1963). Which dimension(s) the infant chooses to attend in preference, discrimination, and memory obviously constitutes a central question in perceptual development. Perhaps the kind of componential habituation advocated by D. J. Miller (1972) will be useful in distinguishing among the visual dimensions to which infants attend and will help to establish their relative order of salience. One of the best examples of this issue derives from recurrent questions concerning the relative salience of color and form. Objects in the natural world are defined by patterns that consist of both physical extents and surfaces qualities, that is, forms and colors. Interest in the relative salience of color versus form derived from early philosophical views on the relative dominance of touch over vision in infancy and from the kinds of information thought to be available through these different modalities (e.g., Berkeley, 1709/1901; Bryant, 1974; Preyer, 1890; Stern, 1924). Several writers believed that the young child’s interest in color vis-a-vis contour, form, or shape more generally was weak (Shim, 1909; Stem, 1924; Valentine, 1942) and, in choice-preference tests, both Baldwin (1895) and Fantz (1961) found that young infants reached for or looked longer at the contour patterning of newspaper print than at a patch of color. However, with little consideration of children’s preferences for particular forms and colors, results such as Baldwin’s and Fantz’s are only suggestive (cf. Spears, 1964). In fact, consideration of the several dimensions of color-form matching, dominance, or salience indicates that color is at least as salient as form at the beginning of life. For example, a task analysis of the requirements associated with matching by color as opposed to matching by form points up the relative ease, rapidity, and facility in color matching. A single glance at any one point on a chromatic stimulus is sufficient to apperceive its color, but inspection of the entire perimeter of a shape may be necessary in accurate matching by form.

164

Marc H . Bornstein

Typically, then, among both children and adults (Farnham-Diggory & Gregg , 1975; Peeke & Stone, 1973), matching by form requires more time and many more eye movements than matching by color. In this sense, chromatic information is more readily available. Discrimination and memory studies conducted with infants as young as 1 month indicate further that they attend to the chromatic information in a colorform pattern (Cohen, 1973; Milewski & Siqueland, 1975). Cohen et al. (1971) showed that infants attend to a novel color or to a novel form, and, as also found by Saayman et al. (1964), novel colors and forms combine to increase attention. For infants, then, color and pattern both seem to represent dimensions relevant to visual information processing. Controlled discrimination studies reviewed above suggest, however, that hue (Bornstein, 1975a), in contradistinction to form (Bryant, 1974), can operate as an absolute code in infantile perception and memory. The preferred dimension of classification, color or form, changes ontogenetically (Katz, 1913; Werner, 1940), as shown by Brian and Goodenough (1929) who performed an extensive color-fonn classification study (n = 474) with subjects 2 years of age through adulthood. They plotted form and color choice as a function of age and found that although form was at first favored, color was preferred by children approximately 4 and 5 years of age, after which form was increasingly preferred. Forty years later, Suchman and Trabasso (1966) found that 4 years and 2 months wa:j the median age of the color-form transition. It has been suggested that exposure to educational institutions and, in particular, to the form-dominant requirements of reading may underlie this critical shift in dimensional attention. For example, Ghanaian children who do not attend school do not change from color to form responding (Davidoff, 1972). Likewise, better problems solvers (Suchman & Trabasso, 1966) and children with higher IQs (Brian & Goodenough, 1929) prefer form earlier. Goesling (1973) also found that retardates tend to sort by color, arid the level of school achievement for deaf children who sort by color lags approximately 3 years behind that for hearing children (Suchman, 1966a). Monkeys tend, too, to sort by color (Warren, 1954), as do adults under stimulus-degraded (i.e., tachistoscopic) conditions (Peeke & Stone, 1973). Independent of the support they afford one or another explanation of the color-form transition, the special-group data and the reading explanation favor the natural priority of color over form in the perceptual hierarchy of young children. Form dominance in 2-year-olds’ classification behavior (Brian & Goodenough, 1929) has been replicated (Melkman, Koriat, & Pardu, 1976); whether color or form predominates as the dimension of preference in infancy is, however, still open to experimental investigation. The relative salience of color or form may be important to establish for their respective roles in learning and early cognitive development. Preference focuses attention on one dimension, and increased attention serves to increase discrimination among instances along that dimension (Suchman, 1966b). Moreover, learn-

Chromaric Visioti in Infancy

I65

ing is hastened when a subject’s preferred attribute is relevant (Trabasso, Stave, & Eichberg, 1969). The implication of this line of research is, then, that among younger children color cues may be valuable in learning situations. Preschool children certainly take greater advantage of color cues than of form cues in concept attainment (Colby & Robertson, 1942; Lee, 1965) and in memory (Daehler et al., 1976). Like Clark (1973), Farnham-Diggory and Gregg (1975) have argued persuasively that perceptual dimensions, such as color, must operate prior to and act to structure conceptual or semantic categories in early congitive development. In support, they demonstrated that young children classify by color faster and more accurately than by other perceptual dimensions or more abstract properties of objects such as function or class equivalence. Color concepts are available even to babies. Bornstein ( 1977c) recently found that 16-week-olds habituated to a single hue would discriminate a novel hue but that comparable babies habituated to a variety of hues would generalize habituation to a novel hue. These results demonstrate “concept” formation and utilization in very young infants (Bourne, 1966). Later in development form appears to be a more salient dimension than color. Children older than age 5 and adults tend to utilize form properties to classify objects in everyday life (Brian & Goodenough, 1929; Farnham-Diggory & Gregg, 1975; Suchman & Trabasso, 1966). At this time, red and blue triangles are both classified as triangles since “of course it’s what the thing is that matters, not what color it happens to be” (Brian & Goodenough, 1929, p. 200). Even though they prefer to focus on forni, children as old as age 9 or 12 will, nonetheless, readily shift their attention to the color component of a pattern if its shape is nonsensical or task irrelevant (Hale & Green, 1976). Likewise, Underwood, Ham, and Ekstrand (1962) found that adults will use the meaning of a word printed in color as a principal memorial cue but will use the color of a nonsense syllable. Among older children and adults, then, color may not be principal, but it certainly represents an important addidional cue and one that is readily available for stimulus discrimination and learning.

e . Color as a Perceptual Cue. among children reading maps:

Nault observed the following propensities

children associate hue change (as from green to brown to blue) with change in quality, and they associate value change (light to dark) with change in quanriry, amount or intensity. For example, many children said that light blue areas indicated shallower water and dark blue areas indicated deeper water. But when a purplish or reddish-blue was used to depict the deepest water category, two-thirds of the children did not associate this with a further depth change, but rather guessed at all sorts of qualitative changes-islands, coral reefs, and so on. (cited in Arnheim, 1971, p. 311)

In general, various of the dimensions of color act as cues to other perceptual phenomena i n vision (Payne, 1964). For example, color influences judgments of weight, size, temperature, and distance. It is certain from research conducted by

166

Marc

H.

Bornstein

McKenzie and Day (1972) that the absolute distance of an object critically determines an infant’s visual attention to it. McKenzie and Day, however, used white cubes only. It is not yet known whether perceptual distance might vary with hue in infants as it does in adults (J. J. Gibson, 1966; McCain & Karr, 1970; Oyama & Yamamura, 1960), and whether the color-depth effect reflects structuralist function or, as Edwards (1955) maintained, experience with objects in the natural world. Hypothesis-testing questions such as these are potentially answerable by color studies with infants and young children. Two experiments provide contrasting opinions on this point. Morgan, Goodson, and Jones (1975) argued from cross-sectional data thai conventional color-temperature assocations, such as hodred, warm/yellow, codgreen, and coldhlue, are acquired relatively late in development and are therefore founded on a “loosely held cultural norm rather than an evolutionary or physiological basis.” Contrasting this domain, Lawler and Lawler (1965) argued on the basis of the consistency of preferences of preschoolers (mean age, 3.7 years) that color-mood associations, such as sad/ brown and happy/yellow, are “biological in nature.”

f . Color Vision Defect and Development. One in twelve (8.2%)Caucasian males is color deficient: This statistic translates into one boy in every classroom in the United States. (Only .4% of Caucasian females are color deficient.) Elsewhere, Bornstein (1977b) has reviewed the effects of color vision deficiency on scholastic, psychosocial, and career development. For example, most commentators, like Gardiner (1973), Heath (1963, 1974), Shearron (1965), Sloan (1963), Snyder (1973), Thuline (1972), and H. Williams (1975), have seen the colordeficient child as at a disadvantage in the classroom. Preschool and primarygrade classrooms are usually quite colorful; texts are typically illustrated with colors; and the teachers often emphasize color. Many preschool and school materials are color coded and thus integrally tied to the child’s color discrimination capacities (Mandola, 1949; Thuline, 1972). Knowing which child is color deficient is obviously beneficial in the school context. Assessment of color deficijency as early as infancy, however, may be beneficial for several other reasons. For example, color vision deficiencies have been used by geneticists and clinicians for purely medical diagnostic purposes, such as marking associated hereditary difficulties (McKusick, 1962; Mendlewicz, Fleiss, & Fieve, 1972). Additionally, diagnosis in infancy is related to theoretical questions in color vision and to the ontogeny and early function of chromatic vision. Both behavioral or electrophysiological assessment of the status of color vision in infancy may be possible. Determination of a neutral region (see Section II,B,2) has, historically, been regarded as a satisfactory diagnostic sign of the status of color vision (Hsia & Graham, 1965; Walls & Heath, 1954) and has been extensively used in animal sensory psychophysics (e.g., De Valois, 1973; Grether, 1939). Because of the desaturated nature of the anomalous trichromat’s spec-

Chrorncrtic Vision in Injuncy

167

trum, examination for the neutral zone is, as well, sensitive to color-weak (anomalous) deficiencies (Chapanis, 1944; Hurvich, 1972; W. D. Wright, 1947). Using the neutral zone as an index, Peeples (1976) addressed the question of whether hereditary color vision deficiencies are congenital by testing five male infants whose maternal grandfathers were reputed dichromats. Three of these 2-month-olds failed to discriminate 496 nm from white. Peeples therefore concluded that color deficiency is present at or near birth and represents a life-long characteristic. Similarly, Bornstein ( 1976a) employed the neutrai-point test to assess the status of color vision in 3-month-olds. This habituation procedure could also provide an accurate and reliable behavioral measure of the status of color vision in early infancy. Although neither method discriminates the type or degree of color deficiency, both adeqtiately discriminate normal from abnormal visual functions and could be relatively easy to operationalize, administer, and interpret. Other “objective” tests of deficiency now available are based on more complicated electrophysiological measures (e.g., Regan & Spekriejse, 1974), and they may be adaptable to early infancy (Barnet et ul., 1965; Lodge et ul., 1969), but only under special circumstances. In children, as in adults, color represents an immediate perceptual dimension, a system of classification, and an often utilized symbolic code. How many of these advantages the color-deficient child lacks during his cognitive development may be questioned. The data suggest that neither IQ nor achievement in school is hampered in color-deficient children (Lorenz & McClure, 1935; Mandola, 1969; Shearron, 1965). As those reared on black-and-white television know, much of the visual information in life is adequately conveyed independent of color. In reality, however, the world of the color deficient is a world apart (Goodenough, 1945); the dichromatic child is indeed part of a minority culture. As he grows he must develop a set of alternative skills: He may utilize brightness, position, or other visual cues to compensate for his color deficiency. The course of this adaptation is essentially unstudied.

111. Summary General interest in the nature of man’s first perceptions is as old as philosophical inquiry, and during the last half century experimental interest in early visual processes has grown steadily. In this time, investigations of pattern vision in infants have dominated the study of the ontogeny of perception, and researchers in developmental psychology have made enormous strides in understanding this facet of early behavior (Cohen & Salapatek, 1975). The history of interest and research in the metaphorical complement of early pattern perception, infant color vision, is nearly as long but, anti1 recently, has not been so successful.

168

Marc H . Bornsrein

The significance of color in perceptual development is patent. Objects in the visual world mainly differ in four ways-location, size, shape, and color-and in visual animals color facilitates signal detection, localization, tracking, and recognition (Walls, 1942). These visual functions are all-important in biological evolution. The human factors literature shows that color coding reduces search time and facilitates identification better than coding based on brightness, size, or geometrical shape (Christ, 1975). Investigators of every hue, then, agree that color enhances common visual functions, including intrinsic contrasts, visibility, object evaluation, and constancy, in the different species that possess it. In addition to its information content, color is known to have influential affective qualities. Colors therefore clarify and accentuate and by their attractiveness are thought to induce vigilence in attention. By virtue of both its salience and information content, color surpasses black and white or monochrome. This chapter has surveyed behavioral data on the early ontogeny of chromatic vision and infomation processing in man. Since detailed pattern perception and color vision are mutually segregated to central vision, and since visual patterns can be differentiated or resolved either by brightness or hue, a major question motivating developmentalists has been, “Do infants see color?” Most early investigations of the development of color vision constituted essentially color preference studies with brightness or other variables uncontrolled. The brightness problem aside, simple color preference studies fell far short of describing or evaluating the basic sensory or perceptual functions characteristic of early perceptual development. Recent anatomical, electrophysiological, and behavioral observations indicate, however, that central vision in infants and young children is primary and that photopic vision in man is functional near the beginning of extrauterine life. Moreover, contemporary investigators have adapted psychophysical techniques and controls to the study of the development of chromatic vision with manifest success. In specific, we have reviewed psychophysical studies (e.g., of brightness discrimination, spectral sensitivity, the neutral zone, and wavelength discrimination) as well as early perceptual and cognitive studies (e.g., of hue categorization, color memory, and color preference). These studies show that infants can see, process, remember, and act on chromatic information in their environment. Much of the young infant’s behavior parallels mature behavior and conforms with expectancies of a trichromatic model of color vision. Indeed, the young child is early adapted to process and use the chromatic information in his environment in reasonably sophisticated ways. His first knowledge of the world grows out of the distribution of his visual attention, and the child, as we have seen, differentiates and responds to different features in his world according to their color, organizes his world by color, just as h e may remember different aspects of it on the basis of their color. Intriguingly, two central questions in this

Chron7citic Vision in infancy

I69

research area, namely, whether the neonate sees color and whether the quality of color vision changes early in life, remain unanswered.

Appendix Addendum to C. H. Darwin’s “Biographical sketch of a young child” (Kosrnos, 1877, I , 367-376. [a]) “Darwin, who has previously criticized Geiger’s color theory in the pages of Kosmos, included in a cover letter to the above article the following observations on the development of the color sense in his own children. The author’s intent is to encourage further observation [Ed.] “Darwin wrote: “ ‘I carefully followed the mental development of my small children, and I was astonished to observe in two or, as I rather think, three of these children, soon after they had reached the age in which they knew the names of all the ordinary things, that they appeared to be entirely incapable of giving the right names to the colors of a color etching. They could not name the colors, although I tried repeatedly to teach them the names of the colors. I remember quite clearly to have stated that they are color blind. But afterwards this turned out t o be an ungrounded apprehension. When I told this fact to another person, he told me that he had observed a rather similar case. The difficulty which small children feel, whether in discrimination or, much more probably, in naming the colors, seems therefore to merit further investigation. 1 will add that formerly it looked to me as if the sense of taste, at least with my own children when they were still very young, differed from the adult Sense of taste; this shows itself by the fact that they did not refuse rhubarb with some sugar and milk which is for us an abominably disgusting mixture and by the fact that they strongly preferred the most sour and most tart fruits, as for instance unripe gooseberries and Holz apples.’ ”

ACKNOWLEDGMENTS Much of the research reported in this chapter was supported by the Grant Foundation, the Carnegie Corporation of New York, and the National Institutes of Health (Grant Nos. 1 F22 MH58197-01 and 1 R03 MH28734-01). Helen G. Bornstein. Charles G . Gross, Lawrence E. Marks, Margaret Ruddy, Davida Y. Teller, Anthony A. Wright, and Michael Lewis all contributed valuable comments to an earlier version of this chapter; Rose Fioravanti, Sharon Olsen, and Carol Smith each aided in its preparation.

REFERENCES Abravanel, E. How children combine vision and touch when perceiving the shape of objects. Perception & Psychophysics, 1972, 12, 171-175. Adkinson, C. D., & Berg, W. K. Cardiac deceleration in newborns: Habituation, dishabituation, and offset responses. Journal of Experimental Child Psychology, 1976, 21, 46-60. Adrian, E. D. The electric response of the human eye. Journal of Physiology (London), 1945, 104, 84-104. Aristotle. De coloribus. In W. D. Ross (Ed.), The works of Arisrotle (T. Loveday & E. S. Forster, trans.). (Vol. 6 ) . Oxford: Clarendon, 1913. Arlitt, A. H. Psychology of infancy and early childhood. New York: McGraw-Hill, 1946.

I70

Marc H. Bornstein

Armington, J. C. The electroretinogram. New York: Academic Press, 1975. Arnheim, R. Visual thinkicg. Berkeley: University of California Press, 1971. A s h , R. N., & Salapatek, P. Saccadrc localization of visual targets by the very young human infant. Perception & Psychophysics, 1975, 17, 293-302. Balazs, E. A. Molecular morphology of the vitreous body. In G . K. Smelser (Ed.), The structure of the eye. New York: Academic Press, 1961. Baldwin, J. M. Distance and color perception by infants. Science, 1893, 21, 231-232. Baldwin, J. M. Mental development in the child and the race. New York: Macmillan, 1895. Banks, M. S., & Munsinger, H. Pupillometric measurement of difference spectra for three color receptors in an adult and a four--year-old.Vision Research. 1374, 14, 813-817. Banks, M. S., & Salapatek, P. Infanr color vision? Critical note. Unpublished manuscript, University of Minnesota, 1975. Barnet, A. B., Lodge, A., & Armington, J. C. Electroretinogram in newborn human infants. Science, 1965, 148, 651-641. Bartleson, C. J. Memory colors of familiar objects. Journal of the Optical Society ofAmerica, 1960, 50,73-77. Beare, A. C. Color-name as a function of wave-length. American Journal of Psychology, 1963, 76, 248 -256. Beebe-Center, J. G. The psychology ofpleasantness and unpleasantness. New York: Van Nostrand, 1932. Belmont, J. M. Relations of age and intelligence to short-term color memory. Child Development, 1972, 43, 19-29. Berkeley, G. A n essay toward a new theory of vision. Oxford: Cluendon, 1901. (Originally published, 1709.) Berlin, B., & Kay, P. Basic color terms: Their universality and evolution. Berkeley: University of California Press, 1969. Binet, A. Perception d’infants. Revue Philosophique, 1890, 30, 580. Binet, A , , & Simon, T. [The development of intelligence in the child.] In H. H. Goddard (Ed.), The development of intelligence in children (E. S . Kiffe, trans.). Baltimore: Williams & Wilkins, 1916. (Originally published, 1908.) Boettner, E. A , , & Wolter, J. R. Tmnsmission of the ocular media. Investigative Ophthalmology, 1962, 1, 776-783. Bond, E. K. Perception of form by the human infant. Psychological Bulletin, 1972, 77, 225-245. Boothe, R.,Teller, D. Y., & Sackett, G. P. Trichromacy in normally reared and light deprived infant monkeys. Vision Research, 1975, 15, 1187-1 191. Bornstein, M. H. Color vision and color naming: A psychophysiological hypothesis of cultural difference. Psychological Bulletin, 1973, 80, 257-285. Bornstein, M. H. Perceptual generalization: A note on the peak shift. Psychological Bulletin, 1974, 81, 804-808. Bornstein, M. H. Hue is an absolute code for young children. Nature (London), 1975, 256, 309310. (a) Bornstein, M. H. The influence of visual perception on culture. American Anthropologist, 1975,77, 774-798. (b) Bornstein, M. H. On light and the aesthetics of color: Lumia kinetic art. Leonardo, 1975, 8, 203-212. (c) Bornstein, M. H. Qualities of color vision in infancy. Journal of Experimental Child Psychology, 1975, 19, 401419. (d) Bornstein, M. H. Infants are trichromats. Journal of Experimental Child Psychology, 1976, 21, 425-445. (a)

Chromatic Vision in Infancy

171

Bornstein, M. H. Infants’ recognition memory for hue. Developmental Psychology, 1976, 12, 185-191. (b) Bornstein, M. H. Name codes and color memory. American Journal of Psychology, 1976, 89, 269-279. (c) Bornstein, M. H. Developmental pseudocyananopsi-a: Ontogenetic change in human color vision. American Journal of Optometry and Physiological Optics, 1977, 54, 464469. Bornstein, M. H. Color-vision defect and development. Manuscript in preparation, 1977. (b) Bornstein, M. H. Discrimination and generalization following young infants familidzation with chromatic stimuli. Manuscript in preparation, 1977. (c) Bornstein, M. H. On the development of color naming: Data and theory. Manuscript in preparation, 1977. (d) Bornstein, M. H. Visual behavior of the young human infant: Relationships between chromatic and spatial perception and the activity of underlying brain mechanisms. Journal of Experimental Child Psychology, 1978, in press. Bornstein, M. H . , Kessen, W., & Weiskopf, S. The categories of hue in infancy. Science, 1976, 191, 201-202. (a) Bornstein, M. H., Kessen, W., & Weiskopf, S. Color vision and hue categorization in young human infants. Journal of Experimental psycho log^: Human Perception and Performance, 1976, 2, 115-129. (b) Bornstein, M. H . , &Marks, L. E. Photopic luminosity measured by the method of critical frequency. Vision Research, 1972, 12, 2023-2033. Bourne, L. E. Human conceptual behavior. Boston: Allyn & Bacon, 1966. Bower, T. G. R. Development in infancy. San Francisco: Freeman, 1975. Bower, T. G. R., Broughton, J . M., & Moore, M. K. Infant responses to approaching objects: An indicator of response to distal variables. Perception & Psychophysics, 1970, 9, 193-196. Boynton, R. M. Color vision. In L. A. Riggs & J . Kling (Eds.), Woodworth and Schlosberg’s experimental psychology. New York: Holt, 1971. Boynton, R. M. Color, hue, and wavelength. In E. C. Carterette & M. P. Friedman (Eds.),Handbook of perception (Vol. 5 ) . New York: Academic Press, 1975. Boynton, R, M., & Gordon, J . Bezold-Briicke hue shift measured by color-naming technique. Journal of the Optical Society of America, 1965, 55, 78-86. Boynton, R. M., & Greenspon, T. S. The distinctness of borders formed between equally saturated, psychologically unique fields. Vision Research. 1972, 12, 495-507. Brackbill, Y. Research and clinical work with children. In R. .4. Bauer (Ed.), Some views on Soviet psychology. Washington, D.C.: American Psychological Association, 1962. Bradbury, H . , & Nelson, T. M. Transitivity and the patterns of children’s preferences. Developmental Psychology, 1974, 10, 55-64. Brian, C . R.,& Goodenough, F. L. The relative potency of color and form perception at various ages. Journal of Experimental Psychology, 1929, 12, 197-213. Bronson, G . Postnatal growth of visual capacity. Child Development, 1974, 45, 873-890. Brown, C. J . The effects of preference for visual complexity on habituation of visual fixation in infants. Child Development, 1974, 45, I !66-1169. Brown, J. L., & Mueller, C. G. Brightness discrimination and brightness contrast. In C. H . Graham (Ed.), Vision and visual perception. New York: Wiley, 1965. Brown, R. W., & Lenneberg, E. H. A study in language and cognition. Journal of Abnormal and Social Psychology, 1954, 49, 454462. Bruner, J. S . , & Koslowski, B . Visual preadapted constituents of manipulatory action. Perception, 1972, 1, 3-14. Bryant, P. E. Perception and understanding in young children. New York: Basic Books, 1974.

172

Marc H . Bornstein

Carroll, J. B. (Ed.). Language, thought and reality. Cambridge, Mass.: MIT Press, 1964. Champneys, F. H. Notes on an infant. Mind, 1881, 6, 104-107. Chapanis, A. Spectral saturation and its relation to color-vision defects. Journal of Experimental Psychology, 1944, 34, 2 4 4 4 . Chase, W. P. Color vision in infants. Journal of Experimental Psychology, 1937, 20, 203-222. Chevallereau, A., & Polack, A. De la coloration jaune de la macula. Annales dOculistique, 1907, 138, 241-253. Child, I. L., Hansen, J. A., & Hornbeck, F. W. Age and sex differences in children’s color preferences. Child Development, 1968, 39, 237-247. Christ, R. E. Review and analysis of color coding research for visual displays. Human Factors, 1975, 17, 542-570. Clark, E. V. What’s in a word? On the child’s acquisition of semantics in his first language. In T. E. Moore (Ed.),Cognitive development and the acquisition of language. New York Academic Press, 1973. Clifford, L. T., &Calvin, A. D. Effect of age on the discriminative learning of color and brightness by children. American Journal of Psychology, 1958, 71, 766-767. Cohen, L. B. A two-process model of infant visual attention. Merrill-Palmer Quarterly. 1973, 19, 157-180. Cohen, L. B., & Gelber, E. R. Infant visual memory. In L. Cohen & P. Salapatek (Eds.), Infant perception: From sensation to cognition. New York: Academic Press, 1975. Cohen, L. B., Gelber, E. R., & Lazar, M. A. Infant habituation and generalization to differing degrees of stimulus novelty. Journal of Experimental Child Psychology, 1971, 11, 379-389. Cohen, L. B., & Salapatek, P. (Ed:;.).Infant perception; From sensation to cognirion. New York Academic Press, 1975. Colavita, F. B. Human sensory dominance. Perception & Psychophysics, 1974, 16, 409412. Colby, M. C., & Robertson, J . B. Genetic studies in abstraction. Journal of Comparative Psychology, 1942, 33, 385401. Conklin, H. C. Color categorization (Review of Basic color terms: Their universalify and evolution by B. Berlin & P. Kay.) American Anthropologist, 1973, 75, 931-942. Coombs, C. Theory of data. New York: Wiley, 1964. Corah, N . L., &Gross, J. B. Hue, brightness, and saturation variables in color-form matching. Child Development, 1967, 38, 137-142. Coren, S . , & Girgus, J. S . Density of human lens pigmentation: In vivo measures over an extended age range. Vision Research, 1972, 12, 343-346. Cornell, E. H. Infants’ discrimination of photographs of faces following redundant presentations. Journal of Experimental Child Psychology, 1974, 18, 98-106. Courtney, G. R., & Heath, G. G. Color vision deficiency in the mentally retarded Prevalance and a method of evaluation. American Journal of Mental Deficiency, 1971, 76, 48-52. Daehler, M., Bukatko, D., Benson, K., & Myers, N . The effects of size and color cues on the delayed response of very young children. Bulletin of the Psychonomic Society, 1976, 7,6548. Dale, P. S. Color naming, matching. and recognition by preschoolers. Child Development, 1969,40, 1135-1 144. Darwin, C. H. [A biographical sketch of a young child.] Kosrnos, 1877, 1, 367-376. (a) Darwin, C. H. A biographical sketch of an infant. Mind, 1877, 2 , 285-294. (b) Dashiell, J . F. Children’s sense of harmonies in color and tones. Journal of Experimental Psychology, 1917, 2,466475. Davidoff, J. The effect of color distraction on a matching task in Ghanaian children. International Journal of Psychology, 1972, ’7, 141-144. Davidoff, J. The psychological relationship between lightness and saturation. Perception & Psychophysics, 1974, 16, 79-8.3.

Chromatic Vision in Infancy

I73

Dearborn, G. V. N. Motor-sensory development. Baltimore: Warwick & Tork, 1910. DeLoache, J. S. Rate of habituation and visual memory in infants. Child Development, 1976, 47, 145- I 54. DeValois, R . L . Central mechanisms of color vision. In R. Jung (Ed.), Handbook of sensory physiology. Vol. VI1/3A. Cennal processing of visual information. Berlin and New York Springer-Verlag, 1973. De Valois, R. L., Abramov, I . , &Jacobs, G . H. Analysis of response patterns of LGN cells. Journal of the Optical Sociefy of America, 1966, 56, 966-977. De Valois, R. L., Abramov, I . , & Mead, W . R . Single cell analysis of wavelength discrimination at the lateral geniculate nucleus in the macaque. Journal of Neurophysiology, 1967,30,415433. De Valois, R . L., & De Valois, K. K. Neural coding of color. In E. C. Carterette & M. P. Friedman (Eds.), Handbook of perception (Vol. 5). New York: Academic Press, 1975. De Valois, R. L., & Marrocco, R. T. Single cell analysis of saturation discrimination in the macaque. Vision Research, 1973, 13, 701-71 I . De Valois, R. L., Morgan, H. C., Polson, M. C., Mead, W. R., & Hull, E. M. Psychophysical studies of monkey visio-I. Macaque luminosity and color vision tests. Vision Research, 1974, 14, 5367. Dimmick, F. L., & Hubbard, M. R. The spectral location of psychologically unique yellow, green, and blue. American Journal of Psychology, 1939, 52, 242-254. (a) Dimmick, F. L., & Hubbard, M. R. The spectral components of psychologically unique red. American Journd of Psychology, 1939. 52, 348-353. (b) Dobson, V. Spectral sensitivity of the 2-month infant as measured by the visually evoked cortical potential. Vision Research, 1976, 16, 367-374. Dodd, C., & Lewis, M. The magnitude of the orienting response in children as a function of changes in color and contour. Journal of Experimental Child Psychology, 1969, 8, 296-305. Doris, J., Casper, M., & Poresky, R. Differential brightness thresholds in infancy. Journal of Experimental Child Psychology, 1967, 5, 522-535. Doris, J., & Cooper, L. Brightness discrimination i n infancy. Journal of Experimental Child Psychology, 1966, 3, 31-39. Duke-Elder, S . , & Cook, C. System of ophrhalmology (Vol. 3 ) . St. Louis: Mosby, 1963. Ebbinghaus, H. [Memory: A connibution to experimental psychology] (H. A. Ruger & C. E. Bussenius, trans.). New York: Dover, 1964. (Originally published, 1885.) Edwards, A. S. Effect of color on visual depth perception. Journal of General Psychology, 1955, 52, 331-333. Ekman, G. Dimensions of color vision. Journal of Psychology, 1954, 38, 467474. Ekman, G. Contributions to the psychophysics of color vision. Studium Generale. 1%3, 16, 5444. Essock, S. M. Color perception and color classification. In D. M. Rumbaugh (Ed.), Language learning by a chimpanzee: The LANA project. New York: Academic Press, 1977. Eysenck, H. J. A critical and experimental study of colour preferences. American Journal of Psychology, 1941, 54, 385-394. Fagan, J. F. Infant color perception. Science. 1974, 183, 973-975. Fagan, J. F. Infant hue discrimination? Science, 1975, 187, 277. Fantz, R. L. The origin of form perception. Scientific American, 1961, 204, 66-72. Fantz, R. L. Visual experience in infants: Decreased attention to familiar patterns relative to novel ones. Science, 1964, 146, 668670. Fantz, R. L., Ordy, J. M., & Udelf, M. S. Maturation of pattern vision in infants during the first six months. Journal of Comparative and Physiological Psychology, 1962, 55, 907-917. Farnham-Diggory,S., & Gregg, L. W. Color, form, and function as dimensions of natural classification: Developmental changes in eye movements, reaction time, and response strategies. Child Development, 1975.46, 101-1 14.

I74

Marc H . Bornsrein

Fichsel, H. Visual evoked potentials in prematures, newborns, infants and children by stimulation with colored light. Electroencephalography and Clinical Neurophysiology, 1969, 27, 660. (Abstract) Friedman, S. Newborn visual attention to repeated exposure of redundant vs. "novel" targets. Perception & Psychophysics, 1972, 12, 291-294. Gaines, R. Variables in color perception of' young children. Journal ofExperimenta1 Child Psychology, 1972, 14, 196-218. Gaines, R . , &Little, A. C. Developmental color perception. Journal of Experimental Child Psychology, 1975, 2 0 , 4 6 5 4 8 6 , Garbini, A. Evoluzione del senso cromatico nella infanzia. Florence: Del Dott, 1894. Gardiner, P. A. A colour vision test for young children and the handicapped. Developmental Medicine and Child Neurology, 1973, 15, 4 3 7 4 4 0 . Garner, W. R. Attention: The processing of multiple sources of information. In E. C. Carterette & M. P. Friedman (Eds.), Hundb'ook of perception (Vol. 2). New York: Academic Press, 1974. Garth, T. R . Race psychology. New York: McGraw-Hill, 1931. Garth, T. R., & Porter, E. The color preferences of 1032 young children. American Journal of Psychology, 1934, 46, 448451. Geldard, F. A. The human senses. New York: Wiley, 1972. Gesell, A. Infant vision. Scienrific .4merican, 1950, 182(2), 20-22. Gesell, A., Ilg, F. L., & Bullis, G. E. Vision: Its development in infant and child. New York Hoeben, 1949. Gibson, E. J. Principles of perceptual learning and development. New York: Appleton, 1969. Gibson, J. J . The senses considered as perceptual systems. Boston: Houghton, 1966. Gilbert, J. A. Researches in the mental and physical development of school children. Studies of the Yale Psychological Laborator), 1894, 2 , 40-150. Gilbert, J. A. Age changes in color matching. Journal ofGerontology. 1957, 12, 210-215. Glanzer, M., & Clark, W. H. Accuracy of perceptual recall: An analysis of organization. Journal of Verbal Learning and Verbal Behavior, 1962, 1, 289-299. Goesling, W . J . Stimulus presentation methods and ability to shift dimensions in color-form preference of retardates. Perceptual and Motor Skills, 1973, 36, 807-810. Gothlin, G. F. An analysis of the perception defect in published and spectroscopically investigated cases of congenital tritanopia. Acta Ophthalmologica, 1943, 21, 88-96. Gothlin, G. F. Experimental determination of the short wave fundamental color in man's color sense. Journal of the Optical Society of America, 1944, 34, 147-158. Goodenough, F. L. Developmental psychology. New York: Appleton, 1945. Goodenough, F. L., & Anderson, .I. E. Experimental childstudy. New York: Century, 1931. Goodnow, J . J . The role of modalitii:s in perceptual and cognitive development. Minnesota Symposia on Child Psychology, !971, 5, 3-28. Gottesman, M . Stage development of blind children: A Piagetian view. New Outlookfor the Blind. 1976, 70, 94-100. Graham, B. V., Turner, M. E., & IIurst, D. C. Derivation of wavelength discrimination from color naming. Journal of the OpticaI Society of America, 1973, 63, 109-1 1 1 . Graham, C. H. (Ed.). Vision and visual perception. New York: Wiley, 1965. Graham, C. H., & Hsia, Y. The disuiminations of a normal and a color-blind eye in the same person. Proceedings of the American Philosophical Society, 1958, 102, 168-173. Graham, C. H., Sperling, H. G . , Ilsia, Y . , & Coulson, A. H. The determination of some visual functions of a unilaterally color-blind subject: Methods and results. Journal of Psychology, 1961, 51, 3-32. Granger, G. W. An experimental shtdy of colour preferences. Journal of General Psychology, 1955, 52, 3-20.

Chronrtrtic Vtsrorr

in

Infanc?

I75

Grether, W. F. Color vision and color blindness in monkeys. Comparurive Psychological Monographs, 1939, 15 (4, Whole No. 76). Griitzner, P. Acquired color vision defects. In D. Jameson & L. M. Hurvich (Eds.), Handbook of sensory physiology. Vol. VII/4. Visual psychophysics. Berlin and New York: Springer-Verlag, 1972. Guilford, J. P. The affective value of color as a function of hue, tint, and chroma. Journal of Experimental Psychology, 1934, 17, 342-370. Guilford, J. P. There is a system of color preferences. Journal of the Optical Society of America, 1940, 30, 455459. Guilford, J . P., & Smith, P. C. A system of color preferences. American Journal of Psychology, 1959, 72, 487-502. Hale, G. A , , & Green, R . Z . Children's attention to stimulus components with variation in relative salience of components and degree of stimulus integration. Journal of Experimental Child Psychology, 1976, 21, 446459. Hall, G. S . Notes on the study of infants. Pedagogical Seminary, 1891, 1, 127-138. Harris, P . , & MacFarlane, A. The growth of the effective visual field from birth to seven weeks. Journal of Experimental Child Psychology, 1974, 18, 340-348. Heath, G. G. Color vision. In M. J . Hirsch & R . E. Wilk (Eds.), Vision of children: A n optometric symposium. Philadelphia: Chilton, 1963. Heath, G. G. The handicap of color blindness. Journal of the Optometric Association, 1974, 45, 62-69. Helmholtz, H. von [Physiological optics] (J. P. C. S . Southall, trans.). New York: Dover, 1962. (Originally published, 1866.) Helson, H . , & Lansford, R . The role of spectral energy of source and background color on the pleasantness of object colors. flluminafion Engineering, 1967. Hering, E. [Outlines of a theor?, of the light senye] (L. M. Hurvich & D. Jameson, trans.). Cambridge, Mass.: Harvard University Press, 1964. (Originally published, 1878.) Hershenson, M. Visual discrimination in the human newborn. Journal of Comparative and Physiological Psychology, 1964, 58, 270-276. Hochberg, J. Perception: Color and shape. In J . W. Kling & L. A . Riggs (Eds.). Woodworth & Schlosberg's experimental psychology. New York: Holt, 197 1, Holden, W. A,, & Bosse, K. K. The order of development of color perception and of color preference in the child. Archives of Ophthalmology, 1900, 29, 261-277. Holmes, S . K.. Kelton, J. J . , & Pollack. R . H. Visual acuity f o r single lines as a function of hue in children. Paper presented at the 14th annual meeting of the Psychonornic Society, St. Louis, November 1973, Hsia, Y., & Graham, C. H. Colcr blindness. In C. H . Graham (Ed.), Vision and visual perception. New York: Wiley, 1965. Hurlock, E. B. Child development (5th ed.). New York: McGraw-Hill, 1972. Hurvich, L. M. Color vision deficiencies. In D. Jameson & L. M. Hurvich (Eds.), Handbook of sensory physiology. Vol. VW4. Visual psychophgsics. Berlin and New York: Springer-Verlag, 1972. Irwin, 0. Effect of strong light on the body activity of newborns. Journal of Comparative Pqchology, 1941, 32, 233-236. Irwin, O., & Weiss, L. Differential variations in the activity and crying of the newborn infant under different intensities of light: A comparison of observational with polygraphic findings. University of Iowa Studies on Child Welfare. 1934, 9, 137-147. Jacobs, G. H. Color vision. Annual Review of Psychology, 1976, 27, 63-89. Jacobs, G. H., & Gaylord, H. A. Effects of chromatic adaptation on color naming. Vision Research, 1967, 7, 645453. Jastrow, J. The popular aesthetics of color. Popular Science Monrhly, 1897, 50, 361-368.

176

Marc H . Bornstein

Jersild, A. T., Telford, C. W., & Sawrey, J . M. Childpsychology (7th e d . ) . Englewood Cliffs, N.J.: Prentice-Hall, 1975. Jones-Molfese, V. Responses of neonates to colored stimuli. Child Development. 1977, 48, 10921095. Judd, D. B. Basic correlates of the visual stimulus. In S. S. Stevens (Ed.), Handbook ofexperimental psychology. New York: Wiley, 1951. (a) Judd, D. B. Report of the US. Secretariat Commission on colorimetry and artificial daylight. CIE Proceedings, 1951, 1, 11. Paris, Bureau Central CIE. (Data reproduced in G . Wyszecki & W. S. Stiles, Color science. New York: Wiley, 1967. P. 436.) (b) Kalmus, J. Pure (unique) green and a neutral zone in the spectrum of colour defectives. Annals of Human Genetics, 1972, 35, 375-377. Karmel, B. Z., Hoffman, R. F., & IFegy, M. J . Processing of contour information by human infants evidenced by pattern-dependent evoked potentials. Child Development, 1974, 45, 39-48. Karmel, B. Z., & Maisel, E. B. A neuronal activity model for infant visual attention. In L. B. Cohen & P. Salapatek (Eds.), Infant perception: From sensation to cognition. New York Academic Press, 1975. Kasatkin, N. I., & Levikova, A. M. The formation of visual conditioned reflexes and their differentiation in infants. Journal of General Psychology, 1935, 12, 416435. Katz, D. The world ofcolour. London: Tmbner, 1913. Kaufman, L. Sight and mind: An introduction to visual perception. London and New York: Oxford University Press, 1974. Kelton, .I.J., & Pollack, R. H. Visual acuity for single lines us a function of hue and age. Unpublished manuscript, University of Georgia, 1974. Kessen, W., Haith, M. M., & Salapatek, P. H. Human infancy: A bibliography and guide. In P. H. Mussen (Ed.), Carmichuel's manual of child psychology (Vol. 1). New York: Wiley, 1970. Kling, J. W., & Riggs, L. A. (Eds.). Woodworth and Schlosberg's experimental psychology. New York Holt, 1971. Koffka, K. The growrh of the mind: An introduction to childpsychology. New York Harcourt, 1927. Kopp, J., & Lane, H. Hue discrimination related to linguistic habits. Psychonomic Science, 1968, 11, 61-62. Ladd-Franklin, C. Color-introspection on the part of the Eskimo. Psychological Review, 1901, 8, 396402. Ladd-Franklin, C. Colour and colour theories. London: Kegan, Paul, Trench, Trubner, 1929. Lakowski, R. Age and colour vision. Advancement of Science, 1958-1959, 15, 231-236. Lashley, K. S., & Wade, M. The Pavlovian theory of generalization. Psychological Review, 1946, 53, 72-87. Lawler, C. O., & Lawler, E. E. Colormood associations in young children. Journal of Genetic PSychOlogy. 1965, 107, 29-32. Lee, L. C . Concept utilization in preschool children. Child Development. 1965, 36, 221-2223. Lenneberg, E. H. Biological foumfations of language. New York: Wiley, 1967. Lepore, F., Cardu, B., Rasmussen, T., & Malmo, R. B. Rod and cone sensitivity in destriate monkeys. Brain Research, 19'75, 93, 203-221. Lewis, M., & Baumel, M. H. A ljtudy in the ordering of attention. Perceprual and Motor Skills, 1970, 31, 979-990. Liebert, R. M., Poulos, R. W . , & Strauss, G. D. Developmenral psychology. Englewood Cliffs, N.J.: Prentice-Hall, 1974. Lodge, A , , Armington, J. C., Barnet, A. B., Shanks, B. L., & Newcomb, C. N. Newborn infants' electroretinograms and evoked electroencephalographic responses to orange and white light. Child Development, 1969, 40, 267-293.

Chromtrric Vision in Infancy

I77

Lorenz, A . B., & McClure, W. E. The influence of color blindness on intelligence and achievement of college men. Journal ofApplied Psychology, 1935, 19, 320-330. Luckey, G. W. A . Comparative observations on the indirect color range of children, adults, and adults trained in color. American Journal of Psychology, 1893, 6 , 489-504. Ludvigh, E., & McCarthy, E. F. Absorption of visible light by the refractive media of the human eye. Archives of Ophthalmology, 1938, 20, 37-51. Major, D . R . First steps in mental growth. New York: Macmillan, 1915. Mandola, J . The role of color vision anomalies in elementary school achievement. Journal of School Health, 1969, 39, 633-636. Marg, E., Freeman, D. N., Peltzman, P . , & Goldstein, P. J . Visual acuity development in human infants: Evoked potential measurements. fnvestigafive Ophthalmology, 1976, 15, 150-153. Marks, L. E. Sensory processes: The new psychophvsics. New York: Academic Press, 1974. Marsden, R. E. A study of the early color sense. Psychological Review, 1903, 10, 3 7 4 7 . (a) Marsden, R. E. The early color sense-further experiments. Psychologcinl Review, 1903, 10, 297300. (b) Martin, R . M. Effects of familiar and complex stimuli on infant attention. Developmental Psychology. 1975, 11, 178-185. McCain, C . N., & Karr, A. C. Color and subjective distance. U.S. Army Human Engineering Laboratories Technical Memorandum. 1970, No. 20-70. McCall, R . B., & Kagan, J . Stimulus-schema discrepancy and attention in the infant. Journal of Experimental Child Psychology. 1967, 5, 38 1-390. McDougall, W. An investigation of the color sense of two infants. British Journal ofPsychology, 1908, 2 , 338-352. McFarland, W. N., & Munz, F . W. Part 111: The evolution of photopic visual pigments in fishes. Vision Research, 1975, 15, 1071-1080. McKenzie, B. E., & Day, R . H. Object distance as a determinant of visual fixation in early infancy. Science, 1972, 178, 1108-1 110. McKusick, V . A. On the X chromosome of man. Quarterly Review of Biology, 1%2, 37, 69-175. Melkman, R., Koriat, A., & Pardu, K. Preference for color and form in preschoolers as related to color and form differentiation. Child Development, 1976, 47, 1045-1050. Melton, A. W., & Martin, E. (Eds.). Coding processes in human memory. Washington, D.C.: Winston, 1972. Mendlewicz, J . , Fleiss, J. L . , & Fieve, R . R. Evidence for X-linkage in transmission of manicdepressive illness. Journal of the Americun Medical Association. 1972, 222, 1624-1627. Meyer, H. Uber den Gelb-Blausinn der Protanopen und Deuteranopen irn Vergleich mit dem der normalen Trichromaten. Klinische Monttrshliitter fur Augenheilkunde, 1932, 88, 4%-507. Milewski, A., & Siqueland, E. R. Discrimination of color and pattern novelty in one-month human infants. Journal of Experimental Child Psychology, 1975, 19, 122-136. Miller; D. J . Visual habituation in the human infant. Child Development, 1972, 43, 481493. Miller, G. A. The colors of philosophy and psychology. Paper presented at the Conference on Philosophy of Psychology, Cambridge, May, 1974. Miranda, S. B., & Fantz, R . L. Recognition memory in Down’s syndrome and normal infants. Child Development, 1974, 45, 651460. Morgan, G . A , , Goodson, F . E., & Jones, T . Age differences in the associations between felt temperatures and color choices. Americun Journal of Psychology. 1975, 88, 125-130. Munn, N. L. The evolution and growth of human behavior. Boston: Houghton, 1955. Munsinger, H., &Banks, M. S. Pupillometry as a measure of visual sensitivity among infants, young children, and adults. Developmental Psychology, 1974, 10, 677482.

178

Marr H . Bornstein

Mussen, P. H., Conger, J . J . , & Kagan, J . Childdevelopment andpersonality (4th ed.). New York: Harper, 1974. Myers, C. S. Some observations on the development of the color sense. British Journal of Psychology, 1908, 2, 353-362. Nagel, V. A. Observation on the color-sense of a child. Journal of Comparative Neurology and Psychology, 1906, 16, 217-230. Newton, I . New theory about light and colors. Philosophical Transactions of the Royal Society, 1671-1672, 80, 3075-3087. Nickerson, D., & Newhall, S. M. A psychological color solid. Journal of the Optical Society of America, 1943, 33, 419422. Norren, D. V., & Vos, J . J . Speclral transmission of the human ocular media. Vision Research, 1974, 14, 1237-1244. Norsworthy, N., & Whitley, M. T. The psychology of childhood. New York: Macmillan, 1920. Nunnally, J . C., & Lemond, L. C. Exploratory behavior and human development. In H. W . Reese (Ed.),Advances in child development and behavior (Vol. 8). New York: Academic Press, 1973. Odom, R. D. Effects of perceptual salience on the recall of relevant and incidental dimensional values: A developmental study. Journal of Experimental Psychology. 1972, 92, 285-291. Odom, R. D., & Guzman, R. D. Development of hierarchies of dimensional salience. Developmental Psychology, 1972, 6 , 271-287. Osgood, C. E. Merhod and theory in experimental psychology. London and New York Oxford University Press, 1953. Oster, H. Color perception in fen-week-old infants. Paper presented at the Biennial meeting of the Society for Research in Child Development, Denver, April 1975. Oyama, T . , & Yamamura, T. The effect of hue and brightness on the depth perception in normal and color-blind subjects. Psychologia, 1960, 3, 191-194. Pancratz, C. N., & Cohen, L. B. Recovery of habituation in infants. Journal of Experimenfal Child Psychology, 1970, 9, 208-216. Fayne, M. C. Color as an independent variable in perceptual research. Psychological Bulletin, 1964, 61, 199-208. Peeke, S. C., & Stone, G . C. Focal and nonfocal processing of color and form. Perception & Psychophysics, 1973, 14, 7 1-80, Peeples, D. R. Hereditary color deficiencies: Are they congenital? Paper presented at the Spring meeting, Association for Research in Vision and Ophthalmology, Sarasota, Florida, April 1976. Peeples, D. R., &Teller, D. Y. Color vision and brightness discrimination in two-month-old human infants. Science, 1975, 189, 1102-1 103. Peeples, D. R., &Teller, D. Y. White-adapted photopic spectral sensitivity in human infants. Vision Research, 1978, 18, in press. Peiper, A. Uber die Helligkeits- und Farbenempfindungen der Friihgerburten. Archives fur Kinderheilklinik, 1927, 80, 1-20, Peiper, A. Comments upon J . M. Smith’s work, “The relative brightness values of three hues for newborn infants.” Child Development, 1937, 8 , 299-300. Peiper, A. Cerebrcl function in infancy and childhood. New York: Consultants Bureau, 1963. (Originally published, 1961.) Perlmutter, M., & Myers, N. A. Young children’s coding and storage of visual and verbal material. Child Development, 1975, 461, 215-219. Pick, H. L., Warren, D. H . , & Hay, J . C. Sensory conflict in judgments of spatial direction. Perception & Psychophysics, 1969, 75, 103-107. Polikanina, R. I. Development of color vision in newborn infants. Zhurnal Vysshei Nervnoi Deyarel’nosti imeni I . P . Paclova, 1968, 18, 1050-1059.

Chromatic Vision in Infancy

I79

Pollack, R . H. Mueller-Lyer illusion: Effect of age, lightness contrast, and hue. Science, 1970, 170, 93-95. Polyak, S . The vertebrate visual system. Chicago: University of Chicago Press, 1957. Posner, M . I. Cognition: A n introduction. Glenview, 111.: Scott, Foreman, 1973. Posner, M. I., Nissen, M. J . , & Klein, R. M . Visual dominance: An information-processing account of its origins and significance. Psychological Review, 1976, 83, 157-171. Pratt, K. C., Nelson, A. K . , &Sun, K. H. The behavior of the newborn infant. OhioState University Studies, 1930, 10, 44-78. Preyer, W. The mind of the child. New York: Apple!on, 1890. Ray, V . F. Techniques and problems in the study of human color perception. Southwestern Journal of Anthropology, 1952, 8, 251-259 Ray, V. F . Human color perception and behavioral response. Transactions of the New York Academy ofSciences, 1953, 15, 98-104. Redfield, J . The light sense in newborn infants. University oflowa Studies in Child Welfare, 1939, 16, 105-145. Reese, H. W., & Lipsit!, L. P. Experimental childpsychology. New York: Academic Press, 1970. Regan, D., & Spekreijse, H. Evoked potential indications of colour blindness. Vision Research, 1974, 14, 89-95. Rosch, E. H. Natural categories. Cognitiw Psychology, 1973, 4, 328-350. Rosinski, R. R. The development of visual perception. Santa Monica, Calif.: Goodyear, 1977. Ruddock, K. H. The effect of age upon colour vision: I . Responses in the receptor system of the human eye. Vision Research, 1965, 5 , 3 7 4 5 . (a) Ruddock, K . H. The effect of age upon colour vision: 11. Changes with age in light transmission of the ocular media. Vision Research, 1965, 5 , 47-58. (b) Saayman, G . , Ames, E. W., & Moffett, A . Response to novelty as an indicator of visual discrimination in the human infant. Journal of Experimental Child Psychology, 1964, 1, 189-198. Salapatek, P. Pattern perception ir, early infancy. In L. B. Cohen & P. Salapatek ( a s . ) , Infant perception: From sensation to cognition. New York: Academic Press, 1975. Schaffer, H. R., & Parry, M. H. Perceptual-motor behaviour in infancy as a function of age and stimulus familiarity. British Journal of Psychology, 1959, 60, 1-9. Schaffer, H. R., & Parry, M . H. The effects ot short-term familiarization in infants’ perceptualmotor coordination in a simultaneous discrimination situation. British Journal of Psychology, 1970, 61, 559-559. Schallenberger, M. Professor Baldwin’s method of studying the color-perception of children. American Journal of Psychology, 1896, 8, 560-576. Schaller, M. J. Chromatic vision in human infants: Conditioned operant fixation to “hues” of varying intensity. Bulletin of the Psychonomic Society, 1975, 6, 3 9 4 2 . Scheibner, H. M. O., & Boynton, R. M . Residual red-green discrimination in dichromats. Journal of the Optical Society of America, 1968, 58, 1151-1158. Shearron, G . F. Color vision deficiency in primary school children. Sight Saving Review, 1965, 35, l48--151. Sherman, M . , Sherman, I. C., & Flory, C. D. Infant behavior. Comparative Psychology Monographs, 1936, 12, (4, Whole No. 59). Shinn, M. W. Notes on the development of a child. Berkeley: University of California Press, 1909. Shirley, M . M. T,!zefirst two years. Vol. 2. Intellectual development. Minneapolis: University of Minnesota Press, 1933. Siegel, M. h‘. Discrimination of color. I. Comparison of three psychophysical methods. Journal of the Optical Society of America, 1962, 52, 1067-1070.

1 80

Marc H . Bornstein

Siegel, M. H. Discrimination of color. 111. Effect of spectral bandwidth. Journal of the Optical Society of America, 1963, 53, 874-877. Siegel, M. H. Discrimination of color. IV. Sensitivity as a function of spectral wavelength, 410 through 500 m p . Journal of the Optical Society of America, 1964, 54, 82 1-823. Siegel, M. H., & Dimmick, F. L. Discrimination of color. II. Sensitivity as a function of spectral Journal of the Optical Society of America, 1962, 52, 1071-1074. wavelength, 510 to 630 Siegel, M. H.,& Siegel, D. E. Improving memory for color. Bulletin of the Psychonomic Society, 1976, 7, 461-464. Sinsheimer, R. L.The brain of Pooh: An essay on the limits of mind. American Scientist, 1971, 59, 20-28. Sinson, J., & Wetherick, N. E. Cue salience and learning in severely subnormal children: Effects of varying attentional value of the cues employed in a one-trial learning situation. Journal of Mental Deficiency Research, 1972, 16, 112-1 18. Sinson, J., & Wetherick, N. E. Shod-term retention of colour and shape information in mongo1 and other severely subnormal children. Journal of Mental Dejkiency Research, 1973, 17, 177-182. Sinson, J., & Wetherick, N. E. The nature of the colour retention deficit in Down’s syndrome. Journal of Mental Deficiency Research, 1975, 19, 97-100. Skoff, R., & Pollack, R. H. Visual acuity in children as a function of hue. Perception & Psychophysics, 1969, 6, 244-246. Slain, L. L. Testing for deficient color perception in children. International Ophthalmology Clinics, 1963,3, 697-705. Smith, D. P. Derivation of wavelength discrimination from colour-naming data. Vision Research, 197 I , 11, 739-742. Smith, H. C . Age differences in color discrimination. Journal of General Psychology, 1943, 29, 191-226. Smith, J . M . The relative brightness values of three hues for newborn infants. University of Iowa Studies of Child Welfare, 1936, 12, 91-140. Snyder, C. R . The psychological implications of being color blind. Journal of Special Education, 1973, 7, 51-54. Sokolov, Y. N. [Perception and the conditioned refex] ( S . W. Waydenfeld, trans.). New York: Macmillan, 1963. (Originally published, 1958.) Spears, W. C. Assessment of visual preference and discrimination in the four-month-old infant. Journal of Comparative and Physiological Psychology, 1964, 57, 381 -386. Staples, R . Color vision and color preference in infancy and childhood. Psychological Bulletin, 1931, 28, 297-308. Staples, R . The responses of infants to color. Journal of Experimental Psychology, 1932, 15, 119-141. Steinhardt, J. Intensity discrimination in the human eye. I. Journal of General Physiology, 1936,20, 185-209. Stem, W. Psychology of early childhood. New York: Holt, 1924. Sternheim, C. S., & Boynton, R. M. Uniqueness of perceived hues investigated with a continuous judgmental technique. Journal of Experimental Psychology, 1966, 72, 770-776. Stimimann, F. Uber das Farbenempfinden Neugeborener. Annalen Paediatrie, 1944, 163, 1-25. Stubbs, D. F. Frequency and the bmin. Life Sciences, 1975, 18, 1-14. Suchman, R. G. Color-form preference, discriminative accuracy and learning of deaf and hearing children. Child Development, 1966, 37, 439-45 I . (a) Suchman, R. G. Cultural differences in children’s color and form preferences. Journal of Social Psychology. 1966, 70, 3-10. (b) Suchman, R . G., & Trabasso, T. Color and form preference in young children. Journal of Experimental Child Psychology, 1966, 3, 177-187.

w.

Chromuric Vision in Infancy

181

Teller, D. Y., & Peeples, D. R. Acuiry and spectral sensitivity in human infants. Paper presented at Spring Meeting, Association for Rescarch in Vision and Ophthalmology, Sarasota, Florida, April, 1974. Teller, D. Y., Peeples, D. R., & Sekel, M. Discrimination of chromatic from white light by two-month-old human infants. Vision Research. 1978, 18, in press. Thomas, H. Unfolding the baby’s mind: The infant’s selection of visual stimuli. Psychological Review, 1973, 80,468488. Thuline, H. C. Color blindness in children. Clinical Pediatrics, 1972, 11, 295-299. Trabasso, T., Stave, M., & Eichberg, R . Attribute preferences and discrimination shifts in young children. Journal of Experimental Child Psychology, 1969, 8, 195-209. Tracy, F., & Stimpfl, J. The psychology of childhood. Boston: Heath, 1909. Trincker, D., & Trincker, 1. [The ontogenetic development of brightness perception and color vision in man. 1. The development of brightness vision in infants.] In Y. Brackbill & G . G. Thompson (Eds.), Behavior in infancy and early childhood. New York: Free Press, 1967. (Originally published, 1955.) Trincker, D., & Trincker, 1. 1955. (Cited in Peiper, 1961.) Tronick, E. Stimulus control and the growth of the infant’s effective visual field. Perception & Psychophysics, 1972, 11, 373-376. Tronick, E., & Clanton, C. Infant looking patterns. Vision Research, 1971, 11, 1479-1486. Underwood, B. J., Ham, M., & Ekstrand, B . Cue selection in paired-associate learning. Journal of Experimental Psychology, 1962, 64, 405409. Valentine, C. W. The colour perception and colour preferences of an infant during its fourth and eighth months. British Journal of Psychology, 1913-1914, 6 , 363-386. Valentine, C. W. The psychology of early childhood. Cleveland: Sherwood Press, 1942. Valentine, C. W. Experimental psychology of beauty. London: Methuen, 1962. Vannas, S., & Tier, H. Observations on structures and age changes in the human sclera. Acta Ophthalmologica. 1960, 38, 268-279. Vernon, M. D. The psychology of perception. Baltimore: Penguin Books, 1%2. Verriest, G. Further studies on acquired deficiency of color discrimination. Journal of the Optical Society of America, 1963, 53, 185-195. von Frisch, K. Bees: Their vision, chemical senses. and language. Ithaca, N.Y.: Cornell University Press, 1964. von Helversen, 0. Zur spektralen Unterschiedsempfindlichkeit der Honigbiene. Journal of Comparative Physiology, 1972, 80, 439472, Wald, G . Human vision and the spectrum. Science. 1945, 101, 653458. Walls, G . L. The vertebrate eye and its adaptive radiution. Bloomfield Hills, Mich.: Cranbrook Institute of Science, 1942. Walls, G. L., & Heath, G . G. Neutral points in 138 protanopes and deuteranopes. Journal of the Optical Society of America, 1956, 46, 640649. Walls, G . L., & Mathews, R. W. New means of studying color blindness and normal foveal color vision. University of California Publicurions in Psychology, 1952, 7 , 1-172. Walters, R. H., & Parke, R. D. The role of the distance receptors in the development of social responsiveness. In L. P. Lipsitt & C. C. Spiker (Eds.), Advances in child developmenr and behavior (Vol. 2). New York: Academic Press, 1%5. Walton, W. E., Guilford, R. B., & Guilford, J . P. Color preferences of 1279 university students. American Journal of Psychology, 1933, 45, 322-328. Warren, J. M. Perceptual dominance in discrimination learning by monkeys. Journal of Comparative and Physiological Psychology, 1954, 47, 290-292. Watson, J . B. What the nursery has to say about instincts. Pedagogical Seminary, 1925, 32, 293 -327.

1x2

Marc H. Bornstein

Weale, R. A . The aging eye. London: Lewis, 1963. Welch, M. J . Infants’ visual attention to varying degrees of novelty. Child Development, 1974,45, 344-350. Werner, H. The comparative psychology of mental development. New York Harper, 1940. Whorf, B. L. Collected papers on metalinguistics. Washington, D.C.: Foreign Service Institute, 1952. Wickelgren, L. W. Convergence in the human newborn. Journal of Experimental Child Psychology, 1967, 5, 74-85. Williams, H. Congenital colour blindness and its detection in children. Developmental Medicine and Child Neurology, 1975, 17, 24:1-25 I . Williams, L. G. The effects of target specification on objects fixated during visual search. Acta Psychologica, 1967, 27, 355-360. Wooley, M. T. Some experiments on the color perceptions of an infant and their interpretation. Psychological Review, 1909, 16, 363-376. Wooten, B . R . Photopic and scotopic contributions to the human visually evoked cortical potential. Vision Research, 1972, 12, 1647-1660. Wooten, B . R. Infant hue discriminaltion?Science, 1975, 187, 275-277. Wright, A . A. Psychometric and psychophysical hue discrimination functions for the pigeon. Vision Research, 1972, 12, 1447-1464. Wright, A . A,, & Cumming, W. W. Color-naming functions for the pigeon. Journal of the Experimental Analysis of Behavior, 1971, 15, 7-17. Wright, W. D. Researches on normal and defective colour vision. St. Louis: Mosby, 1947. Wright, W. D. The characteristics of tritanopia. Journal ofthe Optical Society ofAmerica, 1952,42, 509-52 1. Wyszecki, G., & Stiles, W. S . Color Science: Concepts and methods, quantitative data and formulas. New York: Wiley, 1967. Yendovitskaya, T. V . , Zinchenko, V. P., & Ruzskaya, A. G. [Development of sensation and perception.] In A . V . Zaporozhets & D. B. Elkonin (Eds.), The psychology of preschool children ( J . Shybut & S . Simon, trans.). Cambridge, Mass.: MIT Press, 1971. (Originally published, 1964.) Zeaman, D., & House, B. J. The role of attention in retardate discrimination 1earning.h N. R. Ellis (Ed.), Handbook of mental deficiency. New York: McGraw-Hill, 1%3.

DEVELOPMENTAL MEMORY THEORIES: BALDWIN AND PIAGET'

Bruce M . Ross and Stephen M . Kerst THE CATHOLIC UNIVERSITY OF AMERICA

I . INTRODUCTION: JUSTIFICATION AND SCOPE

1E4

I1 . J . M . BALDWIN'S THEORY OF MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . MEMORY ASSUMPTIONS AND LOGICAL DISTINCTIONS . . . . . . . . . . . . B . IMITATION AND MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C . ASSIMILATION, ACCOMMODATION. AND THE SCHEMA . . . . . . . . . . . . D . RECOGNITION MEMORY AND AFFECTIVE MEMORY . . . . . . . . . . . . . . .

186 186 188 190

I11 . PIAGET'S THEORY OF MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . SCHEMES . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... C . RECOGNITION MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D . RECONSTRUCTIVE MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. LANGUAGE AND MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F . MEMORY AWARENESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G . AFFECTIVE MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

195 195 198 199 201 204 205 208

IV . COMPARISONS BETWEEN MEMORY THEORIES . . . . . . . . . . . . . . . . . . . . . . . .

210

V . TOPICAL MEMORY RESEARCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . POSSIBLE THEORETICAL CONTACTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . RECOGNITION AND IMITATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C . SCHEMA THEORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D . IMAGERY AND MEMORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E . MEMORY AWARENESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . REFERENCES . . . .

193

212 212 214 217 220 224 226

'The writing of this chapter was supported by funds from the Boys Town Center for the Study of Youth Development of Catholic University. Washington. D.C. 20064 . I83

I84

B r w e M . Ross and Stephen M. Kerst

I. Introduction: Justification and Scope The present chapter deals with the only two theories of memory development that have some claim to completeness. Such a claim is only relative, of course, and it is only in going beyond the piecemeal approach of other theorists and filling in theoretical gaps wi,th ingenious speculative extrapolations that Baldwin and Piaget are able to at least sketch the bolder outlines of a theory of memory development. It should be emphasized that for neither theorist was or is memory the main issue. Baldwin Was concerned with mental development generally and Piaget, as is well known, is genetic epistemologist who has written on memory primarily as it is necessary to do so while developing a theory of the changing structures of intelligence. A prime question must be answered first as to the relevance of Baldwin, a pre-World War I theorist, in any chapter of the late 1970s published under the heading of “advances.” A rationale justifying inclusion is all the more necessary because his approach is primarily that of philosophical psychology in which assertions are not only unbacked by data, but there is often uncertainty as to how evidence about a particular theoretical position could be obtained. Nevertheless, there are several reasons why such a revival is worth consideration even now. One is simply the interest of his ideas and novelty of his approach, and that which we would wish to emphasize, the ability to pose questions of theoretical value for later theorists. Many of his ideas are out of date, some even amusingly antiquated, and many ideas are taken directly from dated views of his contemporaries; but no one until tlhe advent of Piaget has attempted such a thorough developmental synthesis in either an empirical or theoretical vein. Memory is not the main focus of Baldwin‘s theory, but gleaning from his writings the fairly large number of observations he makes about memory does have the advantage of sticking with the more psychological and potentially testable part of Baldwin’s writings. Another attraction is to trace how concepts in some measure adapted by Piaget from Baldwin have been altered, been enriched, or perhaps dropped in the ensuing years-e.g., such important concepts as schema, imitation, and assimilati on. It should be recognized that Baldwin was far from being the only influence on Piaget. Some of the major concepts of Baldwin were also influential with other French psychologists such as Guillaume and Wallon, and somewhat earlier Piaget’s teacher Clapar8.de, so that all the parallels between Baldwin and Piaget are not necessarily direct. [Broughton and Riegel (1976) have suggested that the personal friendship between Baldwin and Claparede encouraged Piaget’s early interest in Baldwin’s theory.] Not discounting possible indirect influences, Piaget himself has often attributed key concepts to Baldwin beginning with originating the concept “genetic psychology” (together with G . Stanley Hall) and being a pioneer in advolzating the concept of internalization of actions as the

Developinetirul Mernoty Theorre;,

185

necessary precursor of thought. Piaget’s borrowings were in more unalloyed form in the 1920s when his theory was in a less developed and systematic form. Nevertheless, quite recently in a 70-page summary of Main Trends in Psychology, Piaget (1970) found space to note that Baldwin was the first to grasp clearly the importance of developmental psychology for social psychology and sociology and also that the notion of the interiorization of social taboos into a “superego” was formulated before Freud by Baldwin (independently also by P. Bovet). Several other ideas more specific to cognitive psychology whose origin Piaget attributes to Baldwin will be noted in the following exposition. There is, to be sure, an imbalance in source material in presenting the views of one of the world’s best known and most prolific psychologists in tandem with those of a shadowy figure from the past whose major works have been out-ofprint until recently. Indeed, the suggestion that Baldwin was perhaps the most significant precursor of Piaget has been more of a rumor than a documented fact. Perhaps we are mistaken, but to our knowledge the present essay is a pioneer effort in comparing the later Piaget with Baldwin, although our attempt is confined to memory theory. Historical assessments of Baldwin have been somewhat disparate but not very favorable. It is possible, however, that a reevaluation may be in the offing as something of a revival of interest in Baldwin’s theories appears to be taking place.’ Boring’s (1950, p. 531) judgment was: “Baldwin’s felicitous literary style, surpassed only by James’, gave a transient vitality to his ideas; but his effect was not permanent.” Roback (1952) mentions that Baldwin was a facile writer but that “often his writings are overcast by a thin mist” (p. 179). Part of the difficulty in obtaining a uniform judgment about Baldwin is that he wrote both popular and highly technical works. This is by way of cautioning that just as James’ stylistic polish did not prevent ambiguity in his writing, so it is also with Baldwin. Thus, some of our interpretations of Baldwin could be given alternate readings and may in places be i n The major focus is not historical but on the ideas about memory of the two theorists with the hope that comparison will illuminate interesting similarities and *We are indebted to John Broughton for calling to our attention a number of as yet unpublished papers by scholars delving into various aspects of the wide gamut of Baldwin’s thought. From the standpoint of developmental psychology and Piagetian theory, Broughton’s ( 1975) dissertation outlining Baldwin’s theory of developmental stages is particularly interesting. 3Baldwin’s cosmopolitan and somewhat colorful life (hy academic standards) has not worked in his favor in transmitting his theories to future generations. Both Boring and Roback are i n agreement in giving the reader gossip where he might have expected some elucidation of Baldwin’s far-fromsimple ideas. As examples: “He probably holds among psychologists the record for close approach to persons of royal blood” (Boring, 1950, p. 547). “Baldwin was the most decorated and most translated American psychologist with the exception of William James, and he apparently set great store on being elected to, or joining, many learned societies abroad” (Roback, 1952, p. 179).

Bruce M . Rossand Stephen M. Kerst

186

differences. We would, in fact, be handicapped in drawing historical conclusions because we stress the theoretical outlook that Piaget has formulated in the last decade, although some earlier works that we reference are also important. We do not attempt to present Piaget’s theory compressed into a summary outline; there are already an ample supply of such digests. Similarly, we also limit ourselves to only that small part of Baldwin’s work that is germane to memory theory. We do not attempt any overall summary and we omit altogether such central ideas as his enthusiastic Darwinism applications. For comparison purposes we make a number of allusions to Bartlett’s memory theory. (There is a closer historical connection than might at first appear, since Baldwin wrote a collaborative article with G. F. Stout, an important early influence.on Bartlett.) In Sections I1 and 111 we present the major features of Baldwin’s and Piaget’s theories of memory. In describing each theory a number of coniparisons are made between them, but Section IV is devoted specifically to comparisons between the two theories. In Section V, several topics that have been considered in the previous sections are reconsidered in terms of current research studies. As active research areas are involved, no conclusive interpretations are given, but our review samples what makes these areas of continuing interest and why thzoretical problems have endured.

11. J. M. Baldwin’s Theory of Memory A.

MEMORY AS,SUMPTIONS AND LOGICAL DISTINCTIONS

Baldwin was never precise as to the exact ages at which he places different memory phenomena-but only that certain functions cannot be performed without the prerequisite functioning h e describes. Nevertheless, it is quite clear that some of the most important problems he defines originate during the first two years of life and subsequently become ramified by the enlargement of an environment that includes a strong social component. Given the reality of reasonably accurate human memory, three broad problems are posed as to how this achievement is possible. These are: (1) What is the nature of the development of memory representation? (2) How is control gained over memory? and (3) What gives memory its persistence and validity? 1. The infant develops hi!; memory through the differential retention of occurrences of external objects that derive lrom the events of his daily life. Initially, Baldwin maintains, the infant’s memory is only of entire happenings; a whole event is apprehended as a single cornplicated mental occurrence or object. (For Baldwin, using an act-psychology terminology, the contents of memory are invariably mental objects.) Memory progress occurs by the breaking up of these large undifferentiated happenings into smaller separable units. To a great extent

Developnien/til Memory Theories

187

the breaking up results from conscious individuation, but the early separation is entirely between items in the external occurrences, not between the child and his thought objects. This latter separation is a later and more sophisticated achievement. It can be noted in this account of representation that priority is given to social interactions over either persons or perceptual events. Baldwin speculates that once memory images have been formed they can yield representations of an external object, even when the object is present to an observer, so that an object can simultaneously be perceived and remembered. By this Baldwin means more than the potential for recognition memory to occur; he believes that it is quite justified to assert that a memory may be fully constituted in the presence of a perceived object so that one can be confused as to whether he is dealing with a remembered or a perceived event. 2 . The most important way in which control is gained over memory is by manipulation of order context. The term order context has a special meaning and is not to be confused with the serial order in which external events gccur. Rather, order context refers to the sequential conversion of mental representations from the initial percept to the memory image, which is the original order of conversion of the constructed memory image. But with a well-formed memory image there can also be some shifting of the memory back toward the original perceptual experience. The child does not actually reconvert his memory image all the way back to a perceptual experience, but there is a feeling of manipulability that is in contrast to the “stubborness” or uncontrollableness of the perceptual experience. This feeling of control or at least of a relation between the reinstated memory image and the original perceptual experience-Baldwin describes it as a kind of “fitting on again” to the original sense modality-gives a special mark to memory experience that sets it off from other experience. Baldwin contrasts his view of an intrinsic, qualitative feeling of a relationship with older theories of a quantitative type that describe memory experience as like perception but of a weaker intensity. Control of memory is classified as a mediate manipulation because it is at one remove from the original experience since actual control is shared between external events and internal discursive operations similar to those performed in thinking. Other types of experience, some of which will be discussed, are immediate; these include experiences of an imaginative, esthetic, and mystical nature. The hallmark of immediate experiencing is that it is usually nonschematic and dominated by affective content. 3. Baldwin tends to treat the persistence and validation of memories together and to consider their attainment as a further extension of the just discussed conversion potentiality of memory. It is only the memory validation process which Baldwin considers in detail, but he implies that memories must be validated in order to persist. Validation of the memory process can occur in three stages, which are labeled physical, social, and psychical stages of confirmation.

I88

Bruce M . Ross and Stephen M . Kerst

The first stage is direct meniiory validation where the original external occurrence or object is reinstated in acluality and confirmed by perceptual experience quite independently of other possible confirming social or mental experiences. The second stage is social; physical confirmation is dismissed in favor of social experience or at least types of subjective experience that are fairly universal and therefore widely understood. The child goes to parents and others to verify the truthfulness of his own memory and imagination. However, there is much mental content, Baldwin maintains, that cannot really be considered to be part of memory. The tendency of the child is to take this content on faith and consider it to be valid until disproved. The third stage, psychical confirmation, includes the conversion of social material or objects into a mental content that is verified by being remembered. The memory image involved is one that cannot itself be validated by second-stage social Confirmation. Third-stage validation depends on revived inner experiences and in this regard is like the second stage, but third-stage confirmation tends to deal with private, singular events, such as earlier memories or contents of imagination, that escape the possibility of physical or social conversion. Baldwin recognizes that memory development and especially memory validation by the child is not always straightforward, and he mentions several difficulties children experience along the path of development. The young child suffers under the handicap that there is no corrective given by the grotesqueness of an experience; absurd qualities do not invalidate images from representing memories, since appropriate reality limits are not yet available. At the same time, memory and imagination may be woven together into a continuous context that Baldwin compares to a lawyer’s contextuation of evidence in the same way that the client without intent to deceive comes to believe that what did not occur is true. Further, the stage of social confirmation and validation is particularly prone to error. As a result the child merges his own experience and that of others to form a larger aggregate in such a way as to illustrate that the child has become “credulous and suggestive to a scandalous degree.” Somewhat later in development, difficulties are encountered in validating memories because the strong motivations that occur lead to a conflict between perceptual and memory experiences, and as a consequence there is some loss of memory control. B

IMITATION AND MEMORY

It is unavoidable in considering Baldwin’s theories not to give special emphasis to the topic of imitation. In many ways it is the key concept for Baldwin, as he gave it wide application in the sociological area as well as the psychological; at the same time, his overextension of the imitation concept led to its being the most generally criticized of his ideas [e.g., Sewny (1945); Sewny also cited J. Dewey in a similar veiin]. In any event, imitation is both a precursor of

memory and a functional mechanism that continues to interact with memory after the capacity for memory has come into existence. The reason memory is not itself the earliest function is bound up with Baldwin’s definition of memory as necessarily a conscious experience, since memory always implies an identification of something past. The earliest form of imitation is what Baldwin designates as the “circular reaction,” a form of imitation characterized by a muscular reaction that reproduces the same stimulating conditions that triggered the reaction in the first place. Thus, a second similar reaction can recur and the circularity can continue to sustain itself through further repetitions. As Baldwin puts it, “the effect of imitation, it is clear, is to make the brain a ‘repeating organ,’ i.e., to secure the repetitions which on all biological theories the organism must have, if it is to develop” (Baldwin, 1920, p. 251). As a simple example, Baldwin suggests a visual stimulus that is reacted to by motor movements that continue to reinstate it. This imitative, self-sustaining type of reaction Baldwin finds at every level in the phylogenetic scale (possibly with an analog even in plants). At the lowest level, pleasure and pain are not necessary for selective circular reactions4; at a higher level, pleasure and pain enter into conscious selective or inhibitory reactions so that new reactions can be made available for repetition and eventually become habits. The circular reaction is the necessary preparation for the memory function. In memory the thing remembered is itself absent; yet as the appropriate movement reaction comes about just the same, Baldwin conceives of this sequence as a more sophisticated form of circular reaction. The aim is similar to that of the nonrepresentational circular reaction-to get in touch with a desirable stimulus or avoid an undesirable one. Imitation cannot be subsumed by memory because memory is more consciously representing. In the circular reaction the sought after object or action, even when successfully reinstated, does not yield a representation because imitation is brought about by action while memory stems directly from perception. Additionally, imitative results are not validated by the three-stage confirmation process previously described for memory. A paradoxical aspect of the circular reaction is that at the very young age at which the circular reaction first appears, Baldwin considers it a conscious experience, though nonmemorial, while at older ages the circular reaction becomes an agent of habit and consciousness lapses. From the standpoint of the child, a fundamental distinction to be acquired at an early age is the necessary dualism between the inner and 4Piaget borrowed Baldwin’s concept of circular reaction and elaborated it into a three-level process that describes some infant behavior during the sensorimotor stage (Piaget, 1952, 1954). The elaboration is not treated in this chapter because Piaget has not emphasized this process in relation to

memory.

Bruce M . Ross and Stephen M . Kerst

190

the outer. This distinction approximates that between subjective and objective, but what is designated inner or outer shows systematic change with development. The child is initially unable to make any inner versus outer distinction; this undifferentiated state Baldwin calls adualism. (Piaget has frequently used this concept, with acknowledgment of its coinage by Baldwin.) Imitation plays an important role in bringing about the first inner versus outer separation. Surprisingly, however, it is failed imitation that is most important, since imitation is most productive of an inner-outer opposition to the extent that imitative actions do not actually yield the expected duplication of external objects. Later, when the child does achieve memory representations, he cannot help but be aware that the individuating context of memory with its emphasis on specificity and singularity is markedly different from the nondifferentiating context of the consciously imitative. Nevertheless, at this early stage what is inner and what is outer is not the same for the child as for the adult. As belonging to the outer, the child lumps together sense perception and what Baldwin calls fulfilled memory, contents of memory that can be validated by at least one of the three confirmation stages. The opposing cla5s of occurrences includes a “sort of unfulfilled pseudo-memory’’ which because it is unconfirmed and ineffectual is grouped with images in general. Thus, this early dichotomy depends on whether a particular mental process is instrumentally satisfying. It is only later that the distinction between inner and outer is revised and all memory including accurate memory is designated as “inner.” At this later point the child can get images of the outer world at will which, though inner, are confirmed satisfactorily as regarding outer meaning also. A major characteristic of Baldwin’s theorizing is that implications for social development are often drawn from cognitive development. In the three-stage memory confirmation process, for example, the child who is often unsure of appropriate perceptual validation comes to infer that others also possess memories, hence a distincbion between inner and outer, and that these memories frequently have validating powers superior to his own. Further, a result of carrying out successful imitation is for the child to advance his concepts of other selves since he finds that other people have inner representations and that the outer may in general become inner. Thus, successful imitation plays a key role in the child’s knowledge of how others can also think. C.

ASSIMILATION, ACCOMMODATION, AND THE SCHEMA

The processes of assimir‘arion and accommodation that bulk large in Piaget’s theory are undoubtedly the most important continuation that Piaget makes of Baldwin’s ideas. These concepts did not originate with Baldwin: “assimilation,” for example, as a more limited psychological process was described by Wundt. For Baldwin, assimilation gives the “general statement of all the forms, nets,

Developmetital Memory Theories

191

modes of grouping, which old elements impose upon the new” (Baldwin, 1920, p. 293). Every act of perception requires assimilation to constitute mental objects from it. There is even, Baldwin acknowledges, a familial resemblance to the older Herbartian conception. More directly for our consideration of memory, Baldwin asserts that assimilation is at the base of recognitions and of those illusions which are but mistaken recognitions. All perception is a case of assimilation, but literal duplication of the same objective content can never be experienced twice in quite the same way because the motor shadings of attention vary each time. Assimilation is also the necessary basis of the earliest form of association. When stimulations are multiple, as with external objects, motor reactions are reduced to orderly habitual discharges to give them a sense of unity or identity. This process of association by assimilation has a motor basis with elements associated together in memory because they are used together in action. Further occasions for association arise when the individual becomes less dependent on particular external objects or events and more capable of remote and substitute stimulation. For Baldwin any two elements connected in consciousness are connected only because they have motor effects in common. These motor effects are on a continuum with, at one extreme, similar clusters of motor effects resulting in signs of sameness that lead to recognition. At the other extreme the looser and less effectively connected clusters brought about by assimilations result i n associations. Accommodation is, like assimilation, also a term with a variety of previous theoretical applications. Baldwin, although he disagrees with their theories, cites in particular H. Spencer and A. Bain. In Baldwin’s broadest usage, accommodation refers simply to individual reactions that are adaptive. What Baldwin describes as a general formula for accommodation is the “principle by which an organism comes to adapt itself to more complex conditions of stimulation by performing more complex functions” (Baldwin, 1920, pp. 454-455). All functions which the individual has learned-speech, handwriting, piano playing, learning to act-fit under this rubric. But at a more basic level accommodation is part of a circular reaction or of acts of imitation. However, if a circular reaction or imitation continues, accommodation no longer plays a part and habit takes over, habit being defined rather generally as the tendency of an organism to continue with greater and greater ease processes which are vitally beneficial. Through habit, one becomes capable of the internal revivals of memory. In summary, Baldwin is fairly definite about what constitutes imitation and assimilation, but such key terms as accommodation and habit are given less precise definitions. There is no indication of an equilibrium model with a balance drawn between assimilation and accommodation as is found in Piaget (195 l), where imitation is characterized by an excess of accommodation. Once a memory representation is possible, accommodation takes on a new role of guiding conscious experiencing, which leads to the attention function and, on

I92

Bruca M . Ross and Stephen M . Kerst

occasion, to volition by means of the coordination of all the motor elements involved. All along, it should be emphasized, Baldwin scrupulously attempts to give a physiological motor or efferent nervous discharge basis to accommodation; it is excessive motor discharge that enables accommodation to occur, and volition is possible because of incipient motions and vibrations of memory copies with other copies remembered or perceived. For Baldwin, unlike Piaget, accommodation is not the complement and opposite of assimilation so much as it is of habit. Although accommodation is essential in forming habits, its reference is to new movements and therefore is a prospective reference, while habits have retrospective reference in that they are concerned with old, already performed movements. The new movements of accommodation also come into conflict with old movements so that the habits of which they are components disintegrate. A further point to make specifically about the accommodative function of attention is that attention vanes according to content, that is, according to type of imagery involved-visual, auditory, motor, etc. It naturally follows that memory also varies according to type of imagery; but Baldwin emphasizes that while this view is commonly held about memory, psychologists have been inconsistent and few relate type of imagery to attention. If an individual or a species has a dominant modality, it will, of course, be reflected in both attention and memory. Although not particularly fond of illustrations drawn from arumal behavior, Baldwin presents the rabbit as an example of an olfactory type with, to be sure, concomitant motor responses. “The constant movement of the tip of the snout in many such animals when exploring for food, etc., by smell, shows the development of delicate smellmotor reflexes analogous to our eye-motor reflexes and the horse’s ear-motor” (Baldwin, 1920, p. 446). Baldwin points out that his concept of the schema can be traced to Kant. The schema is defined by Baldwin as an “experimental” mental object and, rather like the accommodation process, its orientation is on prospective meaning. It is a product of imitation that is lrequently constructed during play. At first a schema usually possesses only a provisional meaning that may subsequently be enlarged or altered. The schema is “held and controlled with the express psychic proviso or reservation that its meaning is yet to be madeup” (Baldwin, 1906, p. 165). In fact the schema is such an indefinite placeholder that a single schema may possess alternative meanings. Baldwin carefully distinguishes a second function of the schema; it is individuating in setting apart a specific mental content, but it accomplishes this function in a way that is different from either perceptual or memory individuation. In any event the meanings of a schema may be transitory because they are merely instrumental. The schema, then, is a product of the way the child goes about his pragmatic concerns and deals with the hypothetical and with the future. As one might expect, such a necessary structure appears at a very early age, but schemas that occur in infancy are not yet adequate logically and can later be supplanted by general meanings.

Developmentul Memory Theories

I93

The conclusion to be drawn is that Baldwin’s schema has surprisingly little to do with memory except in the negative sense that he indicates schemas are characterized by greatly loosened memory control as compared with other types of activity. Baldwin does mention that there may be habitual schemas and even assimilation of new elements to old habitual schemas, but this idea is (in contrast to Piaget) not followed up. At a young age, past experience can sometimes be taken as literally representing future experience so that the child’s literal memory is itself the schematic basis of the future, but such a naive view does not long persist. All in all, in regard to the application of schemas to memory, Baldwin’s position contrasts sharply with Piaget’s and is almost opposite to Bartlett’s (1932). D. RECOGNITION MEMORY AND AFFECTIVE MEMORY

The reader of Baldwin will find recognition paired under a variety of headings that are much broader in range than the topic of recognition memory by itself. But behind his discussion of recognition and assimilation, and attention, and meaning, and occasionally even a direct discussion of recognition memory, there is a consistent point of view. Recognition for Baldwin is a process broader than memory since it is also an important part of the thought process in that it confirms the individuation of mental objects available through perception as well as memory. But recognition is dependent on elements of content brought together by assimilation so that there is a common core of content at each occurrence of an event or object, even though every element is not identical at each occurrence. Recognition makes events detachable, or as Baldwin phrases it “liftable,” and thereby also establishes contextual meaning, although this does not preclude the possibility of committing constant errors. Baldwin distinguishes three types of recognition of progressively increasing difficulty. The developmentally earliest type is absolute recognition in which there is only a sense of familiarity without recognition of additional relational features. This type of recognition, performed also by lower animals, can occur without memory images and is thus close to the sphere of primary attention. A second type of recognition is class recognition with the focus on the class of objects the individual item represents; Baldwin equates one form of conceptual thought with class recognition. The third type is individuating recognition with the focus on delineating item singularity. Additionally, Baldwin believes-even as some current research demonstrates-that under favorable circumstances a content can be reproduced by an individual without recognition occurring. With respect to memory error, not only dkjb vu but even simple failure of memory are both classed as memory illusions. Baldwin also forms three categories when he classifies types of recognition in regard to the persistence of objects or events. “Present sameness” refers to the recognized sameness of present objects, whether presence is continuous or inter-

194

Bruce M . Ross and Stephen M . Kerst

rupted. “Remote sameness” occurs when an image is the same as a remote or absent object, although present sameness may be included as a component. “Recurrent sameness” occurs when there is successive recurrence, as in the repeated tickings of a watch declared to be the “same” sound. Presumably, in this last example there are a series of remote sameness fulfillments which are substituted for verified recognitions. Baldwin gives an unusually detailed example (worthy of a philosopher of ordinary language) of the distinctions between remote and recurrent sameness. The distinctions deal with the analysis of the proposition: “I remember this pen-it is mine” as compared to “I remember my p e w i t is this,” (Baldwin, 1906, p. 155) which Baldwin reduces to the formulas: “this is the same as it” versus “it is the same as this.” The first example, depending on circumstances, can be a case of either present or remote sameness, but the second example, since it involves an actually detached and recognized memory content apart from the object presented, is definitely an example of recurrent sameness, or as Baldwin alternatively calls it “sameness as recurrence.” However, even knowledge of persistence gained in this way is not yet fully adequate to give the judgment of identity, which Baldwin places at a still higher level of understanding. It is clear that, in general, because recognition is intrinsic to the development of meaning it is far from being simply the most passive form of memory. Recognized objects can be not only external sense objects but also factual propositions and memory images; i n the full sense of the term, recognition occurs only with cognitive objects, not with conative or affective functions. That recognition at a level above familiarity is attended by complexities is well illustrated by Baldwin’s treatment of absolute pitch recognition. He considers both auditory and visual imagery among musicians and, characteristically, concludes that absolute recognition is possible owing to the revival of motor associates of former acts of attention. Another characteristic Baldwin attribution is that one’s sense of self is implicated in the recognition process since recognition also bears with it a “feeling of ‘warmth,’ ownership, self-reference” (Baldwin, 1920, p. 301) because “just this motor element it is that carries along with it the habitual attention strains, and these attention strains are in large part the stable, ‘identical’ element in the sense of self” (Baldwin, 1920, p. 301). Baldwin places strong emphasis on affective memory. Although he acknowledges indebtedness in this area to contemporary French psychologists, particularly Ribot, and to the empathy theory of the German psychologist Lipps, the systematic framework into which affective memory is cast is clearly Baldwin’s. For him affective states have their own revival and recognition but, as just intimated, recognition is a somewhat restricted process as compared with cognitive recognition. This is so, in part, because there is no second-stage social confirmation or third-stage psychic confirmation in which one affect could be substituted for another. But an even more basic difference is that cognition and affection are understood differently, in that cognition must be preceded by

Developmentcil Meinon1 Theories

I95

generalization while affect has the immediacy of a singular and unique experience. Thus, separation between the cognitive and affective is fairly complete. although obviously to some extent the mechanisms of thought are available for casting affective experiences into cognitive descriptions. Following Ribot, Baldwin espouses an explicit affective logic that is largely defined by its differences from traditional logic; there is no principle of excluded middle or contradiction, no universal logical implication and deduction, and “syllogistic processes are very undeveloped in the domain of feeling” (Baldwin, 1911, p. 131). Some veridical communication about affective matters is possible, however, because of the Occurrence of afective generalization. By affective generalization is meant a wholistic, patterneci organization built up from detached and isolated motor tendencies. It is these patterned affective organizations which are attributed to others by a sort of primitive projection (Baldwin’s term is “ejection”). Thus, though the intimate qualities of affective and conative experiences remain personal and unshared, the general forms of hope or fear, impulse or desire, can be meaningfully attributed to others. Since Baldwin places great emphasis on esthetic experience, it is not surprising that affective memory is emphasized in an esthetic context. The development of esthetic experience is a mature achievement since “in aesthetic contemplation we have the fullest revelation of what reality means” (Baldwin, 191 1, p. 257). In large measure such a revelation can be accomplished because esthetic values are assumed to reconcile apparent epistemological dualisms that puzzle the individual intellectually when dealing with merely cognitive relations. Works of art, according to Baldwin, have their distinctive effects because they are expressive; thus, their chief impact is made by affective representation and revival. Baldwin postulates a progressive continuity between memory for isolated and unorganized affects and memory for emotional experience in the service of esthetic ends. He asserts that the build-up of emotions follows the form of the medium; for music with its discursive form, emotion moves from part to whole, while pictorial form produces emotional movement from whole to parts. The role of memory in emotional experience is somewhat simpler in responses to music than it is in the visual arts. In contrast to music the order of presentation is uncertain in the visual arts and actual objects are frequently given representations that require cognitive interpretations. But in spite of his extended treatment of esthetic experience, Baldwin does not directly trace the developmental progression involved.

111. Piaget’s Theory of Memory A.

OVERVIEW

Piaget’s memory theory has been presented in considerable detail with copious supportive data in two full-length books on imagery and memory (Piaget Sr

I96

Bruce M . Ross and Stephen M . Kerst

Inhelder, 1971, 1973). One important claim of the theory has been particularly followed up: That long-term memory performance for a specific cbject display improves over a period of months or even more than a year. Measurable retention improvement is possible because the child has acquired higher level schemes that enable him to restructure his original memory impressions in a more organized way. Short theoretical expositions particularly emphasizing this aspect of the theory appear in Piaget (1968) and Inhelder (1969). In two comprehensive reviews Liben (1976, 1977) has summarized the empirical studies bearing on this claim and has discussed the status of the developmental coding problem from a theoretical and methodological standpoint. Therefore, this very active topic will be largely excluded from our discussion. In a recent review (Ross & Furth, in press) Piaget’s memory theory was briefly outlined in terms of definitional criteria and a sixfold dichotomous classification. This classification can be repeated here to locate Piaget’s description of memory functioning within the framework of Piaget’s cognitive theory. What is described by thefirst term of each dichotomy is what Piaget identifies as “memory in the narrow sense”-the traditional psychological problem of describing recall and recognition of events or objects encountered in the past. The classification dichotomies for conservation of the past in terms of this kind of memory are: (1) individually acquired versus hereditarily transmitted; (2) cognitive-behavioral versus physiological-somatic; (3) accommodative-figurative versus assimilative-operative; (4) objective-symbolic versus undifferentiated-sensorimotor; (5) singular versus generalizable; and (6) reference to the past versus reference to the present or future. A further clarifying distinction is to point out what Piaget means by “memory in the wide sense.” This is simply the conservation of knowledge schemes, in Piaget’s sense of “schemes.” Such schemes embody generalized knowledge such as practical actions and logical and mathematical concepts rather than reference to a specific past occurrence; therefore, they are part of intelligence and so not subject to forgetting in the ordinary course of affairs. While a scheme is the structural generalization and therefore “abstract,” its actual manifestation by figurative instruments-perceptual, imitative, or imagic-can be embodied in alternative manifestations or schemas (e.g., Stacey & Ross, 1975). Characteristic of Piaget’s emphasis on development is the classification of types of mnemonic performance that first become functional at different developmental levels. Recognition has a wide range beginning at the level of a simple extension of reflex action, e.g., regaining again a moving object that an infant has tracked with his eyes and momentarily lost sight of. Next is recognition of signs as signifiers in the habits and acts of the sensorimotor intelligence where the sign, say a word or gesture, is linked to a sensorimotor scheme. At a higher level recognition is tied to classification. This gives the possibility of definite nonrecognition as well as positive recognition where a specific class entry is

Developmental Memory Theories

197

being sought. But there may be cases where “recognition” blends into judging and highly schematized comparisons so that any specific memory component is absent. Piaget and Inhelder give the example of “recognizing” Vivaldi’s style i n an unknown musical score as such a nonmemorial recognition. However, there are all sorts of mixtures of recognition and judgment that are possible with both schematizing identification and individualizing identification (reminiscent of Baldwin’s recognition categories). Obviausly these latter types of recognition appear only after reconstruction and recall are already well-practiced forms of memory. After the possibility of the child’s performing recognition but before he can manage recall, Piaget posits the intermediate-level performance of memory by reconstruction. There are four types of reconstruction, beginning with sensorimotor imitation-imitation of others, of oneself, or even of an object. Reconstruction is a kind of recall by actions, which must be internalized in images before recall through representations can come about. At a higher level is the reconstruction of the result obtained by an action i n which the activity has not been embodied by a sensorimotor scheme. At an easier level this occurs where copying in the presence of the object has taken place prior to the child’s attempt at reconstruction and at a more difficult level where reconstruction occurs without prior attempts at characterizing. Finally, a higher type of reconstruction is that performed by means of schematized actions such as transitive sequences. But here distinctions cannot be cut too fine; for if a clear identification is not made with a specific occurrence in the past, results can be attributed to a generalization from the scheme without the necessity for the retention of actual situational components. An example can occur with a seriated series of sticks of different lengths; if lengths are so similar that seriation has to be carried out by pairwise comparisons, this is the enactment of an operational scheme; but if particular sticks or stick lengths are remembered and placed in order, then reconstructive memory has played an important role. Particularly because verbal recall is avoided, those grades of mnemonic recall that are developmentally later than the beginnings of recognition and reconstruction are easily viewed as extensions of actions and related to action schemes; thus there is direct continuity between reconstruction and recall. At the lowest level is the memory image of a schematized action as found, for example, in the rotation of an object, where in one experiment memory was found to be as good for recall as for reconstruction. At a higher level is the recall by images of actions that are less regular and not easily schematized. At the highest level and cited as the purest example of recall is that kind of recall where objects or events are unrelated to actions in any direct schematizable way, though schemes will always play some role in retention, whether they are piecemeal or even largely in error. The above listing of different examples of memory-processing situations for recognition, reconstruction, and recall is not meant to be an all-encompassing

198

Bruce M . Ross and Stephen M . Kersr

inventory but illustrates a sequence of difficulty in experiments for which data have been obtained. The general notion is maintained that some recognitive memory appears before any reconstructive memory, and that in turn reconstructive memory appears before the earliest type of recali. However, schemes appear and are utilized with all types of memory. This leads to a consideration of the wide variety of organized mental structures that Piaget refers to as “schemes.” B . SCHEMES

The multiplicity of schemes is one of the most confusing aspects of Piaget’s theory. Nowhere is this truer than in his description of memory phenomena. Not only are there sensorimotor and operative schemes, but also there are assimilatory and directly mnemonic schemes (Piaget’s characterization of Bartlett’s (1932) “schemata”). Piaget has also written of presentative schemes-these are representational schemes or concepts including some at the sensorimotor level-and procedural schemes-actions that are primarily means to a goal. Memory schemes are characteristically different from schemes applied elsewhere, however, in being highly differentiated and limited in scope. Developmentally, earlier schemes are sensorimotor schemes beginning with those of a perceptual nature. Repetition appears to play an especially important role in maintaining these schemes, which are sometimes characterized as “habits”-the only place in Piaget’s theory where habit is used as a systematic term. Intermediate between the habitual and operative schemes are preoperative schemes of the presentative type that are only tending toward reversibility. It is apparent that scheme designations do not refer to mutually exclusive classifications. The reason that scheme multiplicity is particularly a problem in working out memory mechanisms is that Piaget’s special empirical emphasis is on long-term memory, where the schemes available to a child can change over the period of retention. Thus, earlier schemes disappear, new schemes appear, and deve!opmentally earlier schemes combine to form more comprehensive later ones. (In contrast, the schema-the regulatcry mechanism governing :he mental or graphic expression of the image-develops only slightly with age.) Piaget emphasizes, however, that many of these distinctions cannot be made at the sensorimotor level where schemes and types of activity often remain largely undifferentiated. Piaget writes of the maintenance of schemes through exercising them. That is, it appears that some early schemes embodying immature activities at the preoperative level are not supplanted by more adequate ones but simply wither away. This withering away does not occur with later, operatively adequate schemes, which are conserved by their own generalizing and reproductive functioning. Although this characterization might seem to be a tautology-a scheme remembers itself-what is meant is that the logical consistency or form-quality of a scheme is preserved on an all-or-none basis; for example, there is no possibility

Developinentol Memory Theories

I99

of part retention for the scheme of seriation. Once acquired, such a scheme would ordinarily be retained unimpaired through adulthood until senility or other organic deficit. It is emphasized that only the mature operative schemes have a claim to such relative immortality. Nevertheless, Piaget asserts that apart from this automatic self-conservation, the maintenance of schemes cannot be ful!y explained on a functional psychological basis. As schemes are axiomatic for the system one would have to leave the behavioral system and find a neurobiological and biochemical basis to account for the memory permanence of knowledge. Thus, Piaget takes a surprising interest i n physiological work with RNA and interprets this work as favoring the short-term and long-term memory division, with the physiological organization of long-term memory more or less permanently structured but in a distributed system. It is apparent that the scheme idea i n Piaget’s formulation is rather different from its earlier formulation by Raldwin. The Piagetian scheme has a structural permanence that is not found in Baldwin’s prospective, impermanent, “semblant” schema with its provisional experimental nature. An illustration of this difference is that Piaget provides evidence that among younger children there can be conflict between incompatible, incorrect schemes with consequent memory distortion. Presumably in Baldwin’s system competition among his schemas would reduce to not much more than alternative hypotheses. Nevertheless, as has been noted, Piaget’s schemes as the major structural features of knowledge possess the possibility of almost every shade of permanence, and there are schemes of almost every degree of generality. In particular, wide-ranging schemes of great generality are brought about by a combining of logically related lower level schemes into what Piaget terms “hierarchical col1igation”-where one explanation binds together several lower-order schemes. At the other extreme there can be specific single-term schemes of identification such as are frequently invoked in recognition memory. C. RECOGNITION MEMORY

Piaget makes the special point that recognition as well as other forms of memory necessarily depend on schemes. The perceptive or sensorimotor schemes underlying recognition are different from the representative schemes underlying recall, but a given figural memory may be based on one of several schemes. Where Piaget differs from other theorists is that recognition, reconstruction, and recall are not simply methods for studying memory but, as outlined above, are tied into a rough developmental stage sequence. The schemes appropriate to memory recognition that have been subjected to experimental examination by Piaget are those that deal with spatial operations or what Piaget terms “infra-logical” operations. Such operations are involved in the perception of spatial partitions and spatial proximities in a manner reminiscent of the way in

200

Bruce

M .Ross

and Stephen M . Kersr

which configural features are described in the Gestalt laws. But, in spite of apparent similarities, Piaget is not a Gestalter and is utterly opposed to the Gestalt “good figure” concept, particularly as a spontaneous tendency in memory retention. For Piaget no mental organizational factors can have the automatic, nativist cast of the Gestall laws. Rather, Piaget emphasizes the overt, constructivist activity of eye movements in children’s performance of recognition, although he himself has not reported eye-movement data. As in Gestalt psychology, the Piagetian-inspired memory program is thus far almost exclusively limited to visual phenomena, although some extensions have been suggested that will be mentioned shortly. The schemes as the chief structures of intelligence are formed by assimilation, while the individualizing of entities necessary for memory is accomplished by accommodations which do nor create any schemes. Nonetheless, Piaget periodically reminds the reader that assimilation and accommodation never occur separately but any mental activity necessarily entails both. Perceptual schemes, for example, are assimilating insofar as they represent the general tendency to identify and compare, but they are also accommodative insofar as they encourage explanation of the perceived configuration. With this point of view, apart from other considerations, it is inevitable that memory is conceptualized as a specialized function of intelligence rather than what theorists using the information-processing approach term a “decoupled” function. It is, of course, in the emphasis on accommodation that the differentiation of memory from other functions resides. But there are marked differences in accommodative potential for each type of retention. Accommodation is greatest for recognition since fine figurative distinctions can be made with the original model reinstated. The presence of the object to be identified allows schemes to be more mobile and differentiated than with other types of memory. At a very young age imitative accommodation precedes figurative accommodation as Piaget, unlike Baldwin, does not treat imitation as a completely separate and unique explanatory function but incorporates it among the activities that encourage good retention. The ability to perform deferred imitation during the sensorimotor stage is of greatest importance. For Piaget, active participation with materials or events favors retention even when the activity is other than reconstruction. Piaget’s view of the mechanisms of recognition is also in contrast to that given by Baldwin, who classes recognition primarily as an assimilatory phenomenon. In part, this disagreement appears to arise because habit plays the same role of a stable intellectual structure for Baldwin that the scheme plays for Piaget. For Baldwin, assimilation is primarily a gateway mechanism for identification and classification, and the interacting schemas are more like transitory hypotheses. Thus, assimilation has for Baldwin many of the recognitive functions attributed to accommodation by Piaget. Nevertheless, as accommodation for Baldwin is the prime means of breaking up habits and reacting with novel actions to novel

Developmentrrl Memory Theories

20 I

elements in the environment, Baldwin’s accommodation serves a number of the same functions as it does for Piaget, but Baldwin’s description is cast in terms of motor consciousness and subjective attention. D. RECONSTRUCTIVE MEMORY

In the preface of their large book on memory, and elsewhere, Piaget and Inhelder (1973) stress the fundamental importance of the reconstructive memory as a special type of memory fitting between the elementary mnemonic level of simple recognition and the higher mnemonic level of recall. This conclusion is listed among only a handful of major conclusions. Another indication of the importance attributed to reconstructive memory is that in the description of memory-processing situations given in Section III,A, four different types of reconstructive memory were noted. Psychologists often gloss over Piaget’s statements about reconstructive memory because, unlike the finding of improvement of long-term memory, these statements completely agree with what one would expect: recognition memory is ordinarily better than reconstructive memory, which in turn is usually better than recall (or what Piaget terms evocative memory when the recall of nonverbal items is brought about). Offhand, one might attribute the superiority of reconstruction over recall to the memory methods involved; reconstruction could be construed simply as a type of aided memory superior to unaided recall. Piapet does not see it this way and uses a number of theoretical arguments to justify the claim that reconstructive memory is a distinct type of memory. (Note that there is a possibility here for terminological confusion. The claim that reconstruction is a distinct memory form is quite separate from the general characterization of Piaget’s theory of memory as “reconstructive.” In this regard, Piaget himself cites P. Janet as a complete reconstructionist in believing that all memories are invented, in contrast to Freud and Bergson, for whom everything in memory is conserved and consequently there is no need for reconstruction. Piaget believes he is taking a middle position between these two extremes.) To return to the theoretical arguments concerning the importance of reconstructive memory as a specific memory type, arguments supporting this position are provocative even if one is not overwhelmed by the experimental results alleged to prove its uniqueness. Underlying schemes are obviously identical in reconstruction and recall, but the accommodation of these schemes is greatly facilitated in reconstruction. The facilitation is greater because the child manipulates material much more readily at a young age than he does drawings or thought images. Thus, the accommodations that occur to his action schemes in attempting reconstructions are those he performs most easily. Additionally, when the items to be retained contain arbitrary elements that are not encompassed by schemes, retention of these elements is also facilitated by active manipulation.

202

Bruce M . Ross and Stephen M . Kersr

While recognition is rooted in perception, reconstruction is an outgrowth of imitation, which is a precursor of the internalized imitation that leads to image formation. Images are then the vehicle for recall when the original item is absent. Most perceptions, imitations, and images are not primarily memorial, but they are used as figurative instruments in the service of the memory function. There is also a developmental reason given for the superiority of reconstruction over recall. Reconstructions based on actions repeat the genetic order of acquisition from actions to schemes and then to memory images which, according to Piaget, facilitates retention. In contrast, recall works in the opposite sequence with the relative handicap of the weaker tendency of the memory image to maintain itself and trigger supporting schemes. A succession of stages for different memory types is broached, and is said to correspond approximately to the well-known stages of intellectual development. Early recognitions are found during the sensorimotor stage; reconstructions mark the transition from the sensorimotor to the representative stage (at about age 2 years); and recall is given a wide range of action over the representative, preoperative, and operative stages of intelligence. Further, when an object or design is complex enough that several schemes are simultaneously involved, the reconstructive performance of the young child can be better than recognition if the child cannot coordinate his i nfralogical schemes sufficiently to produce comprehension. The attempt has not really been made to support these stage divisions empirically, in that memory experiments have usually been conducted with children ages 4-15, and age 4 follows the initial appearance of the different memory types. With the children used in these experiments, the peak age (not the age of first appearance) for good reconstructive performance as compared with recognition and recall appears to be in the age range 5-8. But overall memory maturity depends on the acquisition of more schemes and the surpassing of reconstructive memory by performing accurate recall. Although imitation is especially important for reconstruction and in its internalized form for recall, Piaget is much more restrictive than Baldwin in his applications of imitation. One reason is that Piaget asserts in his book on imagery that the child imitates only what he understands or is well on the way to understanding. Even so, imitation has an extremely important place, not merely to produce copies, but as a liberating force because in its deferred and internalized forms, imitation is the means for developing beyond the initial sensorimotor stage by producing differentiated figurative and symbolic signifiersimages, symbolic games, etc. These signifiers are the beginning of Piaget's communicative semiotic function, a category broader than language. Piaget emphasizes, in fact, that language is not able to particularize detail adequately for memory identification, and c~onsequentlyimages have a necessary mnemonic role to play. It is interesting in this regard that Bartlett (1932) believes that words

Developmental Memory Theories

203

tend to supplant images because the image is apt to go further than is biologically useful in the individuation of situations. It should be mentioned that as with other systematic terms, Piaget is somewhat vague as to the boundaries of imitation. It is viewed as an extension of accommodation (in the same way that play is an extension of assimilation), but it is also active during exploratory phases of perception in tracing the contours of objects by eye movements. In this form imitation can lead to actions and gestures shaped somewhat by external models. This type of crude imitation is the basis for subsequent deferred and interiorized imitation which is, as described above, the source of both the mental image and the semiotic function; the latter originates when for the first time the child can make a clear separation between the symbolic signifier and the referent or significate. In all these uses it is a little difficult to believe that the child is always acting with some understanding, but perhaps it is intention “in the wide sense,” since the higher apes are also said by Piaget to imitate only what they are well on the way to understanding. A reconstructive problem of broad import that once again is attracting some theoretical interest is the retention of the temporal order of past events. This problem can be studied both naturalistically and in the laboratory. Bartlett (1932) pointed out that J. Ward and other early British functionalists had unsuccessfully tried to account for accomplishment of this activity by postulating a principle of successive rather than contiguous association. Bartlett himself took a different tack and put temporal order as an organizational principle within an individual’s retained schema. The retrieval problem then became how one is able to extricate an isolated item from this rigid sequential context. The answer was the puzzling process of “turning round upon one’s own schemata” that, according to Bartlett, “gives consciousness its preeminent function.” Anderson and Bower (1974, pp. 59-60) are but the latest in a long line of writers to point out the obscurantism implied in this mental calisthenic. For Piaget the problem of retention of temporal order is interesting because temporal order as such is not part of the schemes. Therefore, the individual must often make use of multiple inferences drawn from highly differentiated auxiliary schemes in order to reinstate the temporal order. The parallel can be drawn between spatial and temporal seriations, but while the former is concerned with schemes for an overall system and its transformations, the latter requires fairly exact spatiotemporal localizations. Piaget’s problem is the opposite of Bartlett’s: to construct the order of events rather than to explain the accuracy of retrieving single items. But if a situation that binds several schemes together is meaningless to a child, there will be systematic difficulties in temporal ordination. Piaget and Inhelder (1973, Chap. 13) illustrate this type of error when children attempt to remember causal processes that are incomprehensible to them. To some extent Piaget and Inhelder, like Baldwin, argue that localization in time is essentially

204

Bruce M . Ross and Stephen M . Kerst

attributable to context and the present use for which the localization is intended, but judgments so reached can have a highly variable amount of cooperation between memory and intelligence functions. In effect, Piaget’s theory, like the previous ones, does not go very far as an explanation, but it stresses the complexity rather than the inevitableness of accurate temporal ordering. In earlier experimental work (Piaget, 1969) it was maintained that seriation is so difficult that children below the ages of 7-8 fail in the task of seriating pictures, and children below the age of 8. using pictures fail to construct a story that is chronologically accurate. E. LANGUAGE AND MEMORY

Since Piaget has set himself the task of describing the relation of memory to intelligence rather than memory i n its own right, it is not surprising that he performs no verbal memory experiments at all. Such experiments would have little bearing on his theoretical interests. Language rather than being a formant of thought is simply one of the instruments of the semiotic function, as are images, gestures, symbolic play, and the like. But language, as comprising signs that in themselves are not symbolic when considered in isolation from mental constructions, differs from these other instruments in being, by and large, nonfigurative. However, Piaget asserts that this is not strictly the case for natural languages. Young children often use language as a set of symbols rather than as as a set of signs, and adults use figurative speech such as metaphors. At the onset of language in the infant about the age of 2 language is close to imitation. However, the general developmental course of language after this time shows a steady shift away from dependence on representation by figural components. The difficulty of conceptualizing these distinctions and giving them practical import in memory experimentation is illustrated by Piaget’s classification of the spoken word itself as not figurative, while the memory or acoustic image of the word is. Little wonder, then, that insofar as possible memory experiments with speech and language are avoided. Language is recognized as a continually developing, shifting function, simultaneously classifiable in different ways but not basic to the structures of intelligence. Piaget emphasizes that the logic of coordination of actions is at a more basic level than the logic embodied in language, which appears in its full propositional form somewhat late in development. Piaget has remarked that his role, as compared with other theorists, seems to be to emphasize the insufficiencies of language. One way i n which he makes this emphasis was mentioned in the last section: nonverbal images are often much better for precise memory specification than language, which tends toward generalization. Piaget’s view of aphasia is consistent with his theory. The loss of memory in aphasia impedes the use of language, but it does not basically affect the operative schemes. The same is true for other speech problems that reflect

Devehprtlentd Mernor:v Theories

205

figurative defects such as poor spatial representation. All in all, it is difficult to see from a Piagetian perspective how aphasia and language problems-studied developmentally or otherwise-would ever have any crucial theoretical significance. It was quite otherwise in Baldwin’s theory. In his third and higher type of imitation, which he designated “persistent imitation,” he included functions that at first look might appear to be the opposite of imitation such as speaking, writing, reading, and performing music. Each function has its coordinate disorder-aphasia, agraphia, alexia, and amusia-that represents for Baldwin a breakdown in volition and attention, borrowing from P. Janet’s conception of functional synthesis. For Baldwin, of course, there was practically no limit to what could be classed as imitation, including what would strike others as highly autonomous such as imitating what one asks oneself to do. With higher and more complex memories it appears to remain an open question whether language is a luxury or a necessity for Piaget. He writes of verbal syntactic constraints put on memory and at an operatively higher level of deductive or verbal reconstructions for retention of causal sequences, concepts, and operations. Presumably, even if language is less than essential for intelligence, it may be necessary for the formulation and hence retention of some types of material which require formal operative schemes. At the least, an increase in language mastery would seem to work a great economy in the memory code. It is a little ironic that with his deemphasis of language Piaget likes to talk about memory coding in terms of technical information theory with its strong communicative emphasis; mnemonic progress at all levels, Piaget asserts, goes hand in hand with a tendency to retain the maximum amount of data with the minimum amount of information. It is hard to put much credence in this formulation when data from memory experiments with language are totally lacking. An interesting observation, however, is that redundancy in the memory code depends on the nature of the code, and what is redundant at a developmentally more advanced level may not be so at an earlier level with younger children who are using a more inefficient code. F. MEMORY AWARENESS

Introspections about memory are, according to Piaget, of no value in understanding the memory process. The tendency is, in fact, for awareness to be misleading, since in recall, at least, one has some awareness of images but not of the basic structural schemes. The argument is not made because children are considered to be less competent than adults but because, in principle, individuals can be aware only of the contents, hence end products, of thought and not of the thought structures themselves. Since, as outlined above, memories always need the supporting structures of schemes, subjective introspections of both memories and nonmemorial images are of no psychological explanatory value. Nor does

206

B r u w M . Ross and Stephen M . Kersr

subconscious representation differ from that in consciousness except that in subconscious representation there is a lack of sufficient accommodation to the cognitive or affective scheme. Piaget also quotes Claparede approvingly that awareness occurs when there is a disadaptation since there is no reason for conscious analysis of behavior when it proceeds smoothly. (To Americans, this viewpoint is reminiscent of J . Dewey’s “felt problem” as the occasion necessitating the thinking process.) Outside authority is not needed, however, to justify Piaget’s views on consciousness. Awareness consists of a reconstruction on an upper level such as consciousness of that which is already organized in a different manner on a lower level. Piaget’s collaborators have produced some demonstrations (Piaget, 1973) in this regard as to how inadequate one’s description is of his own motor behavior, such as crawling on all fours. But, as every golfer knows, accurate description of motor behavior is notoriously inadequate and unreliable, and in fact, it was shown in Piaget’s demonstration that many adults also had difficulty in describing simple crawling behavior. Although numerous overt actions early become internalized, it is a gap in Piaget’s description of intelligence that there is no attempt to describe how overt sensorimotor behavior develops at more mature operative levels; how it relates to consciousness may be still more complex. To illustrate how misleading early childhood memories can be, Piaget has recounted a false childhood memory of his own in which he remembered being the object of a kidnapping. He deduces that this memory must be false according to his theory since schemes for recall are not available at the age at which a baby is in its carriage. An infant at this age could only perform recognitions. Nevertheless, like other experimenters, Piaget is not averse to quoting subjects in his memory experiments when their verbal reports illustrate a point he wishes to make. Thus, in an experiment on remembering a partly irregular geometrical configuration (Piaget & Inhelder, 1973, Chap. 20) Piaget makes the point that memory organization is roughly isomorphous with the process of operative schematization by quoting a 6-year-old who says, “I see it all in my mind,” while a 12-year-old says, “ I remember it all by logic.” These quotations also illustrate that as memory codes change, what is purely a function of memory becomes harder to define. The tradition of the experimental laboratory has been to make sure that memory is involved at all costs by administering meaningless material. Thus, the greatest stimulant for the revival of “visual memory” studies following World War I1 was the invention of the “nonsense” shape. In contrast, Piaget like Bartlett favors at least partially meaningful material. As a developmental psyclhologist Piaget tries to bring two different domains together, the naturalistic functions of memory as they occur in the mental economy of the child, hence some of the considerations in this section, and laboratory demonstrations of cognitive memory for visual materials. Naturally, such laboratory experiments give only a limited and specialized sample of potential en-

Developmental Memory Theories

207

vironmental encounters. Part of the difficulty is gotten around by the care with which Piaget frames definitions-hence memory in both the wide and narrow senses and the exclusion of purely biological memory. At the same time Piaget does not intend to account for the retention of everything, for there is a sense in which every word is a content of memory and every image is a memory image. As previously outlined, his concern is not with every possible content of retention but only with those which relate to a specific past object or event. In places Piaget draws this distinction by designating the latter as yielding reproductive images and the former simply as producing memory images which he contends are implicitly based on perceptions of which the images are partial copies or imitations. They would come close to being pure figural representations not tied to schemes; after their initial retention such isolated fragments show little development with age. Given the experimental evidence that the memory code changes with the addition of higher level schemes to the intelligence-sometimes in the direction of error-and that the schemes themselves alter and enlarge, what are the implications for memory? One logical possibility is that the qualitative form in which memory contents will be revealed depends completely upon the conditions present at the time of testing or spontaneous recall. Such a nonstructural view Piaget rejects out of hand. Instead, he favors a view of continuous reorganization both of one’s memories and of one’s ideas of the past; the viewpoint and significance of a common core of past contents are altered by current knowledge and also by current environmental exigencies. It would seem to follow, therefore, that awareness of the “moving target” of memory would be quite limited. But, perhaps as a consequence of the occasion of addressing a psychoanalytic society, Piaget (1973, pp. 3640) has camed this reasoning a bit further by talking about the mechanism of “cognitive repression” analogous to “affective repression.” In carrying out cognitive repression a child unconsciously represses a sensorimotor scheme by dismissing it from awareness because it comes into conflict with securely integrated, consciously held schemes. Such repressions occur frequently because they are held to be more general than awareness of action. (There would seem to be interesting implications here for athletic coaching, one of the few branches of pedagogy that has thus far remained free of Piagetian “applications. ’7 A constantly reorganizing memory could be viewed as bearing a considerable similarity to the Gestalt theory of the “dynamic memory trace” in which both perception and conceptual materials in the course of retention move toward symmetry, balance, and “good figure. ” Piaget rejects the comparison forcefully. In fact from the failure of experimental support for the Gestalt hypothesis he draws the conclusion, consistent with his own views, that memory images are not identical with elementary perceptual structures and it therefore follows that memory schemes must be assimilatory.

208

Bruce M . Ross and Stephen

M . Kerst

While Baldwin stressed the importance of memory in developing one’s concept of self through self-awareness, Piaget at times considers the self almost what earlier psychologists would call a memory illusion. Thus, self-identity is defined as nothing other than continuity within the constant restructuring of memory that is taking place. But it is an invalid argument, Piaget asserts, to invoke this “so-called identity” to prove: that memory mechanisms remain the same while the intelligence continues to develop. Likewise, Piaget argues that though “Iness” identification on the part of someone probing his memory may seem to pertain to the figurative aspect of schemes, this conclusion overlooks the unrealized, highly specific imitative accommodation performed by the utilized schemes. Further, Piaget maintains, unlike Baldwin, that there is no intrinsic feeling quality uniquely identifying memory experiences, nor is there any difference in the impression obtained from a false memory as compared with a tme one, whether the memory is that of recognition, reconstruction, or recall. Finally, in regard to feelings of self-identity, a false memory gives the same impression of “I-ness” as a true one. (Claparkde had maintained, at about the same time that Baldwin was writing, that all memories have in common a feeling pertaining to self that he called “I-ness.”) Both relating memory to the past rather than the present and detecting the true memory from the spurious are not located within the memory experience itself but require the direct collaboration and decision-making of what Piaget calls the practical intelligence, which consists of classifications and spatial operations that are the same as those that enable people to structure their present experiences. G . AFFECTIVE MEMORY

In line with Piaget’s general theory of memory he expresses strong opposition in several places to the idea that there are “raw memories” that possess any validity. No exception is made for affective memories, and he thus stands in clear contrast to Baldwin, who maintained memory immediacy for the affective in contrast to the cognitive. Memory for Piaget is mediated by accommodated schemes whatever the content. Although the details remain vague, Bartlett (1932) gives an important place to the “affective setting,” which functions as an attitude that orients one toward the memory image. Such an affectively laden orientation can sometimes be advantageous in maintaining memory over long periods. On other occasions affectivity rather directly distorts the conventionalized retained memory schemata. But whatever the effect of affectivity it is not itself conceptualized as schemata. Not surprisingly, Piaget states that in actual practice the affective and cognitive mechanisms always remain indissociable although distinct; but as the same statement is frequently made about the relation of accommodation and assimilation and does not prevent extensive analysis of each, such a disclaimer is not in

Developtnentnl Memory Theories

209

itself adequate to account for the lack of description of affective memory content. In several places, in fact, Piaget speaks of affective schemes as clearly not the same as cognitive schemes, but details are lacking. He has also said very little about the origin of affective schemes, although it is implied that action schemes can lead to affective schemes. A little more elaboration is called for concerning the mention in the last section that Piaget sees little difference in structure between contents that become conscious and those that remain in the subconscious ( a deep Freudian unconscious is not really admitted). To the extent that contents can be kept out of consciousness because their potential niche is preempted by other schemes, cognitive repression analogous to affective repression takes place. It appears likely that orthodox psychoanalysts would not accept the analogy because too much nonconflictual or preconscious content would be involved. An important but unanswered question is whether affective repression is itself conceived as a conflict between schemes with integrity of organization viewed as the key to entering consciousness. Also unstated is whether there are interactions between the cognitive and affective, in other words are there affectively toned cognitive schemes? Although Piaget writes of assimilatory schemes that are centered on affective behavior and hence distorted, leaving us unsure whether veridical affective schemes occur at all, the answer seems to be negative as there appears to be little interaction between the cognitive and affective. A fairly strict and thoroughgoing parallelism is maintained: There are two sorts of schemes and two sorts of repressions. (Not that affective and cognitive schemes do not interact, for the work of Spitz is cited as showing that affectivity or its deprivation can be the cause of acceleration or delay in cognitive development.) Such a parallelism may be largely a way of excluding affective material from conceptual consideration insofar as it is possible to do this either as an epistemologist or a developmentalist. But there are many other alternatives if this is one’s goal; it is not at all necessary to make the strong assumption that the affective and cognitive structures that bring about memory are similar but parallel rather than interacting. The limitations of Piaget’s views on memory for affectivity are treated extensively and considerably criticized in a book on developmental memory by Guillaumin (1968). Guillaumin contends that an inadequate theory such as that of cognitive and affective parallelism is inevitable as long as representation is completely dependent on the internalization and interiorization of actions. He finds the qualitative gap too great and the chain of successive montages too long in the linkages between past sensorimotor behavior and present affective memory. In particular, he finds Piaget’s theory of the representation of affectivity in memory inferior to that of A. Rey, who maintained that a leading role is played by proprioception. A curious note is struck, for Americans at any rate, in that for the theory of internalization of actions Piaget is described as the immediate successor and theoretical heir of John B . Watson.

Bruce M . Ross and Stephen M . Kersr

2 10

IV.

Comparison Between Memory Theories

As we have pointed out in the preceding two sections, there are both similarities and differences in Baldwin’s and Piaget’s theories, even if one is in many ways the ancestor of the other. The similarities are qualified, however, because views do not remain unchanged over more than 60 years. Some similarities may be more apparent than real; an idea that has been fruitful for empirical studies survives, while another idea, appearing equally valid at the time, has fallen by the wayside as just another unproved speculation. It has also been the case, however, that more than a piecemeal description of memory development is unusual (witness the incompleteness of psychoanalytic theory in this regard). At the same time anyone who considers developmental memory phenomena will necessarily reach numerous conclusions that he holds in common with other memory theorists. Thus, both Baldwin and Piaget emphasize the uniqueness and singularity of the memory representation. Both also talk about images, but the imagery view was commonplace when Baldwin was writing and is a minority viewpoint today, though currently experiencing a comeback in some information-processing, descriptions of memory. Piaget is, therefore, careful to hedge any definition of imagery by enumerating its functional uses rather than by applying introspective descriptions. It is also apparent that for Piaget images are more a product of the intelligence than for Baldwin, who believed them to be the natural, though transformed, consequents of perception. Both theorists are careful to stress that not all the past constitutes memory and that not everything that can be said to be remembered in popular speech is part of the memory function. Baldwin makes this point differently from Piaget. Memory, in general, is characterized by its context and its confirmation process that differentiates some item or event that occurred in the past. But this confirmation is distinct from imitation on the one hand, with its conscious focus on present action, and habit on the other hand, in which consciousness lapses and routinization takes over. In Piaget’s well-known distinction between memory in the wide and in the narrow sense, consciousness is not a reference point. The schemes are retained as part of their own self-consistency but are, by and large, unknowable structures to the individual. They constitute memory in the wide sense, which deserves the designation of memory only in the technical sense of dealing with knowledge that was acquired in the past. Perhaps there is memory also in some “half-wide’’ sense, in that Piaget has been empirically able to work out semilogical organizations and considers them as stages on the way to completed schemes; the individual would seem to be capable of forgetting as well as remembering these incomplete transformations. But in his view of schemes as self-retaining organizations one also has the feeling, perhaps mistakenly, that he has been influenced by the idea of “generativity” that in linguistics furnishes an option to describing spoken words as merely memory revivals. The study of memory in the narrower (but usual) sense is left the task of explaining how the

Developtnetitol Memot?. Theories

21 I

figurative instruments interact with more permanent schemes to give the high degree of specificity that memory requires. Another area of similarity between Baldwin and Piaget is not only their common belief in the importance of the internalization of actions for engendering thought structures but also their emphasis on motor or efferent activity in general. For Baldwin this is put in terms of excess nervous motor discharge with a description in physiological terms, while Piaget discusses functional motor schemes, particularly during the early sensorimotor stage. It is assimilation of new movements to these schemes rather than higher forms of imitation that lays the basis for further mental development. But the emphasis remains sufficiently or, actions for Droz and Rahmy (1976) to characterize Piaget as, relatively at least, a peripheralist (again a similarity can be noted between Piaget and John B. Watson). Rather than differentiate successively higher forms of imitations as Baldwin did, Piaget has emphasized the coordination and accommodations of schemes and feedback to the individual both from the products of transformations and from the actions of transformations, with imitation playing but a partial role in the transformations themselves. (As an adequate description of Piaget’s theory in regard to transformations is more germane to his theory of intelligence than to memory, we touch on this subject only to identify the contrast with Baldwin’s theory of imitation.) Both Baldwin and Piaget conceptualize recognition as initially primitive and without imagery representation. Brit the similarity extends beyond this with recognition playing a par! in nonmemorial thought processes; for Piaget recognition is an inherent part of the sensorimotor schemes while, as previously described, Baldwin claims recognition as essential for the identification of sameness and persistence in external objects. Where images are invoked, both theorists-in regard to different topics-emphasize the transformational orders of which images are a part (which at a minimum implies for both the rejection of an epiphenomena1 interpretation of imagery). Baldwin stresses the idea that the child can gain validation of a memory by a transformational movement from the memory image to the percept of the specific modality that engendered the image. Piaget emphasizes that the contrary direction actiowscheme+image is the natural order in which memory content is acquired, and the reinstatement of this criginal order in reconstruction gives reconstruction a clear advantage in efficiency over recall. (The special importance of a reproductive stage is, of course, a theoretical distinction unique to Piaget.) In contrast, the importance of the social confirmation of memory is ignored by Piaget in his description of memory, although validating cognition by social transactions has been strongly emphasized elsewhere, particulxly in his early works (Piaget, 1926, 19?0), as necessary to attain social perspectives that diminish an egocentric, cognitive viewpoint.

212

Bruce M . Ross and Stephen M . Kerst

Other differences in approach round out our comparison of the two theorists. Baldwin’s view of direct afflective memory is clearly opposed to Piaget’s frequent assertions that “raw memory” does not exist in anything but an illusory form. But Baldwin expresses the belief that his system is in agreement in the affective domain with what he calls the trend toward “alogicism” of which Bergson is a leader. To Piaget, of course, Bergson is a never failing source of opposition as standing for a theory of fixed and unchanging memories, an objection independent of Bergson’s often condemned dualistic approach to memory that placed many memory functions outside the possibility of scientific study. In that Piaget sometimes labels schemes dealing with interpersonal relations as affective, it is apparent that there is a blurring between the cognitive and affective for Piaget that is lacking in Baldwin. The potential range of the affective domain is extended by Baldwin to take in almost the whole of esthetics interpreted very broadly, and in this sphere he makes any cognitive contributions very subordinate to affective intuitions as well as to affective memory (Baldwin, 1915). In the main, though, differences in memory description between the two theorists are largely due to subsequent progress made in developmental psychology over the many years since Baldwin presented his theory. For example, Piaget in recent years has strongly stressed the definition of a symbol as a nonarbitrary relation between the signifier and the significate, the signifier being classed as figurative and hence falling under the rubric of “memory in the narrow sense.” This characterization of symbols originated with the Swiss linguist de Saussure shortly after the period in which Baldwin was writing. In addition, of course, Piaget has himself developed a probabilistic theory of perception, particularly applicable to the preschool years, that is sufficiently constructivist in outlook to discourage any copy theory conception of perception, let alone memory. Perhaps most important for his ideas about memory, however, are the results of his empirical studies of children’s imagery. A main finding was that imagery development, particularly kinetic imagery, lagged behind what was commonly supposed and was not accurate until the 8- to 9-year old period, if then. Thus, for Piaget the role of imagery in memory as able to produce accurate representations is considerably circumscribed as compared with the easy assumptions about accuracy that Baldwin could make about children’s memory imagery.

V.

Topical Memory Research

A. POSSIBLE THEORETICAL CONTACTS

In this section we use the: word “topical” in a double sense meaning both classifying by themes and that such themes are of current interest. Two excellent critical reviews that cited a rnajority of the recent studies of children’s memory appeared in Volume 10 of this series (Brown, 1975; Hagen, Jongeward, & Kail,

Developmeiittrl Memorv Theories

213

1975). Therefore, we have felt no necessity to review the contemporary literature but, instead, selectively concentrate on issues raised by the previous theoretical sections. To be sure, several studies performed since the previous reviews will be mentioned. We would not like to give the impression that Baldwin was describing memory areas that have no current relevance or that Piaget’s ideas about memory pertain solely to the minority of investigators who either work within his theoretical interpretations or wish to challenge them. Thus, in an attempt to induce some rapprochment between our discussion of theories and today’s empirical studies, we will sample several topics that can be exemplified in recent memory research-recognition and imitation, schema theory, imagery and memory, and memory awareness. The studies we cite are not particularly influenced by Piaget, since our intent is to illustrate areas where a background consideration of the ideas of Baldwin and Piaget is of some interest whether one has a similar, opposite, or totally unrelated point of view. These areas tend to be those where basic methodological and explanatory problems have persisted over several psychological generations. Without exception they are also topics that have shown a resurgence of interest in recent years (e.g., G . Mandler, 1975, on conscious awareness). For several of the areas, renewed interest appears to be a function of contemporary concentration on cognition and mental transformations, of which the information-processing approach is partly cause and partly symptom. It is worth emphasizing in this regard that Hagen er al. (1975) maintain not only that the information-processing approach to psychology is a mentalistic one but also that it is the modern descendant of the act psychology of Brentano and Kiilpe. (Doubtless many psychologists of the information-processing persuasion would not claim such dubious ancestry and would admit no nearer a familial relation than second cousinship.) We have previously noted that the vocabulary used to describe psychological functions by Baldwin frequently employs the traditional act-psychology descriptions of mental objects and processes, although we have retained this form of description in only a minority of instances. The point is that in considering the memory explanations given by Baldwin, however nonphilosophical psychology has become, there is a greater affinity with current research than might first appear. It is certain in any event that Baldwin was the only developmentalist among those closely allied with the act-psychology approach. The challenge Piaget offers to the infomiation-processing approach, insofar as developmentalists assume this increasingly popular point of view, is i n another direction. It is obvious that for particular situations such as those at the boundaries between perception and memory, which are the situations most frequently selected for inforrnation-processing experiments, Piaget’s detailed process descriptions are either incomplete or nonexistent. In part this surely stems from inadequacies i n the theory, but it also arises from studied neglect on the part of Piaget’s theory of short-term and short-lasting processes except where they can

214

Bruce M . Ross and Stephen M . Kern

be used to diagnose a particular state or stage of underlying developmental competence. Often, in fact, substitutions can be made by the child in carrying out similar mental functions so that the specific details of which process was involved are of little or no importance in describing the child’s behavior from a developmental perspective. Up to now and quite in line with this developmental perspective, Piaget’s memory theory has had its chief impact on the study of long-term memory. Can information-processing mechanisms be extended to apply to longer lasting phenomena that show developmental changes and at the same time can the processes described by Piaget be made more precise? It is quite conceivable that both viewpoints may pass each other by and fail to make contact in any serious way. But if consolidation of any theoretical importance occurs, it is most likely to be in the area of children’s memory. B.

RECOGNITION AND IMITATION

We have pointed out that i n both Baldwin’s and Piaget’s theories recognition can vary from a simple to a complex performance. The simpler types of visual recognition have often been considered to be a primitive form of memory that would not be expected to show developmental change, since no retrieval pro1975; Olson, cesses and memory strategies need be involved (Hagen et d., 1973). However, Brown (19’75) cites more recent work which indicates that the failure to find developmental trends is contaminated by ceiling effects that reflect the spectacular ability of children and adults to recognize pictures (Brown & Campione, 1972). In addition, the instructions in a delayed-matching recognition task seem more difficult for children to understand than instructions for recall tasks that are theoretically more complex (Wickelgren, 1974). As a result, memory development may already have occurred before the minimum age at which conventional recognition taslcs are commonly used (3 or 4 years). Studies to be reviewed here indicate that in spite of these methodological restrictions, current experiments show that there are both strategic and nonstrategic developmental changes in recognition memory. Since both verbal and visual memory strategies have been found to improve simple recognition, it does not appear to be a primitive, elemental, and nonstrategic task. Although deliberate verbal labeling strategies can be effective, it is not clear how they work, and it seems unlikely that recognition memory development can be explained in terms of the mere addition of verbal labels. Instead, labeling may direct attention to important visual information to be remembered about a picture. Acquisition of deliberate visual strategies such as “paying attention to details” may improve with age, and greater efficiency in more automatic processes involved in the selection of crucial visual features for schema formation or storage may also be important. Experiments which show nonstrategic developmental changes in recognition have been designed to exclude verbal mediation strategies as explanations for

Developmenrd Memory Theories

215

obtained age differences. For example, Feinman and Entwisle (1 976) found improvement in recognition for photographs of children’s faces over the first, second, thud, and sixth grades when subjects were asked to select 20 previously presented pictures from a group of 40. Similarly, Perlmutter and Myers (1976) found that recognition memory for drawings of familiar objects improved from age 3 to 4, while spontaneous verbal labeling remained constant. Developmental trends are also found with memory for meaningful pictures when the recognition of large numbers (300-600) of items is tested. Hoffman and Dick (1976) found that recognition memory improved for 3-year-old, 7-yearold, and adult subjects. Since younger children’s performance was particularly reduced when the number of alternatives was increased from 2 to 4 , these authors suggest that improvement i n recognition with age may result from the older subjects’ improved efficiency at extracting salient information for storage, rather than from age differences in capacity. Development of strategy-free recognition memory is found in another type of task. J. M. Mandler and Day (1975) tested children and adults for recognition of the left-right orientation of nonsense forms and similar meaningful figures which could be generated in principle by the same procedure used to produce the nonsense forms. Only for nonsense forms were developmental differences found for kindergarten, grade school, and adult subjects, but the orientation of meaningful figures was remembered better than orientation for nonsense forms. This latter difference was apparently not due to simple verbal labeling of the meaningful forms, since subjects hesitated to name them and produced a wide variety of names when asked to label them. This difficulty of naming, however, did not increase the latency of recognition, a result which also suggests that recognition of orientation did not depend on naming. The authors propose that recognition of orientation for meaningful forms as opposed to nonmeaningful forms is easier because subjects have developed a representative schema for the form. Furthermore, instructions which informed subjects to remember orientation improved performance only slightly. This finding suggests that, for meaningful figures at least, orientation is remembered automatically without a deliberate strategy. These results are consistent with Frost’s (1972) findings where incidental memory for orientation was found with adult subjects who were directed to remember the names of meaningful drawings. Studies which show the effects of deliberate strategies on recognition have also involved controls for verbal mediation. Millar (1972) found that 4-year-old preschool children performed better at recognizing nonsense shapes when they were told to visually rehearse them over a 5-second delay. Latencies obtained for labeling were greater than those for recognition, so a verbal mediation explana. tion seems unlikely. Similarly, Tversky and Teiffer ( 1976) tested recognition memory for meaningful pictures where recognition distractors differed from the targets in orientation or detail but not in name. When kindergarten, third-, and fifth-grade children

216

Bruce M. Ross and Stephen M . Kerst

were told to pay attention to details in the pictures, performance improved for the oldest subjects. In addition, a pronounced age trend was found, as scores ranged from 70% to 90% correct foa the youngest and oldest subjects, respectively. It is unlikely that an increase in verbal labeling can account for these results, since distractor pictures had the same names as targets but differed with respect to visual details. Further evidence that memory strategies can affect recognition was found by Nelson and Kosslyn (1976). These authors tested recognition of 12 familiar and 12 abstract forms (e.g., puzzle pieces) by 5-year-olds and adults. Recognition improved with age, even with this relatively small number of items. When subjects were supplied with verbal descriptions of the drawings during presentation the performance of both age groups improved, but adult recognition improved only for the abstract forms. Verbal descriptions facilitated recognition, but it is not clear how this effect was produced, since subjects were not required to retrieve the label for recognition. Recognition latencies did not increase when descriptions were provided, which suggests that subjects may not actually have stored or retrieved the label. Nelson and Kosslyn propose that labeling may favorably influence the kind of perceptual information that is stored, an explanation that is similar to the feature selection hypothesis advanced by Hoffman and Dick (1976). To summarize, the observed effects of strategies and the age trends obtained cast doubt on the assertion that recognition memory depends on basic memory capacity, which shows no developmental improvement (Hagen el al., 1975; Olson, 1973). Research on imitation with children has been directed at the effects of verbal coding on observational learning (Bandura, Grusec, & Menlove, 1966; Coates & Hartup, 1969; van Hekken. 1969) rather than on imitation per se. The emphasis on verbal coding seems to change the task to one of memory for verbal descriptions which can be translated into actions at the time of recall. The youngest subjects used in recent research (4 years) are past the age where nonlinguistic sensorimotor imitation appears as a distinct form of early memory, so that the development of imitative memory has received little experimental attention. In the three studies cited above, children were shown films in which an adult model carried out a series of distinct actions such as building a tower of blocks in a unique way and shooting, a target off the top of the tower with a dart gun. In spite of the similarity of the material to be imitated and some attempts at replication, the effects of verbalization are not consistent across the three studies, and the authors of the two later ,studiesdo not provide compelling explanations for the differences. For example, van Hekken found that irrelevant verbalization (counting), which was included to suppress verbal mediation, did not interfere with learning, while Bandura et al. found that it did. It should be noted that retention of actions occurred in both studies even when verbalization was prevented, which suggests the presence of some nonverbal memory representation.

Van Hekken had classified 6- to 8-year-old subjects as mediators and nonmediators on the basis of their reversal or inconsistent shift performance in a discrimination learning task. Mediators did not outperform nonmediators. Only mediators’ performance was improved by subject-produced verbalizations. This result seems inconsistent with the findings of Coates and Hartup (1969), who found that 4- and 5-year-old children (presumably nonmediators) were aided by either subject-generated or experimenter-supplied verbalizations, while 7- and 8-year-old children (presumably mediators) performed worse with subjectgenerated verbalizations. In this study, a strong age difference between the performance of 4- and 5-year-old and 7- and 8-year-old children was found when they passively observed the model. Although this difference was substantially reduced with subject-generated and experimenter-supplied verbalization, it is not clear from the results of the three imitation studies cited above that the developmental trend can be accounted for by a straightforward verbal-mediation explanation. It may be that verbalization promotes attention or the encoding of relevant actions that are actually represented nonverbally . Work with adults has involved the imitation of actions that are not meaningful to the learner, that is, a sort of action equivalent to the nonsense syllable (e.g., vertical and horizontal movements of various distances i n a field of dots). Bandura and Jeffery (1973) argue that it is necessary to use novel actions that have not been integrated through prior learning in order to study observational learning. They indicate, however, that actions which have no functional value are not easily remembered. The fact that the actions to be remembered in this study were meaningless may explain why experimenter-provided letter and number coding systems for specific movements proved to be effective. Bandura and Jeffery (1973) point out that observational learning in natural settings involves new combinations of familiar activities that are already represented in memory, and consequently these actions are more easily remembered. It is precisely this area of memory for meaningful actions that seems to deserve exploration i n developmental research. The possibility of classifying different actions so that they can be treated as functional equivalents or “paraphrases” of actions would appear to call for analysis beyond the level of “verbatim” repetition which has been used previously. C . SCHEMA THEORY

The concept of some unit or organizing tendency or both by which memory can be encoded into contents that are larger than minimal discriminable segments is an enduring one. Whether one denotes such a unit as a schema or a plan or a conventionalization, it is the natural opponent of trace theories of memory. Wherever phenomenalism gains ascendancy there will be found explanations that tend to support something like a schema; wherever there is emphasis on the

21s

Bruce M . Ross and Stephen M. Kerst

possibility of physiological explanations (at least until recent biochemical theories), there will be an emphasis put on traces and small units or features. The exception that proves the above generalization is, of course, Bartlett’s (1932) schema with its physiological analogy in Head and Holmes’ theory of a postural, neurological schema, which for a previous generation gave the idea of schema acceptance as at least a conceivable possibility. Nevertheless, to date, physiological theorizing has proved to be only a cosmetic appendage for either trace or schema theories. We have taken pains to point out that schema ideas even when restricted to memory theories show considerable variety. Thus, Baldwin’s and Bartlett’s ideas of the schema are quite different from each other; let alone Piaget’s more subtle scheme-schema distinction and its implications. However, we will not be directly concerned with Baldwin or Piaget i n this section. For our purposes Bartlett’s descriptions remain limited because he gathered no developmental data and did not theorize in this regard. In a theoretical article that draws on psychoanalytic memory theory, Paul ( 1 967) attempts an updating of Bartlett’s ideas. He proposes that, developmentally, schemas tend to precede traces as effective memory agents. (One implication is that rote memorization is an achievement, and only the maturer child performs it successfully. Also, he maintains, only the older child is able to give eidetic recall. Since there are both coexisting schemata and traces, Bartlett’s problem of how isolated detail can be retrieved from stable schemita is overcome. In bringing this about, a matching process is of crucial importance. Trace carryovers, sometimes in the form of images, are combined into an organizational form and this construction is matched against a schema. This match “justifies” the schema, and the degree of success of the match is a “reality test” for the recall. It is in this matching process that affective considerations are manifested through the operations of needs and defenses and the expression of cognitive styles. The matching process is, however, neutral toward the proposition that schemata developmentally precede traces. Support is given by the claim that children perceive schematically while adults perceive realistically , with ultimate theoretical justification depending on Piaget’s interpretation of the circular reflex as leading to sensorimotor schemes rather than sensory images. Only a logical rather than a developmental reason is given for postulating traces. “I must know numbers before I can register facsimilies of them for later revival. But then we must provide for a process to subserve the formatiop of this facsimile, the storage of it, and the subsequent revival of it; this process cannot be identical with schematic processes and yet must become intimately associated with them” (Paul, 1967, p. 255). Certainly, Paul seems to be struggling with a dichotomy close to Piaget’s scheme-schema distinction. But Paul’s theory requires less reconstructive memory activity in that the traces that represent the fine details of the memory are themselves stored facsimilies. This is in line with the

Developmentol Memory Theories

219

usual conservationist tendencies of psychoanalytic memory theory. One conclusion is that Bartlett’s theory of a single class of schemata is unacceptable even to those sympathetic to his views; at least a dual process must be postulated in order to account for adequate retrieval. Several information-processing experimenters who have studied visual processing have made use of an analog of the scheme, the concrete “prototype” or quintessential exemplar. Comparisons can be made using the older philosophical terminology; the universal in some limited pictorial domain is represented by a prototype that necessarily includes particulars thought to be representative. In contrast, schemata for Bartlett and especially schemes for Piaget might be loosely characterized as generalizing universals from which particulars are deduced. But prototypes and schemata are alike in that they describe categories germane to thinking as well as memory processes, and they are central to descriptions of stimulus representation. In an experiment with adults, Posner and Keele (1968) had people learn to classify dot patterns that were variants of prototypes they had not seen. When the prototypes were later presented, they were classified more easily than new control dot-pattern variants. In a later experiment (Posner & Keele, 1970) they found that the inferred prototypes were subject to less forgetting over a 1-week interval than the previously learned variants. As there were some losses in memory for the learned variants, but these losses did not affect the ability to recognize the prototypes, the authors argue that prototypes are formed while classifications are being learned rather than at the time of the test presentation of the prototype. The study of visual prototypes was extended by Franks and Bransford (197 1) who described visual configurations i n terms of a prototype-plus-transformation model. In a series of four small experiments Posnansky and Neumann (1976) demonstrated that children as young as age 7 can form prototypical representations from a set of highly similar stimuli. These authors favor an attributefrequency model as against the prototype-plus-transformation model, but as the two experiments crucial to their argument were performed with letter trigrams rather than pictorial items, one can remain unconvinced of the generality of their model. With the current emphasis on items that cannot strictly be said to be remembered, such as inferred classifications and absent prototypes, this type of experimental approach is not restricted to memory studies but spills over into the more general problem of the categorization of things in general. In a series of 12 experiments on categorizing concrete objects Rosch, Mervis, Gray, Johnson, and Boyers-Braem ( 1976) investigated the categorization of objects in natural taxonomic categories. Their favored method was the use of sorting tasks in classifying objects in order to find what they termed “basic-level’’ categorization. An objeci’s “basic level” category is determined by the type of name which is applied to it most frequently, most spontaneously, and with the shortest reaction time. An example

220

Bruce M . Ross and Stephen M . Kerst

would be “apple” as a basic-level name with the name “fruit” at a superordinate level of classification, and the name “delicious apple” at a subordinate level. Two of the 12 experiments were performed with children. It was found using an oddity task that children as young as age 3 classified pictures of objects into basic-level categories in an adult manner. Using a sorting task and more pictures, children of age 5 classified at an adult level; but using superordinate categories, 8-year-olds but not the 5-year-olds classified as the adults did. In current usage, the ability to perform such category classifications draws on the long-term semantic memory of the children. What is more provocative, however, is the opportunity that the results of such classifications offer for future memory experiments. Some degree of calibration should be possible for to-be-remembered items in lieu of administering items judged to be intuitively alike or certified as relatively meaningless. Our brief look at schemalike entities has omitted two obvious research areas, one because there is too much material and the other because there is too little. Although Bartlett’s best-known results centered on memory for prose, this area has expanded too greatly to be feasibly considered here, even though the theories put forward in recent years still largely pertain to adult comprehension. We mention this omission because the possibilities of content analysis in connected discourse include themes, motifs, and plot structures at several levels, furnishing an expansive playground for the schema notion, whether considered as a prototype or a logical organization. The area where there is too little to write about is that of “memory for events” where compression and selectivity must necessarily play an important role in retention. In this area Loftus (1975), using films and postfilm interrogations, has revived the moribund psychology-of-testimony experiment with interesting results on the drawing of false inferences. As yet, however, researchers have not closely interrogated the child witness. D.

IMAGERY A N D MEMORY

Children’s visual imagery has recently been studied from at least five perspectives: (1) imagery as a deliberate mnemonic strategy with verbal materials; (2) spontaneous use of imagery as a representation in long-term associative memory for concrete objects; (3) imagery’s role in retaining perceptionlike analog information in short-term memory for visual materials; (4) the qualitative changes in the information retained in the image; and (5) its retention of perceptual information in long-term memory. Since imagery mnemonics have proven effective, they provoke the questions of whether they actually involve visual mental representations and what developmental changes are involved in imagery. Examples of the five approaches as they relate to these two questions will be treated here. 1. Instructions to generate pictorial interacting mental images from words or separate pictures in paired-associate tasks have been found to improve children’s

recall and recognition performance. Jusczyk, Kemler, and Bubis (1975) found that first- and fourth-grade children (but not adults) performed better on cued recall and recognition of declarative sentences when subjects either generated mental images for the sentences or were shown pictures which portrayed the actions described in the sentences. However, there was no clear evidence that there were qualitative differences i n the mode of mental representations used by the control and imagery groups. Anderson and Bower (1974) have suggested that imagery instructions are effective because they encourage subjects to elaborate meaningful relations in the material to be learned. Indeed, Bower (1970) reports that insbuctions to generate an interacting image for pairs of words improved children’s learning, while instructions to form separate images of the referents of the stimulus and response terms did not. The production of an interaction, rather than the generation of images per se seems to be the effective process here. In a massive study on age-related differences in the effectiveness of verbal and visual elaboration on noun-pair learning i n nursery, kindergarten, and second-grade children, Rohwer, Kee, and Guy (1975) found that elaboration was effective for all ages and that differential development effects for the two modes did not appear. Rohwer etal. concluded that the paired-associate task is not likely to be a useful means for analyzing the role of verbal and visual representations in learning and memory during early childhood. 2. Developmental changes in the dominance of visual and verbal associative memory are found in a different paradigm where imagery is not treated as a mnemonic strategy but as a mode of long-term memory storage. Cramer (1976) presented first- and fourth-grade students either lists of pictures, recorded names of the pictured objects, or pictures accompanied by names. The materials were constructed so that items had strong associates which could be used as distractors in a test of recognition (e.g., knge as a test item,fork as a distractor). In support of the hypothesis that verbal encoding increases with age, first graders made more associative errors in the picture condition than in the name condition, which suggests the dominance of associations among visual representations. Fourth graders, in contrast, made a visibly greater number of associative errors i n the name condition than in the picture condition, which suggests the increasing importance of verbal associations. 3. Other investigators have stressed the perceptual rather than the associative nature of imagery. The similarity between perceptual and memorial comparisons of concrete materials on dimensions such as size, color, shape, and orientation has lent support to the notion that subjects make some kind of “internal psychophysical judgment” among memory representations (Marmor, 1975; Moyer, 1973; Moyer & Bayer, 1976; Paivio, 1975; Shepard, 1975; Shepard & Chipman, 1970; Shepard & Podgorny, in press). Some attention has been paid to developmental aspects of these phenomena, often by simply asking young children to perform a task originally used with adults. For example, Marmor (1975)

222

Bruce M. Ross and Stephen M. Kersr

had 5- and 8-year-old children perform Shepard’s (1975) mental rotation task. In this paradigm, a subject is asked whether two successively presented visual forms are the same shape. The second form may be rotated a specified amount in relation to the original one. It is found with adults that recognition latencies are linearly related to the extent to which the forms differ in orientation. Presumably, a subject must rotate his mmtal-image representation of the first form to correspond to that of the second form before the two can be compared. Quantitative but not qualitative differences were found here. Marmor (1975) found that the same relationship held for 5- and 8-year-old children as for adults, but that the rate at which the mental image can be rotated increases with age. 4. Qualitative developmmtal differences in perception which seem to be related to differences in short-term memory for visual forms have also been found. Chipman and Mendelson (1975) provide evidence for changes in the kind of visual information that is perceived by children and adults. Subjects (4-5 years, 7-8 years, 9-10 years, and adult) were asked to judge which of two nonsense patterns that varied on contour and structure was simpler. Multidimensional scaling procedures showed that there was a developmental increase in the complexity of contour in a pattern that could be perceived as organized. In addition, adults weighted structure much more heavily than even 9- to 10-year-old children in their judgments of simplicity. Apparently, developmental changes in sensitivity to visual structure continue during adolescence. Arabie, Kosslyn, and Nelson ( 1975) report that qualitative developmental differences are found not only in visual perception but i n memory as well. Adults and 5-year-olds made similarity judgments of nonsense forms which were difficult to code verbally. Perceptual judgments of similarity (e.g., “Which of those two is most like that one?”) were made with both forms present, or while the forms were presented sequentially (standard last) for memory judgments. Multidimensional scaling techniques indicated that children and adults differed with respect to both perception and memory judgments. However, for both groups a clear correspondence was :shown between their perceptual and memory judgments. This correspondence suggests that children are not lacking in basic storage capacity for visual information, since the similarity between perceptual and memorial judgments would be disrupted if children’s memory were impoverished. Other factors mlentioned above (e.g., selectivity of feature extraction or schema formation) may need to be considered as explanations for qualitative developmental changes in visual recognition memory. 5. Perceptual analog models for long-term memory [used as a synonym for “knowledge of the world” concerning concrete objects (Paivio, 1 9 7 5 t a n apparent imagery counterpart to semantic memory] are just beginning to be treated developmentally, and it seems fruitful to do so using research with adults as a starting point. For example, Moyer (1973) and others (Moyer & Bayer, 1976) have investigated how subjects make comparisons among symbols that represent

Developmental Memory Theories

223

concrete objects in long-term memory. Adult subjects were first asked to rank animal names in terms of the relative size of the animal. In a task where subjects were later shown pairs of animal names and asked to judge which of the two animals was larger (Moyer, 1973), it was found that reaction time decreased as the difference between the rated sizes of the animals increased. These results are similar to those obtained when subjects compare the sizes of physical stimuli that have actually been presented (Curtis, Paulos, & Rule, 1973). Paivio (1975) argues that visual imagery is the mode in which such analog information is retained about concrete objects and events. In Paivio’s (1971, 1974) dual-coding theory the imagery and verbal systems are viewed as independent but connected domains for representing knowledge. The verbal system deals with discrete linguistic representations, while the imagery system is closely related to perceptual knowledge. An example of the type of data used to support this hypothesis is the finding that size comparisons based on information in long-term memory are faster when the items are presented as pictures rather than words (Paivio, 1975). Kosslyn (1975) provides a particularly explicit metaphor for the nature of the mental image which has been explored developmentally. He describes visual memory imagery in terms of a model involving computer-generated visual displays. In this model, images are not considered replays of unanalyzed sensations or simple “photographs in the head”-a conception of imagery which has received criticism (Pylyshyn, 1973). This model involves a more constructivist approach, since visual imagery in memory is taken to be a construction which is based on abstract units that are retained in long-term memory and are products of perception. These products of perception are not directly available; they are acted on by constructive processes that produce the subjective psychological experience of a mental image. The relationship between the mental image and the abstract perceptual units from which it is constructed, as Kosslyn (1975) proposes, is similar to the relationship between the display on a television screen and the computer program used to generate the display. The visual display (mental image) but not the program (perceptual units) can be seen by the viewer. When the size and complexity of the mental image are manipulated by visual context instructions and by added material, empirical support is found for the parallel between memory and perception in that (1) parts of large images are more identifiable than parts of small ones, and (2) added material to be imaged (e.g., a 4- or 16-cell matrix) retards attribute verification. Similar effects of size are found with first- and fourth-grade children (Kosslyn, 1976) when they are instructed to use mental imagery to retrieve semantic information from long-term memory to verify large and small perceptible properties of animals (e.g., a cat’s head would be large and its claws small). Under imagery insbuctions children responded almost as quickly as adults, but when imagery was not required adults answered much more quickly than children. These results seem congruent with

224

Bruce M . Ross and Stephen M . Kerst

those of Cramer (1976), cited above, which support the hypothesis that a kind of “visual semantic memory” develops earlier than its verbal counterpart. In conclusion, the recent shift in imagery research from paired-associate mnemonics to exploration of parallels between perceptual and memorial processes should provide a more direct approach for describing the nature of mental imagery. Developmentalists should be able to draw more adequate conceptualizations of imagery from current constructivist models (e.g., Kosslyn, 1975; Shepard, 1975) which supercede the more literal “photographs-in-the-head” models which have been heavily criticized (Pylyshyn, 1973). Qualitative developmental differences in perception and memory, as well as possible changes in the relationship between perceptual and memorial processes with age, can be investigated with multidimensional scaling techniques. Such measures provide rich data on age changes as to the qualitative description of information and its organization in fairly short-term visual memory. Possibly, measurement of this type could be used to study long-term retention as separate from general knowledge about perceptual characteristics of concrete objects. This separation is usually not made when the broad classification “semantic memory” is invoked. E. MEMORY AWARENESS

The problem of memory awareness in perhaps its most general form-How does one know what memory is?-has been stated by Wittgenstein (1953) in the Philosophical Investigations: Would this situation be conceivable; someone remembers for the first time in his life and says ‘‘Yes, now I know what ‘remembering’ is, what itfeels like to remember.”-How does he know that this feeling is ‘remembering’? Compare: ‘‘Yes, now I know what ‘tingling’ is.” (He has perhaps had an electric shock for the first time.)-Does he know that it is memory because it is caused by something past? And how does he know what the past is? Man learns the concept of the past by remembering. And how will he know again in the future what remembering feels like? (On the other hand one might, perhaps, speak of a feeling “Long, long ago,” for there is a tone, a gesture, which go with certain narratives of past times.) (p. 231e)

In other words there is a choice between believing that memory is the result of an inferred judgment or of a direct memory experience. It would seem that Wittgenstein believes that it must be inferred because not only is it a little ridiculous to compare knowing memory to knowing an electric shock, but shortly before the quoted passage on the same page Wittgenstein writes: “Remembering has no experiential content.--Surely this can be seen by introspection? Doesn’t it show that there is nothing there, when I look about for a content?” But as Wittgenstein indicates by his concluding parenthesis, there can be some doubt. He does not, for example, dispose of the possibility of direct experience of memory as neatly as he previously demonstrated that “to understand” could not

Developtnentnl Memory Theories

225

be descriptive of a specific mental process. The above quotation was not Wittgenstein’s only treatment of memory awareness as an epistemological puzzle. On the next-to-the-last page of the Brown Book Wittgenstein (1958) tilts the case a little more in favor of feelings of memory. But isn’t there also a peculiar feeling of pastness characteristic of images as memory images? There certainly are experiences which I should be inclined to call feelings of pastness, although not always when I remember something i s one of these feelings present.-To get clear about the nature of these feelings it is again very useful to remember there are gestures of pastness and inflexions of pastness which we can regard as representing the experiences of pastness. (p. 184)

Here Wittgenstein suggests that even though memory may not be a direct feeling, sometimes there are at least symbols that are experienced as standing for the past and so denoting memory images as different from other images. Wittgenstein (1958) illustrates this further by naming a specific German tune played with the right expression as the “most elaborate and exact expression of a feeling of pastness that I can Imagine” (p. 184). Similar opposition between a theory of direct versus assimilated experience has been discussed previously in regard to affective memory with Baldwin favoring the former view and Piaget the latter. It is a curious coincidence that Wittgenstein (1953) emphasizes the term “assimilate” when arguing for the nonexperiential identification of memory. “I get the idea of a memory-content only because I assimilate psychological concepts. It is like assimilating two games. (Football has goals, tennis not.)” (p. 231e). Quite apart from the correct determination of such theoretical disputes, psychologists can always take evasive action and view such equivocation as a fecund basis for gathering interesting data. How do individuals themselves conceptualize their own memories and what are they experiencing? That is, there exists a commonsense psychology of cognition, and naive conceptions of memory must be a major component of such a psychology. A start has been made in studying these “metamemory” phenomena developmentally by Flavell and his associates (Flavell & Wellman, 1976; Kreutzer, Leonard, & Flavell, 1975). Thus far, empirical work has concentrated on younger children, though not below age 4,both as to their knowledge of their individual memory capabilities and as to what they believe to be true about memory in general, or at least as distinct from other mental processes. Also not a necessity but characteristic of this work as a first effort is that it has been closely tied to memorization tasks of the simplest and most straightforward type. This allows comparisons between actual and estimated performance, but in a number of the reported experiments the individuals who make the estimates do not themselves always carry out the actual task. Although not unexpected and touched on by both Baldwin and Piaget, the thoroughgoing confusions between perception and memory for preschool chil-

226

Bruce M . Ross and Stephen M. Kerst

dren are perhaps the most important and consistently supported findings to date. The area of affective memory so far remains untouched by the metamemoe approach, but it seems unlikely to remain so long.

REFERENCES Anderson, J. R., & Bower, G. H. Human Associative memory. Washington, D.C.: Hemisphere, 1974. Arabie, P., Kosslyn, S. M., & Nelson, K.E. A multidimensional scaling study of visual memory of 5-year olds and adults. Journal of Experimental Child Psychology, 1975, 19, 327-345. Baldwin, J. M. Thought and things. Vol. 1: Functional logic or genetic theory of knowledge. New York: Macmillan, 1906. (Reprinted New York: Arno h e s s , 1975.) Baldwin, J. M. Thought and things. Vol. 2 Experimental logic or genetic theory of thought. New York Macmillan, 1908. (Reprinted New York Amo Press, 1975.) Baldwin, J. M. Thought and things. Vol. 3: Interest and art being real logic. I : Genetic epistemology. New York: Macmillan, 1911. (Reprinted New York: Amo Press, 1975.) Baldwin, J. M. Generic theory of reality being the outcome of genetic logic as issuing in the aesthetic theory of reality called pancalism. New York Pumam, 1915. (Reprinted New Y o k Arno Press, 1975.) Baldwin, J. M. Mental development in the child and the race (3rd ed., rev.). New York Macmillan, 1920. (Originally published, 1906.) Bandura, A,, Grusec, J. E., & Menlove, F. L. Observational learning as a function of symbolization and incentive set. Child Development, 1966, 37, 499-506. Bandura, A., & Jeffery, R. W. Role of symbolic coding and rehearsal processes in observational learning. Journal of Personality and Social Psychology, 1973, 26, 122-130. Bartlett, F. C. Remembering: A study in experimental and social psychology. Cambridge, Eng.: University Press, 1932. Boring, E. G. A History of experimental psychology (2nd ed.). New York Appleton, 1950. Bower, G. H. Analysis of a mnemonic device. American Scientist, 1970, 58, 4%-510. Broughton, J. M. The developmeni' of natural epistemology in adolescence and early adulthood. Unpublished doctoral dissertahon, Harvard University, 1975. Broughton, J. M., & Riegel, K. F. Developmental psychology and the sev. Paper presented at the meeting of the New York Academy of Sciences, New York, April 1976. Brown, A. L. The development of memory: Knowing, knowing about knowing, and knowing how to know. In H. W. Reese (Ed.), Advances in child development and behavior (Vol. 10). New York Academic Press, 1975. Brown, A. L., Campione, J. C. Recognition memory for perceptually similar pictures in preschool children. Journal of Experimental Psychology, 1972, 95, 55-62. Chipman, S. F., & Mendelson, M. J. The development of sensitivity to visual structure. Journal of Experimental Child Psychology, 1975, 20, 41 1 4 2 9 . Coates, B., & Hartup, W. H. Age and verbalization in observational learning. Developmental P~chology,1969, 1, 556-562. Cramer, P. Changes from visual to verbal memory organization as a function of age. Journal of Experimental Child Psycholog.y, 1976, 22, 50-57. Curtis, D. W., Paulos, M. A,, &Rule, S. J. Relation between disjunctive reaction time and stimulus difference. Journal of Experimental Psychology, 1973, 99, 167-173. Droz, R., & Rahmy, M. Understanding Piaget. New York International Universities Press, 1976. Feinman, S., & Entwisle, D. R. Children's ability to recognize other children's faces. Child Development, 1976, 47, 506-510.

Developmental Memory Theories

227

Flavell, J. H., & Wellman, H. M. Metamemory. In R. V. Kail & J. W. Hagen (Eds.), Memory in cognitive development. Hillsdale, N.J.: Lawrence Edbaum Associates, 1976. Franks, J. J., & Bransford, J. D. Abstraction of visual patterns. Journal ofExperimenta1 Psychology, 1971, 90, 65-74. Frost, N. Encoding and retrieval in visual memory tasks. Journal of Experimental Psychology, 1972, 95, 317-326. Guillaumin, J. La genese du souvenir. Paris: Presses Universitaires de France, 1968. Hagen, J. W., Jongeward, R. H., Jr., & Kail, R. V., Jr. Cognitive perspectives on the development of memory. In H. W. Reese (Ed.), Advances in childdevelopmentand behavior (Vol. 10). New York Academic Press, 1975. Hoffman, C. D., & Dick, S. A developmental investigation of recognition memory. Child Development, 1976, 47, 794-799. Inhelder, B. Memory and intelligence in the child. In D. Elkind & J. H. Flavell (Eds.), Studies in cognitive development: Essays in honor of Jean P iaget. London and New York: Oxford University Press, 1969. Jusczyk, P. W., Kemler, D. G., & Bubis, E. A. A developmental comparison of two types of visual mnemonics. Journal of Experimental Child Psychology, 1975, 20, 327-340. Kosslyn, S. M. Information represented in visual images. Cognitive Psychology, 1975,7,341-370. Kosslyn, S . M. Using imagery to retrieve semantic information: A developmental study. Child Develqment, 1976, 47, 434444. Kreutzer, M., Leonard, C., & Flavell, J. H. An interview study of children’s knowledge about memory. Monographs of the Society for Research in Child Development, 1975, 40(1, Serial No. 159). Liben, L. Memory in the context of cognitive development: The Piagetian approach. In R. V. Kail, Jr. & J. Hagen (Eds.), Perspectives on the development of memory and cognition. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1976. Liben, L. Memory from a cognitive-developmental perspective. In W. Overton & J. Gallagher (Eds.), Knowledge and development. Vol. I: Advances in research and theory. New York: Plenum, 1977. Loftus, E. F. Leading questions and the eyewitness report. Cognirive Psychology, 1975,7,560-572. Mandler, G . Consciousness: respectable, useful, and probably necessary. In R. L. Solso (Ed.), Information processing and cognition: The Loyola symposium. Hillsdale, N.J.: Lawrence Erlbaum Associates, 1975. Pp. 229-254. Mandler, J. M., & Day, J. Memory for orientation of forms as a function of their meaningfulness and complexity. Journal of Experimental Child Psychology, 1975, 20, 430-443. Marmor, G. S. Development of kinetic images: When does the child first represent movement in mental images? Cognitive Psychology, 1975, 7, 548-559. Millar, S. Effects of instructions to visualize stimuli during delay on visual recognition by preschool children. Child Development, 1972, 43, 1073-1075. Moyer, R . S. Comparing objects in memory: Evidence suggesting an internal psychophysics. Perception & Psychophysics, 1973, 13, 180-184. Moyer, R. S., & Bayer, R. H. Mental comparison and the symbolic distance effect. Cognitive Psychology, 1976, 8, 228-246. Nelson, K. E., & Kosslyn, S. M. Recognition of previously labeled or unlabeled pictures by 5-year olds and adults. Journal of Experimental Child Psychology, 1976, 21, 40-45. Olson, G. M. Developmental changes in memory and the acquisition of language. In T. E. Moore (Ed.), Cognitive development and the acquisition of language. New York: Academic Press, 1973. Paivio, A. Imagery and verbal processes. New York: Holt, 1971. Paivio, A. Language and knowledge of the world. Educational Researcher, 1974, 3, 5-12.

228

Bruce M . Ross and Stephen M. Kerst

Paivio, A. Perceptual comparisons through the mind’s eye. Memory and Cognition, 1975, 3, 635647. Paul, I. H. The concept of schema in memory theory. In R. R. Holt (Ed.), Motives and thought: Psychoanalytic essays in honor of David Rapaport. Psychological Issues, 1967, S(Monograph 18/ 19). Perlmutter, M., & Myers, N. A. Recognition memory in preschool children. Developmental Psychology, 1976, 12, 271-272. Piaget, J. The language and thought of the child. New York: Harcourt, Brace, 1926. Piaget, J . The child‘s conception of physical causality. London: Kegan Paul, 1930. Piaget, J. Play, dreams and imitation in childhood. New York: Norton, 1951. Piaget, J. The origin of intelligence in the child. New York: International Universities Press, 1952. Piaget, J. The construction of realiry in the child. New York: Basic Books, 1954. Piaget, J. On the development of memory and identity. Barre, Mass.: Clark University Press, 1968. Piaget, J. The child‘s conception of time. London: Routledge & Kegan Paul, 1969. (orig. publ. 1946.) Piaget, J. Main trends in psychology. In Main trends of research in the social and human sciences (Part 1). The Hague: MoutonAJnesco, 1970. (Published separately, Main trends inpsychology. New York: Harper, 1973.) Piaget, J. Affective unconscious and cognitive unconscious. In The child and realify. New York: Grossman, 1973. Pp. 31-48, Piaget, J., & Inhelder, B. Mental imagery in the child. New York: Basic Books, 1971. Piaget, J., & Inhelder, B. Memory and intelligence. New York: Basic Books, 1973. Posnansky, C. J., & Neumann, P. G. The abstraction of visual prototypes by children. Journal of Experimental Child Psychology, 1976, 21, 367-379. Posner, M. I., & Keele, S. W. On the genesis of abstract ideas. Journal ofExperimenta1 Psychology, 1968, 77, 353-363. Posner, M. I . , & Keele, S. W. Retention of abstract ideas. Journal of Experimental Psychology, 1970, 83, 304-308. Pylyshyn, Z . W . What the mind’s eye tells the mind’s brain: A critique of mental imagery. Psychological Bulletin, 1973, 80, 1-24. Roback, A. A. History of American psychology. New York: Library Publishers, 1952. Rohwer, W. D., Jr., Kee, D. W., &Guy, K.C. Developmental changes in the effects of presentation media on noun-pair learning. Journal of Experimental Child Psychology, 1975, 19, 137-152. Rosch, E., Mervis, C. B., Gray, W. D., Johnson, D. M., & Boyes-Braem, P. Basic objects in natural categories. Cognitive Psychology, 1976, 8 , 382439. Ross, B. M., & Furth, H. G. Children’s memory: Research and theory. In J. Piaget, P. Mounoud, & J. P. Bronckhart (Eds.), Psychologie, Encyclopedie de la Pleiade. Paris: Gallimard, in press. Sewny, V. D. The social theory of James Mark Baldwin. New York: King’s Crown Press, 1945. Shepard, R. N. Form, formation, and transformation of internal representations. In R. L. Solso (Ed.), Information processing olnd cognition: The Loyola symposium. Hillsdale, N.J .: Lawrence Edbaum Associates, 1975. 4. 87-122. Shepard, R. N., & Chipman, S. Second-order isomorphism of internal representations: Shapes of states. Cognitive Psychology, 1970, 1, 1-17. Shepard, R. N., & Podgorny, P. Cognitive processes that resemble perceptual processes. In W. K. Estes (Ed.), Handbook of learning and cognitive processes. New York Academic Press, in press. Stacey, 3 . T., & Ross, B. M. Scheme and schema in children’s memory of their own drawings. Developmental Psychology, 1975, 11, 3 7 4 1 . Tversky, B., & Teiffer, E. Development of strategies for recall and recognition. Developmental Psychology, 1976, 12, 406410.

Developmental Memory Theories

229

van Hekken, S. M. J . The influence of verbalization in observational learning in a group of mediating and a group of non-mediating children. Human Developmenf, 1969, 12, 204-213. Wickelgren, W. A. Memory. In Psychology and fhe handicapped child (DHEW Publication No. (OE) 73-05000). Washington, D.C.: U.S. Government Printing Office, 1974. Wittgenstein, L. Philosophical invesfigarions (translated by G . E. M. Anscombe). New York: Macmillan, 1953. Wittgenstein, L. The blue and brown bookr. New York Harper, 1958.

This Page Intentionally Left Blank

CHILD DISCIPLINE AND THE PURSUIT OF SELF: AN HISTORICAL INTERPRETATION

Howard Gadlin' UNIVERSITY OF MASSACHUSETTS

I. INTRODUCTION.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

231

11. THE NINETEENTH CENTURY-THE TRANSFORMATION OF THE TRADITIONAL FAMILY . , . , . . . . . . . , . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

235

III. THE NATURE OF COLONIAL CONTROL . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . .

236

IV. CONTROL IN THE JACKSONIAN ERA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

238

V. THE IDEOLOGY OF THE FAMILY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

240

VI. CHILD REARING IN THE MODERN E R A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

244

DISCIPLINE . . . . . . , . . . , .

248

VIII. PERSONHOOD AND CHILD DISCIPLINE IN CONTEMPORARY AMERICA . .

253

IX. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

259

V11. CHILD REARING IN THE MODERN ERA-CHILD

.

260

., , .., , ., ., , .. , , , , ., . . , . .. .. . . .. .. . . . ... . .. ... .. . .. . .

261

X. A NOTE ON THE ROLE OF CHILD DEVELOPMENT AND PSYCHOLOGY.. REFERENCES

, ,

I.

Introduction

Contemporary literature, both popular and professional, abounds with discussions of the crisis in the family. Of special concern are the methods of 'I am endebted to Jan Dizard, Harold Raush, and Cathy Portuges for numerous discussions of the issues considered in this chapter. 23 I

232

Howard Gadlin

discipline and training which function as cornerstones of socialization practices. Of course, this concern is not new. Indeed, as Joseph Featherstone (1974) has observed: “ . . . our pervasive: sense of crisis concerning children and the family is quite traditional, being part of a worried conversation that began in the 1820’s and has continued to the present” (p. 163; cf. Bane, 1973; Beck, 1973). This “worried” conversation has also been a contradictory one. On the one hand, the literature speaks of a great uncertainty about the future of the family and a deep uneasiness about the way we treat our children. On the other hand, although there are always those who critically compare the spoiled offspring of the present with the exemplary children of the past, the dominant tendency, especially in the twentieth century, is for each generation of parents to congratulate itself for its enlightened treatment of its children. Child-rearing practices of past generations are seen as inhumane or uninformed. This makes sense because the pace of social change practically guarantees that the methods whereby any given generation is reared will1 be outmoded by the time that generation begins to raise its own children. Caldwell (1964) has noted that in the twentieth century major trends i n infant care lasted about 20 years. But the current inappropriateness of past modes of child re:aring is no guarantor of the wisdom or correctness of present methods, and parents and experts may well have a reasonably strong sense of what they ought not to be doing, without any firm confidence about what they should do. Thus, there is good reason behind both poles of this contradictory national discussion about chilld rearing. The purpose of this chapter is to locate this discussion within a critical histoncal review of socialization pra.ctices in the United States.’ I will attempt to relate changes in ideals and techniques of socialization to changes in the dominant conceptions of self and self-fulfillment. These in turn will be connected to transformations in the structure and function of the family within a broader context of social, economic, and political changes. My main concern will be with the changes in the ways people attempt to direct the development of their children, with special consideration given to issues of discipline and control. It is the transformation of the function of the family in the context of modernization32My original hope had been to integrate a consideration of differences in child rearing across class and between boys and girls throughout the chapter. This did not prove possible, in some cases because there simply is not enough detailed information. More significantly, if recent data are any indicator of past complexity the class and sex differences are, in some respects, enormous and could not be incorporated into a chapter of chis length without it becoming a mere catalog of disciplinary techniques. I have written as if the practices under consideration applied equally well to boys or girls; obviously they were mast pertinent to boys. Still, especially for past epochs we are examining advice more than behavior (see Mechling, 1075), and it appears quite certain that the ideals that informed child-rearing advice in the past were ideals formulated with boys in mind. 1 suspect they still are to a very large degree. At any rate, a more differentiated examination of child-rearing practices will have to await further work. 3I use modernization simply as a one-word designation for the development associated with the rise of indushial capitalism and not as it I S used by the structural functionalists (see Lasch, 1975b).

Child Discipline and the Pursuit of Self‘

233

i.e., industrialization and urbanization-which will give shape to the discussion which follows (cf. P. Laslett, 1971). It is important that this analysis not be taken as an argument that the changes in the family are caused by urbanization and industrialization. It is more appropriate to speak of industrialization and urbanization as providing the context within which the meaning and purposes of family life are transformed. As for causal relationships, we cannot really say. The old myth about the nuclear family being the product of industrialization has been gradually put to rest; the nuclear family, in most of Europe, precedes industrialization (see, e.g., P. Laslett & Wall, 1972). But the nuclear family is not the modern family. Clearly, some changes in the family, both here and in Europe, come before industrialization and urbanization; we might even say that they provide the groundwork for these developments (Shorter, 1975). At the same time, it is equally apparent that the effects on the family of urbanization and industrialization are enormous. It is safest to conclude that the relationship between changes in the family and larger social change is dialectical-changes in each are involved in and require changes in the other. Furthermore, modernization is not a process which occurred once and for all in the late nineteenth century. Throughout the nineteenth century, and to some extent even into the early years of the twentieth century, “a profound tension existed between the older American preindustrial social structure and the modernizing institutions that accompanied the development of industrial capitalism. . .” (Gutman, 1973, p. 540). If we consider as well immigration and the ethnic, regional, and other sources of diversity, it becomes all too apparent that any portraits drawn are necessarily idealized types. The characterizations drawn here then do not necessarily present the child-rearing patterns found even in the majority of families-such a determination is, in fact, not generally possible. Rather, I have attempted to reconstruct past epochs with some knowledge of subsequent history in mind, emphasizing those aspects of parental control of young children which were centrally related to what now appear as the major themes in changes in personal identity through American histor9 (Potter, 1954; Stannard, 1971). It is especially important when describing the transformations in the family, the elimination of many of its traditional functions, and the elaboration of its concern with child rearing, not to overly romanticize the quality of family life in 4Although I have reviewed almost all of the past 20 years’ literature on the history of the family and the history of childhood as well as the past 10 years’ literature pertaining to child socialization and parental control of young children, this i s not an attempt to collate all that has been said on the subject. The review is selective and interpretive. Before proceeding further a cautionary note is needed. In the historical review especially, it has been necessary to concentrate on the middle-class, white, and for the most part Protestantssimply because it is they for whom most information is available. Even within these limits, however, it is necessary to homogenize phenomena which were, in actuality, far more diverse. At any point in time, backward- as well as forward-looking childrearing practices coexisted with the styles which are here offered as representative of particular epochs.

Howard Gadlin

234

the past, particularly in relation to the treatment of children. Certainly there was much that was oppressive and constraining, if not abusive, in child-rearing practices in the traditional family. Burgess’ description of the “. . . transition from traditional institution type of family to the person-centered unit of companionship . . .” captures very succinctly some of the advantages for children of the modern family which he states are: . . . characterized by such concepts as respect for the person (both child and adult), satisfaction in personal interaction, pride in growth and development, and a permissive, growth-promoting type of guidance as opposed to the more traditional attempts to “make” children conform to patterns of being neat and clean, obedient and respectful, polite and socially acceptable. (Duvall, 1946, p. 202-203)

At the same time, while recognizing that some features of modem child-rearing practices may legitimately be considered advances or improvements over traditional modes, it is absurd to argue, as has deMause (1974) and some of his followers, that: “the history of childhood is a nightmare from which we only recently have begun to awaken. Dehlause proceeds, after attempting to document that nightmare, to announce the dawn of a golden age of child rearing in which parents are characterized as oriented toward helping their children along to the full actualization of their human potential. Somewhere between nostalgically romanticizing the past and glorifying the present there lies the recognition that modern child-rearing practices represent an “. . . adjustment by the family to changed circumstance” (D. Calhoun, 1973). These new circumstances meant far-reaching conflict and with that conflict both Post and newly found selves (Thompson, 1966, pp. 93-94). And it should be noted that differences in historical interpretations of the changes in child-rearing practices are not merely academic disputes. Parental control of young children is a microcosm of the methods of control appropriate to a given society, at a given moment in its history. The way in which control is exercised over children is important as a reflection of adult and societal relationships, and as a model for the interactions a child will be expected to experience as he or she matures. Thus, there is a special importance to understanding child-rearing practices: Misunderstanding the forms of control of young children risks a similar misunderstanding of the nature of social control in our adult lives. Misunderstanding the forces that require those forms of control risks misunderstanding the limits of adult freedom (Gadlin, 1976; Rothtnan, 1971). There have been numerous descriptions of the transitions in the control of young children in American history (Borstelmann, 1976b; Bremner, 1970; deMause, 1974; Miller & Swanson, 1958; J. E. Senn, 1955; Smith & Gadlin, 1975). Borstelmann (1976b) has remarked: “All viewers of the changing family landscape see a marked change of orientation over time from strict, authoritarian, obedience-oriented parental dominance to a child-centered, low or indirect parental control approach, along with considerable ambivalence about indulgence in ”

Child Discipline and the Pursuit of Self

235

the process” (p. 7). Especially striking are the changes in the nineteenth century, for they are associated with a wide range of social and economic changes (Bailyn, 1960).

11. The Nineteenth Century-The Transformation of the Traditional Family5 There developed in the nineteenth century a new awareness of children; their welfare became “the preeminent goal of family life” (Bremner, 1970, p. 344). As emphasis on children’s welfare grew, it became a legal as well as social consideration. For example, until the 18OOs,custody of children was customarily given to the father as the head of the family. As the interests of children came to be seen as more important than traditional parental authority, custody was sometimes given to the mother when that was deemed in the best interest of the child. Welfare of children also became a primary justification-and limitation-of their socialization and control. In writings of the early nineteenth century parental authority becomes less absolute, tempered by considerations of the development and individuality of the child. There is a certain irony here, for the heightened interest in children develops at precisely that time in the nineteenth century when the family becomes a more private institution than it had been previously, and child-rearing practices are no longer considered matters of public policy within each community. This change is part of the beginning of the development of the United States from the primarily rural and agricultural society, to a predominantly urban and industrial nation. Even in 1850 only 16% of the American labor force was engaged in either manufacturing or construction. Four-fifths of the population was rural and two-thirds of the workers were farmers. By the turn of the century these ratios had been essentially reversed (Furstenberg, 1966). It is difficult for us to imagine the pervasiveness of the changes wrought within the process of modernization, for our families are the results of a succession of adaptations to those changes (cf. Hunt, 1970). Economically, the family has gone from serving as the center of production to its current role as a unit of consumption (Weinbaum & Bridges, 1976). Socially, the family had served as the pivotal institution in a network of closely interrelated households; currently it plays a greatly reduced role as the private home base ‘In the analysis that follows I will be speaking of changes in the American scene between 1825 and 1975. For convenience, the earlier portion of that period will be referred to as the Jacksonian era (Jackson was President from 1829-1837). The period of intensive industrial development in the United States is after 1850. The changes in the family discussed here begin on a small scale in the Jacksonian America and spread rapidly with industrialization through the remainder of the century. With industrial development comes the growth of urban centers. To cite just one statistic, there was no city larger than 50,000 people in 1790, but by 1830 more than one-half million people were in cities at least that large (Rothman, 1971, p. 57).

236

Howard Gadlin

from which its members leave or to which its members retreat for most of their social interactions and personal satisfactions (Gadlin, 1976a). This represents a profound change from the conditions of family life i n premodern America where personal life was intermeshed with public existence and intervention in malfunctioning families was customary. For example, a colonial child might be placed in another home if the authorities deemed he or she was not receiving a proper upbringing. Furthermore, it was fairly common practice for children, especially boys, to be placed in the households of families other than their own where, depending on economic status, they served as apprentices or servants and where they lived and spent a major portion of their childhoods. By the end of the nineteenth century this had all changed, people’s lives were divided between their work and their homes, and there emerged in the family an unprecedented arena of private life “. . . and consequently a decrease in direct social control over and social support for the performance of family roles” (B. Laslett, 1973, p. 489). The meaning of these changes becomes clearer when we look specifically at some of the differences in the control of young children as practiced by colonial and Jacksonian Americans. The Jacksonian era is the first period in American history in which we find the existence, on a significant scale, of modem families and a modern family ideology, i.e., families in which the goal of individual self-realization overshadows community solidarity and stability (Shorter, 1975, p. 19).

In. The Nature of Colonial Control The avowed purpose of controlling the child was the breaking of his will, the necessary restructuring of his natural evil propensities which reflected Puritan beliefs in the originally sinful nature of man. The primary social model of parental control was theological: A father is to his child as God is to human beings. From the theological model comes the justification for strict parental control: “Until a child will obey his parents, he can never be brought to obey God” (Greven, 1973, p. 78). Thus disobedience becomes a “moral disorder” and must be treated and controlled accordingly. The colonial model of parental control is also martial, i n that parents and children are frequently at war, working at inherently incompatible crosspurposes. J. S. C. Abbott, a well-known minister, spoke of parendchild conflicts in terms of the “victor” and the “vanquished.” Stressing the importance of the mother’s never losing a battle, he noted that, “if the child then obtains a victory, it is almost impossible for the mother afterward to regain her authority” (Greven, 1973, p. 123).

Child Disciplitir rind the Pursuit of Serf

231

Control was also considered a highly rationalprocess, as “rational” was then defined. This did not mean that the methods of control of children should be reasonable (i.e., not exceedingly harsh), but rather that punishment should be addressed to the rational faculties of the offender, even if that punishment be administered through harsh physical discipline. J. S. C. Abbott wrote that punishment should be inflicted “with composure and with solemnity” (Greven, 1973, p. 126) instead of in a flush of unreasonable anger. Finally, the nature of control was completely authoritarian. A parent’s authority was to be reciprocated by a child’s obedience; neither was to be qualified by explanation, considerations, or conditions. Authority rather than understanding should be the immediate cause of obedience: “Obedience is absolutely essential to proper family government. Without this, all other efforts will be in vain. Neither is it enough that a child should yield to your arguments and persuasions. It is essential that he should submit to your authority” (Greven, 1973, pp. 119-120). Physical punishment of children was an expression of authority which it was almost the parent’s duty to exercise. Parents controlled their children directly and physically, and with certainty of the propriety of their actions. An aggressive or disobedient colonial child was not isolated, but humiliated in front of his peers. The public aspect of discipline in the colonial era was important for the adults and children alike. Just as children were punished before their peers, offending adults were publicly displayed in the stocks. Such punishments reflected a means of social control based on public shame rather than on internal guilt. The manners taught to and expected of colonial children were also an aspect of parental control and were essentially rites of obedience. It is difficult to assess the age at which colonial parents began to exercise these kinds of control over their children. Demos states that control of children appeared between the ages of 1 and 2. Abbott, as always the voice of severity and rectitude, warned parents: “NEVER THINK THAT YOUR CHILD IS TOO YOUNG TO OBEY” (Greven, 1973, p. 124). He felt that if canaries and dogs can be taught to obey, children can as easily learn obedience. He describes one woman who, for the first year of her child’s life, “. . . would be obedient to the wishes of the child. But, by the time the child was one year of age, she considered it old enough to be brought under the salutory regulations of a well-disciplined family” (Greven, 1973, p. 125). The nature of control, in the colonial era, was thus ideally hierarchical, authoritarian, rational, educational, and reciprocated by unquestioning obedience. The means of control were external, in that they were based on physical punishment and on shame rather than on guilt, formal, in the emphasis on manners and on the positional authority of the parents, and direct, in that the parent dealt directly and physically with the child. A parent controlled his child’s aggression

238

Howard Gadlin

and disobedience, when the:y occurred, through prescription and direct punishment; he did not carefully manipulate the environment to make it more difficult for the child to transgress (Earle, 1899; Greuen, 1966; Stannard, 1974). In this way, control of young children well represented the dominant means whereby social cohesion was maintained. Colonial society itself was hierarchical, authoritarian, and directed toward reason (as it was then understood). It was not only children but all persons who expected control from their surrounding community. In that sense there was a rea: parallel between the control of young children and the control of adsults. The child who matured then was not selfcontrolling in the same way we think of ourselves as, or expect our children to become, self-controlling (cf. Beales, 1975; Demos & Demos, 1969).

IV. Control in the Jacksonian Era More notable than the changes in the contents of Jacksonian child-rearing practices are the changes in tone and sources of child-rearing advice. Prior to the nineteenth century, in Ameriica, almost all writing on children was done by men who were ministers. The nixteenth century presents us with a great proliferation of child-rearing manuals, a good many of them secular, and many of them by women. There is an entirely different tone to the nineteenth century tracts and it reflects a changed conception of childhood as well as different notions as to how children should be raised (K.ett, 1971; Strickland, 1973; Wishy, 1968). Thus in the Jacksonian era, the introduction of “mother’s books” strongly affected the type of advice on socialization and control of children available to mothers. Through them, the models for child rearing and family interaction were de-theologized and democratized, expressed in political and more democratic metaphors (Mergen, 1975). The growing importance dluring the 1800s of children’s welfare necessitated a different approach to their control and discipline-one more gentle, more situational, and more careful of thle child’s self-image. Control became environmental or atmospheric as well as personal, a recognition of the silent influences of the child’s surroundings. His environment could be arranged and manipulated to furnish a Christian atmosphere in which aggression and disobedience would be less likely to occur. For children as well as adults, the concept of “moral architecture” became popular, reflecting widespread belief that the architecture and structure of the environment bear a direct relationship to actions occurring within it. Since children were no longer seen as heavily tainted with the evil of original sin, parents were advised to concentrate on preventing the Occurrence of unwanted behavior, rather than solely punishing wrongdoing. Contained within and reinforcing this new ease in disciplining children was the discovery of a “natural” potential in children. The Transcendentalists believed that proper nurturance of this potential could produce previously unimagined

Child Discipline and the Pursuit of Serf

239

levels of personal development. Still, by contemporary standards the proscribed ideals of child rearing which emerged early in the nineteenth century do not seem completely different from those which dominated the colonial period and they certainly d o not appear completely permissive. It is somewhat difficult, then, to understand what all the fuss was about. How could European and American observers have spoken so harshly about the nature and behavior of children in the United States? (Rapson, 1965). How could there have been so widespread a sense of alarm about the disintegration of the family and the dissolution of parental authority? The differences lie perhaps not so much in changes in what people were actually doing to and with their children but rather in the changed context in which the family found itself and the uncertainty people felt about the possible efficacy of the child-rearing practices available to them. How can we know this was the case? Why would the lifting, or even the softening, of controls create such panic? There are indications that, for the new urban dwellers and middle-class, there was a general sense of deep personal confusion. Removal of the constraints of traditional culture left people alone, so it seemed, to construct and maintain their identities. Indeed the construction of identity became a personal project. People. no longer knew who they were simply from knowing where and to whom they were born. Such persons uncertain about who they should or could become turned eagerly toward any institutions or techniques that promised assistance in the task of identity construction and maintenance. This interest is reflected in the fascination with techniques and methods of self-control and self-development which is found in almost all spheres of Jacksonian life--schools, prisons, asylums, boarding houses, the sex-reform movement, the communal movement, special diets, phrenology, moral architecture, etc. (Brown, 1974; Clark, 1976; Strong, 1973). These concerns, of course, were reflected in a special way in the rearing of children, especially in middle-class families who, in addition to the uncertainties and confusions facing most newcomers to the urban scene, had the additional burden of the traditional desire to pass on to the next generation what they owned (see Shorter for an extended discussion of this). We must remind ourselves what socialization is about to fully understand the apprehension that lurks behind the rhetoric of children’s welfare in the nineteenth century discussions of child rearing. Inkeles (1969) puts it this way: “The critical test of the socialization process lies in the ability of the individual to perform in the statuses-that is, to play the roles-in which he may later find himself” (p. 616). In traditional societies socialization centers around the passing on of relatively stable elements of culture; the child must become like the adults who raise him and individuality, self-discovery, and personal choice, as we know them, make little sense in such systems. In the nineteenth century, with the decay of traditional society, most parents were confronting, without long-established cultural supports, a relatively unique circumstance-the inappropriateness of traditional adult roles and therefore traditional modes of child rearing. In modern

240

Howard Gadlin

societies, children need different skills, attitudes, and characters than their parents and parents cannot blitheby raise their children to be like themselves (Inkeles, 1955). Indeed as modernization proceeds there are more people who simply do not know what kind of world their children will have to face, there is no room for persons who will replicate the past, obsolescence is built into all aspects of people’s lives-their personality traits and values as well as their skills and commodities. Rather it becomles necessary to produce individuals who will create new familial and cultural patterns. As Seeley, Sim, and Loosley (1956) have pointed out, parents and children must learn to accept new forms of social behavior unlike those previously known and practiced. This requires not only different personality types but, as Kluckhohn (1952) observed, “the kind of family which permits individualistic expression and allows its members to go free of bonds that would tie them to particular people and places.” In such conditions, where one cannot know how to treat one’s child simply by seeing the society and knowing one’s location in it there are two other places to look-at the nature of the child and at the techniques and procedures which promise to help the child face an essentially unknowable future. It is in Jacksonian America that this search blegins to become a cultural obsession. And it is the family, its members increasingly alone with one another, often in regions far removed from those of their kin, and frequently in communities of strangers for whom neighborhood has only geographical meaning, that this obsession is developed into a ruison d’etre. All this is not to suggest, of course, that nineteenth century parents simply and gleefully tossed aside traditional notions about children and their care. Quite the contrary, much of the passion that infuses these discussions about children indicates quite the opposite. Writing about Britain, Harrison (1961) has noted: “Traditional social habits and customs seldom fitted into patterns of industrial life, and they had. . . to be discredited as hindrances to progress” (cited in Gutman, 1973, p. 541). If anything the: contrasts between traditional and modem life and the pressures on traditional family styles must have been greater in the United States than in Europe. In the United States, supports for traditions were much weaker, while immigration amd westward migration combined with the complexities of urbanization and >industrializationto undermine traditional ways (cf. Calhoun, A. W., 1960; Wiebe, 1969).

V. The Ideology of the Family But the nineteenth century fascination with the nuances of childhood and child rearing must reflect more than the growing uncertainty of parents as to what to do with their children. The interest in childhood functions also in an ideological manner serving as a benediction for the changes which had occurred in family

Child Discipline and the Pursuit of Self

24 I

life. The separation between public and private life means, in Berger’s (1965) words, that society’s members: experience this dichotomization as a fundamental ordering principle of their everyday life. Identity itself than tends to be dichotomized, at the very least, in terms of a public and private self. . . . The typical option has been. . . to assign priority to their private selves, that is to locate the “real me” in the private sphere of life. If the “real me” is to be located in the private sphere, then the uctivities of this sphere must be legitimated as decisive occasions for self-discovery. (pp. 36-37)

The discovery of children’s unique nature and the concern with their welfare and upbringing is then part of that legitimation. But in legitimizing this focus on childhood, the very split between public and private institutional spheres is sanctioned. In sum, the discovery of the importance of childhood, of the rights of children, of their natural capacities, and of the significance of the methods by which children are raised reflects the confusion of parents, serves to assuage that confusion and, simultaneously, functions as the legitimation of the very changed circumstances which provide the context for the new sets of problems with which parents are confronted. There is another aspect to the changed nature of family life that both further reveals the ideological meaning of the new understanding of childhood and, at the same time, uncovers another important dimension of the problem of discipline and child rearing in modern society-one that remains with us to this day. With the sundering of private from public life, as Lasch (1975a) has observed, the supposed satisfactions of private and family life are held out in compensation for the deprivations and dissatisfactions suffered in the realm of work and with the impoverishment of community life. Early in Jacksonian America the ideal of the home as a refuge took hold. Jeffrey (1972), looking at the nineteenth century home as analogous to a utopian retreat, has remarked: Our ancestors thus were encouraged to nurse extravagant hopes for the domestic realm. Whether they regarded home as an utter and permanent retreat from life in a shocking and incomprehensible social order, or as a nursery and school for preparing regenerate individuals who would go forth to remake American society, they agreed that domestic life ought to be perfect and could be made so. (p. 22)

These words from Eliot’s Lectures to Young Women (1855) well represent the middle-class spirit of the time: The foundation of our free institutions is in our love, as a people, for OUT homes. The strength of our country is found, not in the declaration that all men are free and equal, but in the quiet influence of the fireside, the bonds which unite the family circle. . . . From the corroding cares of business, from the hard toil and frequent disappointmen of the day, men retreat to the bosom of their families, and there, in the midst of that sweet society of wife and children and friends,

242

H o word Gntllr II

receive a rich reward for their industry, and are reminded that tneir best interests are inseparable from public and social morality., . . (cited in Jeffrey, 1972, p. 21)

In the popular literature numerous passages speak eloquently to the painfulness of urban life and the needs people brought to their increasingly private home lives. Of course even disregarding the sentimentality with which home life was extolled it was an unachievable idea(cf. Model1 & Hareven, 1973), and in almost the same breath that the family was praised as the locus of personal satisfaction, we must remember, it was also blamed as the point of origin for crime, delinquency, insanity, and almost ,all other forms of deviance. So began a trend which continues into the present. But especially significant for our purposes here are the implications of the way in which the split between the public and the private came to be understood in terms of an antagonistic opposition between them. In a modern society, with both personal identity and the social order divided into public and private real realms, and both public and private realms seen at least partly i n opposition to each other, the task of socializing the young becomes highly complex and contradictory. As we have noted, some of the complexity is because it is impossible to anticipate the kind of world for which they are being trained, the skills they will find useful, and the personal traits which will be desirable. But in addition, if the realms for which children must be prepared are understood as essentially antagonistic, then the socialization practices which each realm requires may be incompatible, or parents must become highly ambivalent about their role in the entire process of socialization, or both. Put most crudely, what it takes to turn out a child who will be publicly successful may be quite different from what it takes to produce a child who will be personally happy and fulfilled. Of course the easy way out is to create an individual who finds personal fulfillment in public success. But as Berger (1965) has noted, “this option is not very seductive for the great masses of people in the middle and lower echelons of the occupational system’’ (p. 36). Indeed, given the historical record, and considering the nature of work under industrial capitalism (see Braverman, 1974, for a brilliant discussion of this issue), this option is not even a possibility. Another alternative could be to develop people who found fulfillment i n comlmunity life but, as we have seen, this option too is highly unlikely given the waiy i n which traditional society breaks down under modernization-it is community life which disappears. As we have seen, these were not the paths along which American society developed, and the public and private worlds offered very different satisfactions. And clearly the characteristics necessary for success in the world of workcompetitiveness, distrust, otisessiveness, etc.-were not those most likely to guarantee a warm and loving home environment of the type eulogized in the popular and professional literatures of the time. The incompatibility of the two realms manifests itself in mainy areas of life in the Jacksonian era. One of the

Child Discipline nnd the Pursuit of Self

243

most striking is the sexual segregation which developed almost to a caricature in the nineteenth century middle-class home (Welter, 1966). In the child-rearing literature it takes the form of a differentiation between the acculturating and the actualizing functions of child rearing. I shall use actualizing to refer to the fostering or nurturance of the individual characteristics of the child which are understood, typically, in terms of the unique capacities of the child which are requisite to personal development. By contrast, acculturating will refer to the training or nurturance of those traits which are understood in terms of cultural values and adult roles requisite to the maintenance and continuation of the society. Obviously, acculturation and actualization are inseparable from each other and the seeming opposition between the two reflects a particular cultural formation in our society.6 In other words, I use the terms to refer to categories within which people in American society have come to understand and evaluate their child-rearing practices. The separation of the two dimensions of child rearing is fairly mild and incomplete at first, but it is exacerbated over time to the point where at present some of our writing sounds as if actualization and acculturation were incompatible. The development of this alleged incompatibility can be traced in the changes in our notions of child discipline since the Jacksonian period. One quote from Emerson illustrates nicely the way in which the American understanding of individual independence creates problems for any socializing scheme we might imagine: “Let us feel if we will the absolute insulation of men.. . . Let us even bid our dearest friends farewell, and defy them, saying ‘Who are you? lJnhand me: I will be dependent no more’ (cited in Bridges, 1965, p. 7). [Compare Watson (1928) when he concludes “We have tried to sketch . . . a child as free as possible of sensitivity to people and one who, almost from birth, is rclatively independent of the family situation” (pp. 186-187).] Speaking of the same phenomena, Jeffrey has remarked: ”

Even when they affirmed that the family ought to concern itself with moral training of its members, they still emphasized that this would occur through gentle, loving techniques. Thus the ends of family life were emphatically individualistic, libertarian, even anarchic. In the perfected American home. . . individual freedom. . . could at last be obtaineGand without a corresponding increase in social disorder. (p. 37)

Of course it did mean a decrease in social order, if nowhere else than within the family. It is hard to imagine that families which shared even partially Emerson’s ideal could achieve the kind of loving closeness and harmony which was simultaneously held up as an aspiration of the family. And it is difficult, too, to envision a family inspired by that ideal producing children who could be understood as obedient and respectful in the traditional sense of those terms. ‘Of course, to recognize their inseparability does not mean to imply that both functions are fully and adequately contained within each other.

244

Howard Gadlin

Now before we go into the last quarter of the nineteenth century and on into the present, it will be helpful to review the contradictory position of the family that resulted from the changes in the meaning of family life that had occurred during the early years of modernization. As the public sphere developed into a competitive and, so it seemed, immoral world filled with self-contained, heartless individuals, the family was idealized as a private sanctuary of warimth, intimacy, and love and as the breeding ground for morality and individualism. The family was thus caught in a bind through its need to produce persons who had some chances of success in that public world toward which the family was, at best, ambivalent and, more typically, fearful and critical. As Lasch (1976) notes, the family had to produce a person who could “live in precisely the cold and ugly world from which the family provides a haven” (p. 52) (cf. Walkowitz, 1972). During this time, both the public and private spheres do converge on at least one point-the preeminent value of individualism and the importance of selfdevelopment. Now we must be especially careful here because the notion of the self has itself been changing, so self-development means today something very different than in Jacksonian American or at the turn of the century. To briefly characterize the changes in emphasis from then till now, we might say we have moved from self-discovery and self-discipline in Jacksonian America, toward self-knowledge and the development of the productive self at the turn of the century, and toward self-expression, self-fulfillment, and selfactualization today. (We must remember these are emphases, not absolute categories .) Over these years, then, I believe we find an increasing tendency to justify or rationalize techniques of socialization in terms of their putative value in helping to actualize the potential of the child as that potential can be ascertained in terms of the prevailing notions of the child’s nature. This has become the American way-the tension between the public and the private is translated, in our considerations of child rearing, into a diminishing emphasis on the culture-transmitting, role-preparing features of socialization. In any event, if we look mlDre closely at the changes in disciplinary tactics and ideals from the late nineteenth century to the present, we can examine the conflict between the acculturating and actualizing dimensions of child-rearing practices as they vary with transformed notions of selfhood and as these, in turn, change in coordination with other social changes (cf. Twombly, 1975).

VI.

Child Rearing in the Modern Era

There is one major exception to the general tendency of looser disciplinary control over young children which must be considered in any examination of modern socialization practices. It appears that modern parents “get to” their

Child Discipline ond rhe Pursuit of Self

245

children much earlier, more thoroughly, and even more fundamentally than did their traditional predecessors. In earlier periods the object of control was the child’s disobedient and aggressive expressions. This holds even for the initial half-century of the modern era. Demos (1970) suggests that the first two years of a colonial child’s life were years of relative indulgence (cf. Whiting & Child, 1953). Early in the modem era, however, the object of parents’ attention shifts somewhat, and they become concerned with the bodily functions of infants. Parental control, then, is redirected to the child’s most fundamental relationship to the impulses and gratifications of his or her own body and, therefore, to the child’s most fundamental sense of self. In the nineteenth century there was movement toward earlier toilet training, weaning, prohibitions of infant sexuality, a decline in demand feeding, and disapproval of cuddling and physical indulgence of infants (Danzinger, 1971). Thus, at the time when parents were moving toward relative indulgence of children’s aggression and disobedience, they were simultaneously becoming less indulgent of their infants. Neuman (1975), in a study of the relationships between concepts of childhood and parental attitudes toward masturbation, demonstrates how this stricter control of infants was perfectly compatible with the concern with childhood and children’s welfare that developed in the nineteenth century: “The emergence of a concept of childhood based upon the innocence and weakness of the child required that parents deny the reality of infantile and child sexuality and at the same time prescribed long periods of supervision and control of the child at home and school” (p. 19). And while the tendency i n the late nineteenth and early twentieth century literature is toward less severe child discipline, the trend in the infant care literature is toward stricter impulse control. Three independently conducted historical surveys in the 1950s (Stendler, 1950; Vincent, 1951; Wolfenstein, 1953; see also Miller & Swanson, 1958; Borstelmann, 1976a & b) demonstrate that it is not until late into the 1920s and early in the 1930s that we find a movement toward a more indulgent treatment of infant sensuality.’ And it is not until the 1940s that this movement becomes a clear-cut trend dominating the treatment of most middle-class children. These trends are illustrated in Table 1 from Wolfenstein’s study of Infant Cure, an exceedingly popular advice publication of the United States Children’s Bureau, and Table I1 from Stendler’s review of three women’s magazines between 1890 and 1948. It is especially important to note that even in those periods where impulse control is fairly strict, the justification is given either in terms of the appropriateness of strict control to the child’s nature and needs, or in terms of the kind of adult restrictions which will be imposed on the child’s basic nature as he or she ’1 will use infant sensuality to refer to the sexual, oral, and elimination activities of infants.

246

Howard Gadlin

TABLE I “ Trends in Infant Care ~

Severity in the handling of

1914-1921

1921-1929

1929-1938

1938 to 194245

1942-1945 to 1951

Masturbation Thumb-sucking Weaning Bowel training Bladder training

Decreases Constant Increases Increases Increases

Decreases Decreases Incrcases Increases Decreases

Constant Constant Constant Decreases Decreases

Decreases Decreases Decreases Decreases Decreases

Constant Decreases Constant Decreases Decreases

nFrom Wolfenstein (1953, p. 12711.

matures. Rarely are such discussions informed in any serious way by a shared sense of what would make community life both possible and desimble. (Indeed, one could almost say that the !iocial consequences, or at least what seem to be the social consequences, of our child-rearing practices always appear to come as a surprise to each generation of parents when it first gets a sense of the collective effects of the means by which they have reared their children.) TABLE I1 Percentage of Methods Recommeinded for Two Aspects of Child Training as They Appeared in Three Women’s Magazines Analyzed in 10-Year lntervals for 1890-1948” ~

1890

Discipline (reward a n d or punishment Provide a good home influence Ignore undesirable behavior Look for cause and plan accordingly Invoke divine aid Feed properly Miscellaneous

1900

1910

1920

1930

1. Guiding Character or Personality Develooment 34 38 18 14 34

Tightly schedule, cry it out Loosely schedule Self-regulate, “mother” “From Stendler ( 1950).

1940

1948

28

2

61

53

30

12

14

5

3

12

5

9

0

12

18

4

0

0

0

I

14

48

84

15 0 4

20 0 5

15 0 12

0

0

0

0

50 3

0 21

0

2

0 8

100

75

33

0

0 0

0 25

0 66

0 100

0

11. Infant Disciplines 22 71

100

78

0

0

23 0

Child Disciplitie und the Pursuif of Self

247

At any rate, the rationales for strict control of infants, like those later rationales for more indulgent control of infants, justify strict control in terms of the actualization of the child’s potential. It is of particular interest to follow the changing contents of those justifications for the treatment of infants because they reveal, among other things, the role played by the emergence early in the twentieth century of professional and scientific disciplines that were concerned with the study of child development. [See Sears (1975), for a thorough, if somewhat congratulatory history of the child development movement.] Throughout the nineteenth century and well into the 1920s parents justified strict control of infants in terms of the need to develop persons with moral character. Of the early twentieth century, Miller and Swanson (1958) observe: Middle-class Americans shifted from nursing at the breast to bottle-feeding. At the Same time that they were giving children more freedom for self-determination. Similariy, they seem to have accepted the idea that corporal punishment was to be minimized to avoid crushing the child’s spirit and independence. The use of rigid schedules of feeding appears to have risen after these other developmentsprobably reaching its peak of acceptance during the 1920‘s and the early 1930’s. (p. 27)

Lomax (1975) points out that a concern with physical health as well as psychological considerations underlay the interest in scheduling. Parents, she remarks, “. . . were being told that the essentials of effective infant care were minute attention to hygiene, clockwork regularity in feeding, bathing and putting the baby to sleep, fortified by the introduction and training of good habits at the earliest possible opportunity” (pp. 2-3). Strict infant control was also justified by reference to the role of bodily discipline and regimentation in producing the independence nzcessary for success in the public world. Miller and Swanson (1958) remind us of Watson’s understanding of the benefits of beginning urinary control between the third and fifth week: “The end result is a happy child free as air because he has mastered the stupidly simple demands society makes upon him.” Watson then continued: “. . . An independent child because all during his training you made him play and work alone a part of the time, and you have made him get out of difficulties by his own efforts. The only person in life who is effectively original is the person who has a routine and has mastered a technique” (cited in Miller & Swanson, 1958, p. 18). The growth of the child development field changed the terms within which parents understood their treatment of infants. and eventually contributed, as well, to a change in the actual treatment of those infants. Stendler (1950) describes this change very nicely: The mother of 1890 was chiefly concerned with the development of good moral character; the term “well-adjusted personality” was unknown in the literature of her day. And not such vaguely defined products as “well-adjusted” and “secure” children were sought but rather

Howard Gndlin

248

child-training practices were designed to produce children who would exemplify the Victorian ideals of curtesy, honesty, orderliness, industriousness, and generosity; character, not personality, was the focal point. (p. 125)

With the switch from character to personality, literature on child rearing, whether advising strict or indulgent treatment of infants, tends to explain the recommended behaviors in terms of their importance in relation to the child’s adjustment, mental health, and personality development.

VII.

Child Rearing in the Modern Era-Child Discipline

Thus we return to our consideration of the changes in child discipline bringing with us a new participant i n the process of historical transformation-the “sciences” of psychology and child development. These disciplines fit quite well into the prevailing cultural emphasis on individual development. The words of the psychologist Judd (1907) express this notion quite clearly: “The individual must seek of his own initiative those higher forms of organization which will most fully realize the possibilities of his life. The highest level of individual organization is reached when mental development becomes a matter of voluntary control” (p. 329). Such an emphasis was quite compatible with the kind of advice that had begun to appear late in the nineteenth century emphasizing the importance of a parent’s re1,ationship with the child rather than control over the child and questioning the importance of obedience as a means to producing autonomous adults. Thus Wiggin (1899) could write: “Blind obedience is not in itself moral . . . our task is to train responsible, self-directing agents, not to make soldiers” (p. 122). Consequently she could recommend, rather than demanding the child’s unquestioning acceptance of adult authority, the usefulness of explanations geared to what the adult feels the child is capable of understanding: “What possible harm can there be in sometimes giving reasons for commands, when they are such as the child would appreciate” (p. 163). Associated with the addition of the scientific perspective was an unusual optimism about the possibilities available to parents and society through the use of child-rearing practices properly inspired by psychology. But the ongoing transformations wrought by the social and economic development which had created the conditions for the emergence of persons characterized by firm internalized standards of right and wrong and a solidly aggressive independence soon led to the social inappropriaiteness of such persons. We need not dwell on the details of this transformation-they were written about with considerable insight in the 1950s, first by Riesman (1950)’ and again sMuch of this analysis was developed in Gadlin and Rubin (1978).

Child Discipline

tirid

the

Pursuit of Self'

249

by Miller and Swanson (1958). Although there were points of disagreement between them (Miller & Swanson, pp. 23-24, 109-119), they are in basic agreement about the outlines of the changes which occur, roughly in the course of the second quarter of the twentieth century (for some critical perspectives, see Degler, 1963; Lipset & Lowenthal, 1961). Riesman describes these changes as a transition from inner-directed to other-directed character types. By contrast with the inner-directed type, the other-directed person acts according to goals set in accordance with his reading of the expectations of contemporaries. Riesman locates the reasons for this change in character type in the changed economic and social circumstances associated with the ongoing developments of capitalism. Riesman (iY50) traces these changes to ". . . a decline in the numbers and in the proportion of the working population engaged in production and extraction-agriculture, heavy industry, heavy transport-and an increase i n the numbers and the proportion engaged in white-collar work and the service trades" (P. 20). Riesman goes on to argue that in such conditions social mobility depends essentially on what others think of a person rather than on who one is and what one does. Consequently fixed traits, which used to be called character, are now indicators of a rigidity which keeps the person from adequately adjusting his or her demeanor to the requisites of the situation. In these circumstances, Riesman argues, parents who aspire to create children with any chance of success in the public world must become parents who can produce people who can adjust to both situations and other persons. Riesman (1950) also observes how this new level of parental uncertainty about how to bring up their children contributes to a heightened interest in the advice of experts and, even more than before, to an attempted understanding of the child's nature and needs (p. 47). Early in the nineteenth century, although parents did not know what kind of persons their children would have to become, they knew, at least that their children would need to stand alone, making a place for themselves in a dangerous world. Knowing the corrupting nature of that world, parents also knew what values and traits would help their children to survive, and perhaps thrive. In principle, the home could provide the nurturance lacking in the world and prepare the child to d o without it. However, with newly emerging conditions presenting a situational diversity and interpersonal complexity previously unheard of, steadfastness could be more of a liability than an asset. Rather a child would need a much more diffuse and vaguely defined ability to enter into and manage situations and relationships. As the child-rearing situation continued to change, it became increasingly clear that it was not character and values that the child would have to bring with him when he left home, but the very parent-child relationship itself. In a trend that continues right up until the present day, the parent-child relationship thus becomes not just a means to an end, but an end unto itself. Indeed, we could almost say that the child becomes not merely the product of particular child-rearing practices, but that, mediated by

250

Howard Gadlin

the processes of development, the person the child becomes is a transformation of the parentxhild relationship. Thus new focus on the parent-child relationship is nicely illustrated in Miller and Swanson’s interpretation of the changes Riesman attempted to characterize. Like Riesman, Miller and Swanson turn to American society’s economic structure to understand the changes in personality type and child-rearing procedures which they observe. They describe these changes as a shift from individuatedmtrepreneurial to welfare-bureaucratic types. Entrepreneurial children, having been taught self-control, addressed the world from a manipulative stance. The bureaucratic child had been urged to be accommodative in stance and self-expressive in behavior (Miller & Swanson, 1958, pp. 57-58). They see the welfare-bureaucratic type developing throughout the twentieth century and really coming to the fore in the 1940s. In a highly suggestive analysis, they follow the changes in the language of employer-employee relations to make sense of the changes in parent-child discipline. Most briefly, they describe the change as a movement, inspired by Taylorism, from a concern with worker discipline to a concern with worker morale, and a shaping of the job to suit the worker so that the lloyalty of the worker is enlisted by the boss. The parallels between Miller and Swanson’s descriptions of the boss-worker relationship and the equivalent discussions of parent-child relationships are striking. And as with management, the emphasis in parental dealings with young children was on the relationship. Miller and Swanson (19%) emphasize the importance of the quality of the relationship: This kind of loyalty and devotion. . . represents the experience of a fundamentally moral relationship. Morals. . , are the code of social rules that grow up to preserve a situation in which people find each other’s presence to be so mutually rewarding and, simultaneously so lacking in threat, that they feel wholly comfortable and spontaneous and seek to preserve their happy and productive state. The problem of management is to establish such a moral relationship with its workers without losing its authority over them. There has been a persistent and analogous problem for parents in ,dealing with children-how to be both authority and benefactor. . . . The management solution, like that of many parents, is one of seeking avidly for benefits it can give without jeopardizing control. , , , The benefits, not the planning, must be the conspicuous thing in the worker’s experience. He must feel that what was done occurred because management was genuinely interested in his welfare, not because the benefits would result in higher productivity. . . . (pp. 53-54)

A remarkable description, the definition of morals is itself as splendid an illustration of the ascendancy of relationships as are the descriptions of the changes they observe. And the terms in which even they describe, quite sympathetically, the welfare-bureaucratic type of child rearing are replete with concern with the same issues that appear in the discussion of labor relations. Morale, loyalty, and satisfaction come to be the characteristics of children that are idealized (cf. Kiefer, 1948).

Child Discipline and the Pursuit of Self

25 1

Through the first 40 years of the twentieth century we can follow in the literature on child discipline first an emphasis on indirect control of the child through manipulation of his environment and then, increasingly, a focus on the nature of the parent-child relationship. At the turn of the century adults were advised to be careful to create an atmosphere, a context, in which a child could “freely choose,” but could make only certain kinds of choices. Wiggin advised parents of a young child to “surround him with the best and worthiest books, and let him choose for himself;” to provide the right conditions for mental growth and then let the child do the growing and to “ choose his toys wisely and then leave him alone with them” (Wiggin, 1899, pp. 56,83-84). Thus a child was to enjoy a power of choice in which not he himself, but his environment was directly manipulated by parents and teachers. Even when fairly severe restrictions on infant sensuality are being recommended, there is a tendency to promote techniques that emphasize the structuring of the environment so as to make the undesired behaviors less likely, rather than any form of punishment. Wolfenstein’s (1951) review of the development of the “fun morality” traces the changes in the kind of environmental manipulations suggested in response to childhood masturbation, from mechanical restraints in the 1914 edition oflnfunt Cure to the calm advice in the 1942-1945 editions: “A wise mother will not be concerned with this. . . . See that he has a toy to play with and he will not need to use his body as a plaything” (p. 18). Assuming even more importance in the course of the twentieth century discussions of child rearing was the psychological quality of the parent-child relationship. At first the emphasis is on the way in which a close parent-child relationship will enhance the parents’ influence on and, hence, control over the child. Turn-of-the-century literature considered the parent-child relationship important because the parent would serve as a model for the child. But as the century proceeds there is more and more attention given to the sense of self the child will presumably obtain from the relationships with the parents. Miller and Swanson (1958) cite from the 1954 Spock: We know for a fact that the natural loving care that kindly parents give to their children is a hundred times more valuable than their knowing how to pin a diaper on just right, or making a formula expertly. Everytime you pick your baby u p . . . everytime you change him, bathe him, feed him, smileat him, he’s getting a feeling that he belongs to you and that you belong to him. (p. 24)

And in their own words, too, this theme is repeated: . . . a successful future requires that child to have confidence in himself and to be confident of others’ regard for him. Sympathy in infancy is believed to build such confidence. Second, the child, and the parents as well, are thought healthier if they allow themselves the expression of their feelings and the satisfaction of sensuous and loving appetites when this is possible without serious consequences. (p. 25)

252

Howard Gadlin

In this brief passage several of the major shifts in emphasis in the literature of child rearing are captured-the predominant importance of the parentxhild relationship; the central importance of pleasure to that relationship; and the overriding concern with mental health as the underlying goal of child rearing. Wolfenstein’s (195 1) “fun morality” paper also pointed to the new importance attributed to pleasure i n the mid-century editions of Infant Care. By 1951, parents were being advised: “They are making a good start if they can enjoy their baby. The child should learn that mother and father are ‘two people who enjoy each other.’ Introducing the baby to solid foods will be ‘fun and amusing’ for the mother and the baby will ‘enjoy the new experience more if you are having a good time”’ (p. 176). For the most part the fun morality continues in today’s literature on child rearing, having permeated even the writings of the champions of behavior control. Thus, the cover of 7oilet Training in Less Than a Day (Azrin & Foxx, 1974) proclaims the book “ a professionally tested new method for successful toilet training in one pleasant and exciting learning period.” And one of the criticisms they level against traditional toilet training activities is the condemnation “old training methods aren’t always fun” (p. 22). The central role played by pleasure both within relationships and in the quality of mental health reflects a new emphasis in the conceptions of self that develop during the modern era; the preeminence of self-fulfillment and self-expression. However, it is of utmost importance to notice that these new developments are best understood as emerging in the context of changes in the public sphere, specifically changes in occupational structure and social organization. The concern with pleasure relates as well to another feature noted in Riesman’s analysis of the newly emerged other-directed character type-consumption orientationthe valuing of the ability to live and enjoy in the present rather than postponing pleasure until the completion of work. A modern, industrialized, capitalist economy needing to grow to maintain profits must shift its emphasis from production to consumption. This shift, in turn, requires a legitimation of pleasure and it is of interest that the emphasis on consumption, beginning in the 1920s, accompanies the relaxation of controls and regimentation of impulse gratification during infancy. Thus there are several changes in the public sphere which converge to make more rational both a continuation of the trend to make the discipline of children less harsh and a reversal of the trend to make the treatment of infants more severe. It is particularly interesting that throughout the various changes and fads in child rearing, there is a continued, indeed an increased insistence upon the actualizing functions of child rearing. Each new change is presented as if it were based on an improved understanding of the child’s nature and needs. But now it is not only techniques of treat.ing and disciplining children which must be geared to what is assumed to be the child’s nature and needs, but the totality of the parents’ relationship to the child (Jones, Stewart, & Winter, 1974).

In such families, the goal of discipline has to be not obedience, nor even moral behavior, but sensitivity to others. With interpersonal sensitivity the goal, parents must be friends, not authorities. When permissiveness was at its peak in the 1940s and early 1950s it appeared as if people expected children to develop that sensitivity if they were reared by parents who displayed such sensitivity toward their children, but demanded little else. This project was doomed to failure and the mid- 1950s saw a resurgence of interest in the placing of limits and expectations on children. Both parents and researchers were discovering that the constellation of child-rearing practices that was labeled child-centered produced children very different than those hoped for. For example, Hoffman (1963) reported that: The following parental behaviors were expected to relate positively to the child’s consideration for others: acceptance of the child, discipline which highlights the consequences of the child’s behavior, and discipline explicitly oriented to the needs of others. Acceptance was also expected to relate positively to the child’s general affective orientation or friendliness.. . . The only hypothesis initially receiving support was the subsidiary one that parental acceptance relates to a positive affective orientation. (p. 586)

Of course, the failure of modes of child rearing should lead us to suspect something larger than a technical problem. As we have seen, changes in selfhood and society usually lie behind our recognition of the weakness in particular practices. It is with that i n mind that we should examine the contemporary period.

VIII. Personhood and Child Discipline in Contemporary America The most important characteristic of contemporary child rearing is the continued diminution of parental authority and responsibility. The trend began, of course, in the late nineteenth century but it becomes increasingly the case that there is less and less that parents can pass on to their children, with any certainty that it will help them in the future. Throughout the twentieth century the increased importance of the schools as socializing agents as well as the increased infiltration into the home of other institutions such as media, professional experts, etc., serve to further shrink the domain which parents can consider their own. It is within this context that the emotional and psychological quality of the relationship between parent and child comes to the foreground. Ironically, the focus on the relationship, in some ways, serves to heighten the parents’ felt responsibility for the way the child turns out at the same time that the parent has less control over the direction of the child’s development. Most striking in the contemporary literature is the detailed examination of all aspects of the parent-child relationship. Behavior, feelings, attitudes, postures,

254

Howard Gadlin

motives, explanations offered, and the parent-parent relationship are all examined for their possible consequences on subsequent child behavior, feelings, etc. Some of this represents the increased sophistication of research and researchers. But increased sophistication in understanding a phenomenon usually reflects an increased prominence of features of that phenomenon which previously had been of lesser significance (cf. Klapper, 1971; White & Watts, 1971). Making sense of the present is an especially dangerous business. Writing in 1958, Miller and Swanson (1958) speculated that they were in the beginning of a new age in which the companionship family would be replaced by a collegial family. These families would be characterized by the deep satisfaction that men and women would find in the specialization appropriate to their sexes; women especially would become recommitted to home life because of the increasing “professionalization of the wife’s functions” (p. 201). As for the children, they were to be the junior members of their parents’ partnership: It is our impression that parents in the bureaucratized family, like those in the older agricultural family, find children a fulfillment, without many of the difficulties that went with the raising of a family under entrepreneurial conditions. For the wife they are a necessary canvas on which professionalized skills of homem.iking can be expressed. For both husband and wife they are an outlet for creative management of social relations, for the parents’ learning new things about themselves, and for demonstrating their conventionality and adaptability and “maturity” in the bureaucratic world. If the family is secure in the new sense that the children can find employment near the home of the parents and, with swift transportation and communication, can be part of the parental families for a long time to come, then there is also a renewed sense of self-continuity and self-realization through children that comes to the couple with a family. (p.

204)

The events of the next 10 years proved each of these impressions erroneous. Still there must be limits to the lessons we can learn from history, and I would like to venture some speculations about current changes in personhood and in the family, particularly as they relate to changes in the disciplining of young children. Several reviews of cunent child disciplinary practices have already been done with admirable thoroughness elsewhere, and I will draw from them somewhat freely (see e.g., Baunlrind, 1971, 1973; Becker, 1964; Clarke-Stewart, 1973; Hoffman & Saltzstein, 1967; Parke, 1974; Staub, 1975). It is apparent from the literature that there coexists in our society a multiplicity of child-rearing styles. It is hard to know whether this diversity reflects the current “absence of professional singlemindedness” (Kessen, 1965), an increased diversity over past times, or simply the fact that the actual child-rearing practices of parents were never so thoroughly researched before, and therefore we overestimated homogeneiity in the past. Probably some of each is true. I will focus here on those practices which seem to reflect the characteristics of families we might describe as moving toward postmodern (Shorter, 1975, pp. 269-280).

Child Discipline

and the Pursuit of Self

255

There seems to be a general consensus that we are in the midst of another transformation of the family (Howe, 1972; Shorter, 1975). Although in many ways the features noted by Riesman, Miller and Swanson, and others in the 1950s are still pertinent to a description of the contemporary family, the meaning of these features seems to have been turned on its head; what were interpreted as virtues in the 1950s (e.g., Miller and Swanson’s predictions above) were, by 1970, understood as vices. Others have noticed the collegial nature of family relations (Weinstein & Platt, 1969, p. 221, 297), but the idea of the couple as partners has developed to the point where the partnership is understood as involving a commitment which shall last so long as each party finds the association personally satisfactory [cf. Rogers’ (1972) Becoming Partners; O’Neill & O’Neill (1972) Open Marriage, etc.]. This has meant, within the family, a more intense concentration on selfactualization, only now there an expectation that the family be the unit for the self-actualization of the parents as well as the children. On the surface, selfactualization sounds like a nice thing, something quite alright for everyone to experience, but we need to examine the changing notion of the self which is to be actualized to see what it might mean for the family and for children. Connected with the changes in the family are the beginnings of changes in the kinds of selves people can be (cf. Clauson, 1966). In the nineteenth century, inner-directedness resulted from the internalization of cultural ideals such as self-control. Perhaps what we are experiencing today is something like the internalization of other-directedness (Gadlin and Rubin, 1978). Riesman had observed how the other-directed person could only evaluate his own qualities and acts in terms of their effects on others. Miller and Swanson (1958) noted how the pressure to play to the approval of many meant difficulty in sustaining close comradeship with even a few. Such a person “. . . must learn to produce a relationship that uses the symbols of genuine friendship as its currency without the actual commitment of the real thing” (pp. 202-203). In many ways, other-directedness cannot survive in a fundamentally competitive system that emphasizes an individuality which is atomistically autonomous and asocial. Too little support from others is forthcoming and what approval one can obtain is often invalidated by the very strength of one’s need for it, and by one’s own consciousness of a split between one’s private and public self. In response to these dilemmas the internalization of other-directedness promises a new level of “freedom” from dependence on others, even without the firm sense of self of the inner-directed person. This emerging personality type has internalized the need for approval in much the same way the inner-directed person internalized certain standards and values. , . , For the individual who is so formed the continuous creation of self, the move towards actualization, expression and growth comes to rationalize the lack of a solid sense of who one

256

Howard Gadlin

is. So long as he or she is sure that the process of becoming is not inhibited or restrained, it is possible to continue to approve of oneself. Guilt derives then, not from the failure to meet moral standards, but the failure to be oneself. (Gadlin & Rubin, 1978, pp. 36-38)

If there is any validity to the above portrayal, we might expect some reflections of such persons both in newly emerging styles of parenting and in the children such people produce. As far back as 1956 John Seeley had argued that parents’ commitment to child-rearingpractices designed to help the child abandon his parents’ generation is predicated on the promise of continued upward social mobility or at least social cohesion and status stability. It becomes clearer each day that people now raising young children represent the first generation of Americans who can no longer assume their children will have a better life than they. Perhaps this realization, vague and ill-formed though it may be for most, combined with the emphasis on self-actualization which has been nurtured since childhood and reinforced through the media, in psychology, etc., leads parents to approach child rearing with very different attitudes. In many respects adults today have unprecedented opportunities for autonomous personal satisfactions in their private lives, and for some in their public lives as well. It is plausible that such parents can concern themselves less with making things better for their children, and direct their energies toward their personal fulfillment. At the least we can expect that such parents would find attractive child-rearingtactics which minimized the degree to which child rearing interfered with personal accomplishments and pleasures. It is of interest that in her recent survey Baumrind (197 1) was hard-pressed to find parents whose child-rearingstyles could legitimately be called permissive (p. 101). She did find a significant number of parents she describes as authoritative. The stance of these parents was one of discouraging emotional dependence and making early maturity demands-a parental style quite compatible with careerism and consumerism (Robinson, 1971). There is another possibility for those attempting to combine raising children and personal actualization and that is to seek satisfactions in child rearing which either compensate for or match the kind of gratifications available elsewhere. At least since the nineteenth century we have had an ideology which emphasizes the satisfactions of child rearing, but the presumed nature of those satisfactions has changed. For the nineteenth century parent satisfaction could reside in the sense of knowing one had done a job well. But as we move through the twentieth century we can notice an increasing emphasis on the emotional and interpersonal gratification parents were presumed to obtain from their relationship with their children, not just from raising them. And today there appears to be a tendency to expect that almost all aspects of child rearing should be gratifying in the sense of being self-expressive, actualizing, and pleasurable. (People who

Child Discipline and the Pursuir of Self

257

have been through Lamaze classes often have been confronted with the expectation that the childbirth experience itself ought to be one of great sensual gratification.) Of course the fun morality which Wolfenstein described represented an early version of the prescription of pleasure in the parent-child relationship. But at that time the fun seemed important because it was instrumental to the creation of a child who felt good about himself or herself; nowadays the emphasis appears to be more directly on the enjoyment of pleasure itself. More important than fun in the relationship is what Henry (1963) once referred to as a situation where: “both parents seek gratification from the children at the level of deep feeling” (p. 145). I believe the expectation of this sort of gratification in one’s relationship with children can be found even in the research literature on techniques of child discipline, albeit in a somewhat convoluted form. We have already remarked that one of the consequences of the diminution of the parents’ role as authority is the ascendency of the concept of the parent as a friend. Most of our literature looks at the implications of this from the point of view of what the child does or does not get in such relationships. There is, obviously, also the question of what the parent does or does not get in relation to his or her child. Most succinctly, as we noted above, if the parent as authority expects obedience and respect, the parent as friend expects sensitivity and love. The child development literature is so caught up in a manufacturing metaphor, i.e., what kind of product results from particular techniques of production, that we lose sight of the fact that typically parents also have to live with their children. It is reasonable to expect that at least to a certain degree parents’ child-rearing practices are shaped by what they hope for in the quotidian relationship with their child. One of the central foci of research into the consequences of different disciplinary styles is the degree of interpersonal sensitivity produced. It is hard to imagine that parents are concerned with their children’s interpersonal sensitivities only because of their survival value in the public world (McBride, 1973; cf. Sears, Maccoby, & Levin, 1957). This concern with the interpersonal is apparent in Baumrind’s (1971) account of the authoritative parent: “She encourages verbal give and take, and shares with the child the reasoning behind her policy. She values both expressive and instrumental attributes. . . . She recognizes her own special rights as an adult, but also the child’s individual interests and special ways.. .” (p. 22). These are techniques, it would seem, intended to create a child who can be a sort of partner in a relationship. The emphasis on interpersonal sensitivity is most apparent in the literature on induction where parents focus on “explanations or reasons for why the child should not behave in certain ways or why he should change his behavior. These explanations focus on the consequences of the child’s undesirable behavior” (Staub, 1975, p. 5). Hoffman, who has asserted the advantages of induction over

258

Howard Gadlin

either power assertion or love withdrawal, has based his preference on the interpersonal consequences of induction: . . , the technique most likely to optimally motivate the child to focus his attention on the harm done others as the salient aspect of his transgressions, and thus to help integrate his capacity for empathy with the knowledge 01 the human consequences of his own behavior. Repeated experiences of this kind should help sensitize the child. . . (who) . . . is thus gradually enabled to pick out on his own, without help from others, the effects of his behavior. . . . (Hoffman & Saltzstein, 1967, p. 5 5 )

It seems plausible that such child-rearing practices, when combined with the cultural emphasis on self-actualization and gratification, might well contribute to the creation of the types of people we described earlier as having internalized other directedness. Finally, in what might be considered the epitome of fusion of discipline and interpersonal sensitivity we have what Baumrind (197 1) has dubbed the “harmonious” parent. Harmonious parents, we are told, are: characterized as having control (i.e., the child seemed to intuit what the parent wanted and to do it) but not by exercising control (i.e., the parent almost never directed or commanded the child). Harmonious parents seemed neither to exercise control, nor to avoid the exercise of control. Instead, they focused upon achieving a quality of harmony in the home, and upon developing principles for resolving differences and for right living. . . . These parents brought the child up to their level in an interaction but did not reverse roles by acting childishly.. . . Harmonious parents were equalitarian in that they recognized differences based upon knowledge and personality, and tried to create an environment in which all family members could operate from the Same vantage point, one in which the recognized differences in power did not put the child at a disadvantage. (p. 101)

If these children match their parents’ intentions they would seem ideal candidates to be considered junior partners in the relationship between adults. And indeed Baumrind reports that several of these families were about to enter communal living situations, situations in which children are typically expected to shoulder adultlike responsibility for their interpersonal relations (see Berger, Hackett, & Miller, cited in Howe, 1972, pp. 159-169; see also Dreitzel, 1973, Part II). There is one last development i n the literature on child development that we must examine briefly because it might be a portent of practices to come. Shorter (1975) identifies as one of the defining features of the postmodern family: “the definitive cutting of the lines leading from younger generation to older-an adolescent indifference to the family’s identity, to what it stands for, that shows up in the discontinuity of values from parent to children. . .” (p. 269). He goes on to discuss the heightened significance of the peer-group in adolescent socialization and relates this to., among other things: “a shift in the willingness of adolescents to learn from their parents. . . as the post-modern-family rushes down upon us, parents are losing their role as educators. The task passes instead

Child Discipline and the Pursuit of Self

259

to the peers, and with its transfer passes as well a sense of the family as an institution continuing over time, a chain of links across generations” (p. 276). There are, in the child development literature, a few intimations that transfer of responsibility for discipline and socialization may pass on to the peer group for pre-adolescent children as well (Staub, 1975; Parke, 1974). It is hard to know what might be the implications of these studies, if any, for actual child-rearing practices, but it was not too long ago that one might have been considered rather peculiar for suggesting that children might learn by teaching. Whatever their implication, studies like this are good indicators of the degree to which our understanding of children, child-discipline, and the role of parents and the family have changed since the emergence of modern notions of childhood and the family. They are also measures of the decline of the role of the family as the actual socializing institution.

IX. Conclusion In a fairly recent review of the “consequences of parental discipline” Becker (1964) concludes: It is apparent that the consequences of disciplinary practices cannot be fully understood except

in the context of the warmth of the parent-hild relation, the prior history of disciplinary practices and emotional relations, the role-structure of the family, and the social and economic conditions under which a particular family unit is living. (p. 202)

If nothing else, it is somewhat heartening that those words could also stand as part of a summary of this historical survey. In addition, I think we have shown that our “permanent sense of crisis concerning children and the family” ought to be related, not to technical failures and inadequacies of the procedures of child discipline and socialization, but to the conceptions of self, family, and society which inform these techniques and constitute the context within which they are given meaning. The ideal of the privatized family and the self who defines autonomy in terms of estrangement from others are, at base, antagionistic to any society that might provide the grounds for equality and a conception of self and family consonant with a sense of community and mutuality. Our current conceptions of self and family originate along with a specific socioeconomic formation-the industrial capitalist society. The crisis of the family is one of the social problems built into that system and we cannot resolve that crisis nor the other problems merely through changes in socialization practices and family organization; it requires a more fundamental change in our conceptions of personal and family life and the social organization which has shaped those ideas.

260

Howard Gadlin

X. A Note on the Role of Child Development and Psychology In his history of child development Sears (1975) states that the field “. . . was formed by external pressures broadly based on desires to better the health, the rearing, the education and the legal and occupational treatment of children” (p. 3). Like most psychologists he equates the intentions of professionals and the needs of those who seek their help with the social functions of their field. There is another side to the story beside the altruistic and empathic response to the social problems psychologists and child development specialists attempt to understand. Attempts to understand our own activities must not forget, in Bronfenbrenner’s (1961) words: “Our national rhetoric notwithstanding, the actual patterns of life in America are such that children and families come last” (cited in Howe, 1972, p. 140). And Featherstone (1974) has concluded: Even the relatively large sums of money spent on education seem to some extent to have been compensation for neglect in other areas such as health. Throughout our history children’s welfare has been subordinated to economic goals, and institutions for children, operating in a context of market priorities and social inequality, have too often turned into scrapheaps. (p. 164)

Although there is some consolation for being on the right side of this societal pattern of neglect, there is i1 more serious dilemma that relates to the way we understand these problems. From the very beginnings of both psychology and child development we have conceptualized phenomena in terms which are themselves at the core of the very problems that inspire our inquiry. In this way our “understanding” serves in part to perpetuate those problems. Reading through the literature of different eras one is struck by the way prevailing popular conceptions of childhood, selfhood, and family functioning are consonant with the parallel conceptions of scientists and professionals, even when we cannot easily notice direct linkages between them (cf. Senn, 1975). Reviewing trends in infant care, Vincent (1951) came to the conclusion: “. . . It is painfully clear that writers in the field of infant care and child-rearing disciplines have been slow to construct a body of data that withstands empirical scrutiny. Instead, they have often reflected changing patterns of thought in middle-class society and reflected changing theories of education and personality formation” (p. 205). Although the first objection has been answered and there is, in some respects, an impressive array of empirical data, it is still the case that the literature of the field, empirical data included, reflects changing patterns of thought in middle-class society. Kessen (1975) has recently warned: Developmental study, not unlike education and television, is deeply embedded in the politics, economics, and ideology of our time. It would be difficult to demonstrate that the dramatic

Child Disciplinr and the Pursuit of Self

26 I

changes that took place in the larger culture’s attitude toward children from 1954 to 1964 to 1976 had much to do with the opinions of developmentai experts and their research. Rather, the research and the opinions were often in large measure instrumentalities of other powers in American life. (p. 101)

But the opinions of developmental experts, although they may not shape the larger culture’s attitudes, do serve to rationalize those attitudes. In a way the empirical basis of today’s writings makes the discussion of child-rearing practices so much more dangerous. For today, more than ever, the dissemination and diffusion of scientific information give parents the impression that their treatment of children is informed by medical and scientific understanding of the essence of children’s nature and needs. The scientific literature thus legitimizes the contemporary child-rearing practices, and in so doing, exacerbates the tendency to ask questions about and seek changes in techniques of socialization without also questioning the society and social roles which those techniques of socialization serve. Thus far in our history, the bringing to bear of scientific methods and expertise on the understanding of childhood and socialization has served primarily to present as a natural and “psychological” phenomenon what is essentially a sociohistorical development. It is a deception we can no longer afford.

REFERENCES Azrin, N., & Foxx, M. M. Toiler training in less than a day. New York: Simon & Schuster, 1974. Bailyn, B. Education in the forming ofAmerican society. Chapel Hill: University of North Carolina, 1960. Bane, M. J. A review of child care books. Harvard Educational Review, 1973,43, 669-680. Baumrind, D. Current patterns of parental authority. Developmental Psychology Monograph, 1971, 4( I), Part 2. Baumrind, D. The development of instrumental competence through socialization. In A. D. Pick (Ed.), Minnesota Symposia on Child Psychology (Vol. 7). Minneapolis: University of Minnesota Press, 1973. Pp. 3 4 6 . Beales R. W., Jr. In search of the historical child: Miniature adulthood and youth in colonial America. American Quarterly. 1975, 27(4), 379-398. Beck, R. The White House Conference on children: an historical perspectime. Harvard Educational Review, 1973,43(4), 653668. Becker, W. C. Consequences of different types of parental discipline. In M. L. Hoffman & L. W. Hoffman (Eds.), Review of child developmenr research. New York: Russell Sage Foundation, 1964, Pp. 169-208. Berger, P. Toward a sociological understanding of psychoanalysis. Social Research, Spring 1965, 26-4 1. Borstelmann, L. J. Changing conceprs of child rearing, 1920s-1950s. Paper presented at the Southeatern Conference on Human Development, Nashville, April 1976. (a) Borstelmann, L. J. Periodization in the history of American childhood. Paper presented at the meeting of Cheiron, Washington, D.C., May 1976. (b) Braverman, H. Labor and monopoly capital. New York: Monthly Review Press, 1974.

262

Howard Gadlin

Bremner, R. H. Children and youth in America (Vols. I , II, and III). Cambridge, Mass.: Harvard University Press, 1970. Bridges, W. E. Family patterns and social values in America, 182S-75. American Quarterly, 1965, 17, 3-11. Bronfenbrenner, U. Toward a theoretical model for the analysis of parent-child relationships in a social context. In V. C. Glidewell (Ed.), Parental attitudes and child behavior. Springfield, Ill.: Thomas, 1961. Pp. 90-109. Brown, R. D. The emergence of urban society in rural Massachusetts: 1760-1820. Journal of American History, 1974, 61, 29. Cdldwell, B. The effects of infant care. In M. L. Hoffman & L. W. Hoffman (Eds.),Reviewof child development research. New York Russell Sage Foundation, 1964. Calhoun, A. W. A social history of the Americanfamily (3 vols.). New York: Barnes & Noble, 1960. (Originally published, 1918.) Calhoun, D. The intelligence of the people. Princeton, N.J.: Princeton University Press, 1973. Clark, C. E., Jr. Domestic architecture as an index to social history: the Romantic Revival and the cult of domesticity in America, 1840-1870. Journal of Interdisciplinary History, 1976, 7( I), 33-56. Clarke-Stewart, K. A. Interactions between mothers and their young: characteristics and consequences. Child Development Monographs, 1973, 38(6/7). Clauson, J. A. Family structure, socialization, and personality. In L. W. Hoffman & M. L. Hoffman (Eds.), Review of child development research. New York: Russell Sage Foundation, 1966. Pp. 1-53. Cohen, S . The assault on Victorianism in the twentieth century. American Quarterly, 1975, 5, 604425. Danzinger, K . Socialization. Baltimore: Penguin Books, 1971. Degler, C. The sociologist as historian: Riesman’s The lonely crowd. American Quarterly. 1963, 15, 483497. deMause, L. The evaluation of childhood. History of Childhood Quarterly, 1974, 1(4), 503-575. Demos, J. A little commonwealth: family life in Plymouth Colony. London and New York Oxford University Press, 1970. Demos, J., & Demos, V. Adolescence in historical perspective. Journal of Marriage and thefamily, 1969, 31, 632-638. Dreitzel, H. P . Recent sociology. No. 5: childhood and socialization. New York: Macmillan, 1973. Duvall, R. Conceptions of parenthood. American Journal of Sociology, 1946, 52, 193-203. Earle, A. M. Childlife in colonial days. New York: Macmillan, 1899. Eliot, W. G . Lectures to young women. Boston: Crosby, Nichols, 1855. Featherstone, J. Children and youth in America. Harvard Educational Review, 1974, 44(1), 160168. Furstenberg, F. Industrialization and the American family: a look backward. American Sociological Review, 1966, 31, 326-337. Gadlin, H. Private lives and public order: a critical view of the history of intimate relations in the United States. Massachusetts Review, 1976, 17(2), 304-33 1. (a) Gadlin, H. Spare the rod: disguising control in American childrearing. Paper presented at the meeting of the American Association for the Advancement of Science, Boston, February 1976. (b) Gadlin, H . , & Rubin, S. Interactionism: a non-resolution of the person-situation controversy. In A. Buss (Ed.), The social context of psychological theory: toward a sociology of psychological knowledge, New York: Irvington, 1978, in press. Greven, P. Family structure in seventeenth century Andover. William andMary Quarterly, 1966.33, 234-256.

Child Discipline and the Pursuit of Serf

263

Greven, P. (Ed.). Child-rearing concepts, 1628-1861. Itasca, Ill.: Peacock, 1973. Gutman, H. Work, culture and society in industrializing America: 1815-1919. American Historical Review, 1973, 70(3), 531-589. Harrison, J. F. C. Learning and living. Toronto: University of Toronto Press, 1961. Henry, J . Culture against man. New York: Random House, 1963. Hoffman, M. L. Parent discipline and the child’s consideration for others. Child Development, 1963, 34(3), 573-588. Hoffman, M. L., & Saltzstein, H. D. Parent discipline and t9e child’s morel development. Journal of Personality and Social Psychology, 1967, 5( 1 ) , 45-57. Howe, L. K. The future of the family. New York: Simon & Schuster, 1972. Hunt, D. Parents and children in history. New York: Basic Books, 1970. Inkeles, A. Social change and social character: the role of parental mediation. Journal of Social Issues, 1955, 11, 212-223. Inkeles, A. Social structure and socialization. In D. A. Goslin (Ed.), Handbook of socialization theory and research. Chicago: Rand McNally, 1969. Jeffrey, K. The family as utopian retreat from the ci:y. In S. Tessele (Ed.), The family. communes, and utopian retreat. New York Harper, 1972. Jones, A. D., Stewart, A,, & Winter, D. G . Socialization and themes in popular drama. European Journal of Social Psychology, 1974, 4, 65-84. Judd, C. H. Psychology. New York: Scribners, 1907. Kessen, W. The child. New York: Wiley, 1965. Kessen, W. Commentary. In J. E. Senn (Ed.), Insights on the child development movement in the United States. Monographs ofthe Society for Research in Child Development, 1975, 40(3/4). Kett, J. F. Growing up in rural New England, 1800-1840. In T. Hareven (Ed.), Anonymous Americans. Englewood Cliffs, N.J.: Prentice-Hall, 1971. Kiefer, M. American children through their books. Philadelphia: University of Pennsylvania Press, 1948. KlappeI, Z. S . The impact of the women’s liberation movement on child development books. American Journal of Orthopsychiatry, 197 I , 41(5), 725-732. Kluckhohn, F. R. The American family, past and present. In Patterns for modern living. Chicago: Delphian Society, 1952. Lasch, C. The emotions of family life. New York Review of Books, 1975, 22( l9), 3 7 4 1 . (a) Lasch, C. The family and history. New York Review of Books, 1975, 22(18), 33-38. (b) Lasch, C. The family as haven in a heartless world. Salmagundi, 1976, 35, 42-55, Laslett, B. The family as a public and private institution: an historical perspective. Journal of Marriage and the Family, 1973, 35, 480492. Laslett, P. The world we have lost (2nd ed.). New York Scribner’s, 1971. Laslett, P., & Wall, R. Household and familv in past time. London and New York: Cambridge University Press, 1972. Lipset, S. M., & Lowenthal, L. Culture and social character: the work of David Riesman reviewed. Glencoe, Ill: Free Press, 1%1. Lomax, E. Early twentieth century advice on infant care: the triumph of mechanistic materialism. Paper presented at the meeting of the Society for Research in Child Development, Denver, April 1975. McBride, A. The growth and development of mothers. New York: Harper, 1973. Mechling, Advise to historians on advice to mothers. Journal of Social History, 9, 1975, 4 4 6 3 . Mergen, B. The discovery of children’s play. American Quarterly, 1975, 27(4), 399420. Miller, D. R., & Swanson, G. E. The changing American parent. New York: Wiley, 1958. Modell, J., & Hareven, T. K. Urbanization and the malleable household: an examination of boarding and lodging in American families. Journal of Marriage and the Family, 1973, 35, 467479.

264

Howard Gadlin

Neuman, R. P. Masturbation, madness, and the modern concepts of childhood and adolescence. Journal of Social History, 1975, 8, 1-27. O’Neill, N., & O’Neill, G. Open Marriage. New York: Evans, 1972. Parke, R. D. Rules, roles, and resistance to deviation: recent advances in punishment, discipline, and self-control. In A. D. Pick (Ed.), Minnesota Symposia on Child Psychology (Vol. 8). Minneapolis: University of Minnesota Press, 1974. Pp. 11-143. Potter, D. People of plenty. Chicago: University of Chicago Press, 1954. Rapson, R. L. The American child as seen by British travelers, 1845-1935. American Quarterly, 1965, 17, 520-534. Riesman, D. The lonely crowd. New Haven, Conn.: Yale University Press, 1950. Robinson, J. P. Historical changes in how people spend their time. In A. Michel (Ed.), Family issues of employed women in Europe and America. Leiden: Brill, 1971. Rogers, C. Becoming partners. New York: Dell Publ., 1972. Rothman, D. J. The discovery of the asylum: social order and disorder in the new republic. Boston: Little, Brown, 1971. Rothman, D. J. Documents in search of an historian: toward a history of childhood and youth in America. In T. K.Rabb & R. I. Rotberg (Eds.), Thefamily in history. New York Harper, 1971. Sears, R. R. Your ancients revisited: a history of child development. In E. M. Hetherington (Ed.), Review of child development research. Chicago: University of Chicago Press, 1975. Pp. 1-74. Sears, R. R., Maccoby, E. E., & Levin, H. Patterns of child rearing. Evanston, Ill.: Row, Peterson, 1957. Seeley, J., Sim, R. A., & Loosley, E. W. CrestwoodHeights. Toronto: University of Toronto Press, 1956. Senn, M. J. E. Changing concepts of child care. In I. Galston (Ed.),Societyandmedicine.New York New York Academy of Medicine, 1955. Pp. 83-103. Senn, M. J. E. Insights on the child development movement in the United States. Monographs of the Society for Research in Child Development, 1975, 40(3/4). Shoner, E. The making of the modern family. New York: Basic Books, 1975. Smith, M., & Gadlin, H.Parental control of young children: perspectives in American history. Paper presented at the meeting of the Society for Research in Child Development, Denver, April 1975. Stannard, D. American historians and the idea of national character: some problems and prospects. American Quarterly. 197 I , 23( 12). 202-220. Stannard, D. Death and the Puritan child. American Quarterly, 1974, 26(5), 456476. Staub, E. The development of prosocial behavior in children. Morristown, N.J.: General Learning Press, 1975. Stendler, C. Sixty years of child training practices. Journal ofpediatrics, 1950, 36, 122-134. Strickland, C. A transcendentalist father: the child-rearing practices of Bronson Alcon. History of Childhood Quarterly, 1973, 1(1), 4-51. Strong, B. Towards a history of the experiential family: sex and incest in the nineteenth century family. Journal of Marriage and the Family, 1973, 35,457466. Thompson, E. P. Time, work-discipline, and industrial capitalism. Past and Present, 1966, 38, 56-97. Twombly, R. L. Saving the family: middle class attraction to Wright’s prairie house 1901-1909. American Quarterly, 1975, 27( I), 57-72. Vincent, C. Trends in infant care ideas. Child Development, 1951, 22, 199-203. Walkowitz, D. J. Working-class women in the gilded-age: factory, community and family lifeamong Cohoes, New York, conon workers. Journal of Social History, 1972, 5(4), 464-490. Watson, J. B. Psychological care of infant and child. New York Nonon, 1928. Weinbaum, B., & Bridges, A. The other side of the paycheck monopoly capital and the structure of consumption. Montly Review, 1976, 28(3), 88-103.

Child Discipline and the Pursuit of Self

265

Weinstein, F., & Platt, E. M. The wish to be free. Berkeley: University of California Press, 1969. Welter, B. The cult of true womanhood: 1820-1860. American Quarterly, 1966, 18, 151-174. White, B. L., & Watts, J. E. Experience and environment (Vol. 1). Englewood Cliffs, N.J.: Prentice-Hall, 1971. Whiting, J., & Child, 1. L. Child training and personality. New Haven, Conn.: Yale University Press, 1953. Wiebe, R.The social function of public education. American Quarterly, 1969, 21(2), 147-176. Wiggin, K. Children’s rights. New York: Houghton, Mifflin, 1899. Wishy, B. The child and the republic, Philadelphia: University of Pennsylvania Press, 1968. Wolfenstein, M. The emergence of fun morality. Journal of Social Issues, 1951, 7(4), 15-25. Wolfenstein, M. Trends in infant care. American Journal of Orthopsychiatry. 1953, 23, 120-130.

This Page Intentionally Left Blank

DEVELOPMENT OF TIME CONCEPTS IN CHILDREN'

William J . F riedman OBERLIN COLLEGE

1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

267

I1 . LOGICAL TIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . TEMPORAL RELATIONS IN INFANCY .............................. B . SUCCESSION AND DURATION IN EARLY CHILDHOOD . . . . . . . . . . . . . . C . SUCCESSION AND DURATION IN SCHOOL-AGE CHILDREN . . . . . . . . . . D . LOGICAL TIME CONCEPTS OF OLDER CHILDREN . . . . . . . . . . . . . . . . . .

269 270 272 274 279

III . CONVENTIONAL TIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . CONVENTIONAL TIME CONCEPTS IN INFANCY AND EARLY CHILDHOOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . CONVENTIONAL TIME CONCEPTS OF 5- TO 7-YEAR-OLDS . . . . . . . . . . C . CHANGES IN MIDDLE CHILDHOOD AND ADOLESCENCE . . . . . . . . . . . .

280

1V. EXPERIENTIAL TIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . SIMPLE JUDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B . COMPLEX JUDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C . DEVELOPMENT OF EXPERIENTIAL TIME ..........................

286 287 290 292

V . SUMMARY AND CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

294

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

296

280 283 284

I . Introduction Psychologists have assimilated the topic of time to each of the subdivisions of their discipline from perception to psychopathology (see Zelkind & Sprug. 1974. for a comprehensive bibliography) . A large number of studies spanning nearly 90 'The author wishes to acknowledge the help of Michael L . Davidson and to thank David Elkind and Michael J . Chandler for their comments on a previous draft of this chapter . 267

268

William J . Friedman

years have treated time as a sensory modality subject to psychophysical investigation. This work has been closely related to the so far unsuccessful speculative and empirical search for an underlying physiological mechanism. Another approach has emphasized modality differences in time perception phenomena, pointing toward their dependence on well-established sensory modalities and, recently, information-processing models of many perceptual effects have been proposed. Time concepts, judgments, and perspective have also been related to a variety of organismic variables including personality, social class, intelligence, sex, and pathology, and situational variables including anxiety, motivation, and physiological state. The emerging picture of psychological time is multifaceted. Time seems to be a physiological, perceptual, cognitive, and social, as well as a natural phenomenon. No single level of analysis adequately describes all aspects of psychological time; it is therefore essential to classify phenomena and to decide which can be treated most fruitfully by a given approach. The developmental psychological approach to the study of time is promising, especially for understanding the interrelation of difference levels of temporal functioning. Between 1920 and 1975 more than 50 studies of children’s time language, concepts, and judgments have been conducted. To a surprising extent, however, these studies have been carried out in relative isolation from one another, with little interchange on the subject of what are the crucial issues or phenomena. A series of reviews (Brackbill, Fitzgerald, & Lintz, 1967; Fraisse, 1963; Goldstone & Goldfarb, 1966; Jahoda, 1963) has done little to rectify the problem. Probably a useful first step would be a classification scheme that could organize the existing developmental time literature. In the present review we will make theoretical distinctions that are tempered by a knowledge of what developmental information is available. These distinctions draw in a general way on three main currents in philosophical analysis: (1) a scientific tradition, from Newton to Russell, that posits an objective temporal flow encompassing all events:, (2) a search, initiated by Wittgenstein, for the meaning of time in common language; and (3) a group, including Bergson and Whitehead, who have focused on experiential aspects of the present. In that one of the purposes of a developmental approach is to understand the relationship between different psychological processes, an overemphasis on the contrasts between the levels of functioning would seem counterproductive. The classification to be offered should be viewed as a tentative analysis in service of a later synthesis. It is the resulting developmental patterns which will support or weaken these distinctions. Three aspects of psychological time will be discussed, corresponding to (1) knowledge of logical temporal relations, (2) knowledge of social representations, and (3) perception and intuition. The first aspect, logical time, pertains to a

Time Concepts in Children

269

system which permits the inference of relations of succession and duration from distinct events. It is closely allied with the scientific tradition that assumes time to be a homogeneous and all-encompassing flow. The second aspect, conventional time, refers to the knowledge of socially shared systems which organize temporal phenomena. These systems are represented grammatically, lexically, by numerical schemes, and by concrete devices. The correspondence between conventional time and the third philosophical approach is only partial, because this aspect has nonlinguistic representations and because certain temporal relational terms seem suited to express features of logical time. The third aspect, which we will call experiential time, refers to subjective impressions of the passage of time. It is closest to the third philosophical focus mentioned above. Philosophers have described the experience of instants and durations, using as data their own free introspection. Here we will restrict our considerations to introspective data in the form of controlled duration judgments. One of the problems in attempting to categorize children’s behavior according to such a scheme is that early forms are often relatively less differentiated than the end-state forms. Prelogical and preconventional time concepts, for example, must often be distinguished somewhat arbitrarily. In the discussion which follows we will consider findings of experimental tasks in the section most relevant to the eventual solution of the task.

11. Logical Time Logical time, as we will use the term, refers to a subset of knowledge about time: It pertains to the Newtonian scheme of correlating a given event with a unique position on a temporal continuum. According to this scheme all events can be ordered on a common scale, and durations can be reckoned as the intervals between points. It is difficult to reflect upon logical time because, at least in Western societies, it pervades adult intuitions about the natural world; we assume that any collection of occurrences can be objectively ordered. However, for the young child (as for Einsteinian time) temporal order is not invariant under certain kinds of transformations. Logical time is an abstraction beyond perceptible order or duration. It is not the memory of events but rather the deduction of temporal relations between them. For instance, older children and adults are able to make inferences of the following sort: If two runners begin a race simultaneously, the winner will have taken less time to complete the course. The person making the judgment need not perceive the relative durations-in fact it is irrelevant whether the race is a sprint or a marathon. Instead h e must correlate the starts and finishes within a common scale. Such judgments require an understanding of ordinal aspects of logical

270

William J . Friedman

time. But adults also understand metric features of logical time; they can decompose a duration into equal units, or measure an interval using a standard velocity. The latter competence is necessary for comparing successive intervals. Logical time is a prominent construct in most scientific disciplines; it is the backdrop against which scientists commonly view process and causality (while process and causality are, in turn, prerequisite to certain scientific definitions of time). Logical time, as studied by Piaget, is also the end-state description of human cognition about time, that is, the standard of maturity against which he assesses genetically prior concepts. Like other concepts which Piaget has studied, logical time concepts are systematic in that diverse items of information are organized into a coherent structure that allows reconstruction or reorganization. Less mature time concepts are presumed to have their own structure but to be more limited in their capacity to systematize perceived or remembered events. Piaget's method of genetic epistemology is to seek the origins of logical thought in ontogenesis, often in the sensorimotor behavior of infants. This method incorporates the assumption of certain organismic developmental theories that the ontogeny of thought is analogous to the phylogeny of species, and that a given form is the product of its lineage and its current mode of adaptation. However, establishing the lineage of a concept is fraught with theoretical and practical problems. Piaget addresses the problem not by administering a simplified time-concept task to infants but instead by examining the way infants adapt to temporal features of their environment. The infant, as the adult, adapts to certain regularities of his world. Some of these consistencies inhere in the temporal organization of the infant's own actions and others in the events that he witnesses. A . TEMPORAL RELATIONS IN INFANCY

All actions are by their nature temporal; actions have beginnings and endings and are articulated in time. But unless actions are purposely ordered to effect some result, we have no evidence that the organism appreciates temporal succession. Conditioned responses and primary and secondary circular reactions (Flavell, 1963) all involve sequential movements but all do not require the subject to externalize the temporalkausal relationships. However, when an end state is obstructed and intervening acts are called for, a primitive externalization of temporal relations is required. I . The Means-End Sequence The ability to order actions i n service of a goal is, according to Piaget,(1952, 1954), an achievement of the fourth stage of the sensorimotor period, beginning at about 8 months. Until this time the infant is unable to defer acting directly on

Time Concepts in Children

27 I

the goal object and fails to perform intermediate actions that are necessary to bring the object within reach. For example, if a pillow is placed between the infant and a watch he desires, he will persist in reaching for the watch but do nothing to move the pillow. The fourth-stage infant orders his behavior according to a distinction between means and ends: First the pillow is pushed aside, then the watch is captured. This ordering of means and ends must be distinguished from primary and secondary circular reactions, for which a fixed order of acting and sensing is not crucial to maintain the behavior. These circular reactions rely on simple contingencies in the environment. If the baby bangs a stick on the side of his playpen, a sound naturally follows. The sound may become linked by the infant to the banging, and the behavior will tend to perpetuate itself. However, the Stage 3 infant does not treat the sequence as the repetition of independent acts, each with its perceptual consequence. Rather, the action and its consequences are bound in an indivisible cycle. Cavanagh and Davidson (1974) have shown that when a novel contingency is learned and exercised by 6-month-old infants, noncontingent presentation of the “consequence” evokes the “prerequisite” act. The temporal sequence salient to the adult seems irrelevant to the infant. Thus, the meansmds sequence represents an important step in adapting the temporal organization of action to the causal structure of the world. According to Piaget (1954), however, the Stage 4 infant’s knowledge of temporal succession exists only in the realm of personal action. It is inseparable from the infant’s practical appreciation of causal relations. Until about 1 year of age, the beginning of Stage 5, he remains unable to take account of successions in which he does not participate. For example, when the Stage 4 infant watches as an object is sequentially hidden in several places, he will most likely not seek it in its current location, but rather check its previous resting places. The Stage 5 infant, in contrast, realizes that the object could only be in its latest location and immediately searches for it there. It is evident from this last example that an objective reckoning of temporal sequences is intimately related to the notions of uniform space and the enduring object (Piaget, 1954). The ability to cognize sequential displacements presupposes knowledge of the object as an enduring entity whose existence is not cancelled by spatial displacements which take it out of sight.

2 . The Representation of Sequence A further improvement in objectifying temporal order accompanies the Stage 6 infant’s advances in memory of events. Through the mechanism of deferred imitation (Piaget, 1962) the 18-month-old becomes able to partially reconstruct past events without perceptual support. The infant of this age can remember the last place he put a toy or where his father or grandfather has gone (Lewis, 1937;

272

William J . Friedman

Piaget, 1954). However, this capacity likely reflects a tendency to focus attention on the end state or result of a series of actions and not the ability to reconstruct a series. True temporal reconsbuctions appear to be a much later achievement. If the infant is bound to the end state in his recollection of a series, then it is useful to distinguish the representation of objective temporal order from the practical ordering of means-ends sequences (and from immediate memory of a sequence). By about 2 years, goal-directed actions embed both the before and after relationship in sequences of three acts. For example, Piaget (1952) described his daughter at 1 year 8 months arriving at a closed door with a blade of grass in each hand. Realizing that she cannot turn the door knob and retain the grass at the same time, “. . .she puts the grass on the floor, opens the door, picks up the grass again and enters” (p. 339). The act of opening the door is, by design, afrer putting the grass down and before picking it up. The practical ordering of actions thus appears more advanced than the representation of the temporal order of events much as the young child’s practical concept of space is more advanced than his egocentric representation of space. 3. The Limits of Sensorimotor Time The 2-year-old’s conception of time is extremely limited when compared with the adult end state. Logical time requires the representation of events in a single uniform series. The 2-year-old can retrace a sequence of displacements from the immediate past, but in the case of events in the more remote past and without perceptual support the child seems to focus on the end of a sequence. We must conclude that the child of this age is not yet capable of truly representing an isolated series, let alone the all-inclusive metric series of logical time. B . SUCCESSION AND DURATION IN EARLY CHILDHOOD

During the years following the late sensorimotor period, speech provides an increasingly important means for exposing the child to the conventional concepts embedded in language. The English language contains several words used by adults to denote logical temporal relations of succession, simultaneity, and duration. 1 . Early Comprehension of Temporal Relational Terms

A number of recent studies of the time language of preschool children have focused attention on the terms before and a f e r (Amidon & Carey, 1972; Clark, 1971; Coker, 1975; Ferreiro & Sinclair, 1971; W. Friedman & Seely, 1976; Keller-Cohen, 1975). This research indicates that the terms are incompletely understood as temporal relational terms and little used by 3 and 4-year-olds. However, in one study (W. Friedman & Seely, 1976) even 3-year-olds showed a comprehension significantly above chance levels for several of the temporal

Time Concepts in Children

273

relational terms. There is some disagreement as to whether the comprehension limitations generally found reflect syntactic complexities inherent in the tasks used or inadequate temporal relational concepts. Clark (1971) has argued that 3-year-olds are not yet able to distinguish between the meanings of before and a f e r , but Amidon and Carey (1972) and Coker (1975) believed the presence or absence of adequate concepts is masked by the children’s difficulty with complex syntactic structures. 2 . Nonverbal Tasks In spite of the difficulties that 3- and 4-year-olds show in the comprehension of temporal relational terms, they are adept at nonverbal imitation of the sequence or simultaneity of two events. Keller-Cohen (1975) found that by age 4, 73% of the subjects correctly imitated the order of two actions modeled by the experimenter. It is not clear whether this competence is distinguishable from the Stage 6 infant’s recollection of series in the immediate past. An adequate test of the young child’s ability to represent order would require a longer delay between modeling and imitation.

3 . The Acquisition of Temporal Relational Terms Between the ages of 4 and 5 there appears to be a marked improvement in children’s comprehension and production of temporal words which relate two events. Clark (1971), Keller-Cohen (1975), and W. Friedman and Seely (1976) have found the achievement of high levels of competence in the comprehension of the terms before and after by 5 years. Coker (1975) showed that of children only slightly older (5% years) 60% consistently answered correctly questions of the form, “What did I show you before/after X,” based on a memorized threeevent sequence. Cromer (1968) found in the spontaneous speech records of two children an increase in the use of complex ordering with “before” and “after” as conjunctions at about 4% years. Ferreiro and Sinclair (1971) similarly observed that 4- to 5-year-olds begin to use adverbial expressions such as “before” and “after” in their description of pairs of events, and Keller-Cohen (1975) found that by about age 5 most subjects correctly answered a “when” question using a relational temporal order term. 4 . Cognitive Interpretations What is the nature of the transition between ages 3 and 5 in the ability to understand and use temporal relational terms? Cromer (1968) and Ferreiro and Sinclair (1971) attributed the child’s progress between 3 and 5 years to general cognitive advances in the ability to decenter, which mark the onset of the intuitive substage of preoperational thought. According to this interpretation the child now has the flexibility to go beyond focusing on one event and to consider the relationship between event A and event B. Ferreiro and Sinclair (1971) found,

William J . Friedman

214

however, that the temporal relationship cannot be preserved when the subjects are required to start the sentence with event B. For example, 5-year-olds who assert “A before B” are not then able to restate the information in the form “B after A.” Thus, even these early relational concepts are quite limited when compared to those of older children and adults. Unfortunately, there is little independent evidence for this 4-to 5-year transition in nonlinguistic studies. In fact there is no research that fills the gap between Piaget’s infant work and his studies of the temporal seriation of 6- to 10-yearolds. If, as Piaget (1955) suggested, the preoperational child “gradually attempts to execute the equivalent of sensorimotor acts on the plane of thought” (p. 3 3 , we would expect to find continuity in the 2- to 5-year-old’s ability to represent temporal order. The dramatic improvement in the ability to express temporal relations in speech during the fifth year may depend on some prior, as yet untapped, cognitive process ,, or alternatively semantic and syntactic progress may facilitate the representation of temporal relations.

5 . Early Duration Concept:r There is little information on either the linguistic or nonlinguistic representation of duration by young children. Studies based on speech records indicate that the expression of duration in the speech of preschoolers is uncommon and rudimentary (see Section 111). Two studies have dealt with the duration concepts of preschool children using mainly nonlinguistic responses. Weil ( 1975) found that kindergarten children showed above-chance performance on a discrimination learning task in which relative duration of motion of a pair of race cars was the critical attribute. However, even their moderate success was limited to the concept “more time” (vs. “less time”). The kindergarteners were also unable to benefit from an explicit statement of the rule during a later training phase. Longobardi and Wolff (1973) investigated the ability of 5- and 8-year-old subjects to imitate and explain the movement of two race cars, equated for either speed or duration of motion. In the task two unequal lengths of track were either “start-aligned” or “end-aligned. Thus, an important variable was the influence of particular spatial cues associated with starting or ending the motion. Of interest here is the result that the younger subjects could imitate the equal duration response only when the tracks were “start-aligned. Older subjects could imitate equal duration of motion either with start- or end-aligned tracks. In the case of moving objects the preschoolers’ concept of duration appears to be intermeshed with spatial notions. This has also been an important finding of Piaget’s work on logical time concepts in children. ”

”

C. SUCCESSION AND DURATION IN SCHOOL-AGE CHILDREN

An extensive series of studies conducted by Piaget (197 1) and collaborators is a major source of information on the limitations of logical time concepts in early

T i m Corn-epts in Children

275

school-age children and the subsequent changes through middle childhood. Three sets of tasks from Piaget’s volume on time concepts will be considered in this section.

1. General Features of the Piagetian Tasks The design of the first two sets of tasks can best be understood in the light of the definition of logical time that Piaget has arrived at: Logical time is a system that entails the mutual reckoning of succession and duration of discrete events to accord with a unitary series. This definition accounts for two features of these Piagetian time tasks. First, since logical inferences require a unitary reconstruction of discrete events, temporal order or duration must not vary with spatial position. These tasks therefore involve motions of controlled velocities through space. The control of velocity allows the empirical dissociation of temporal and spatial duration cues. For example object A , traveling faster than object B , may stop first yet end up ahead of B. Second, discrete events must be integrated into a single reconstruction. Therefore, each task employs pairs of events-for example, pairs of cars racing, beakers filling, or imaginary trees growing. A third set of studies dealt with the concept of age and judgment of the duration of actions in light of the results of the first two sets and will be mentioned only briefly.

2 . Communicating Vessels Task The apparatus for the first set of tasks consisted of two communicating vessels, one on top of the other, with a stopcock regulating the downward flow of liquids. In the course of a demonstration the top flask emptied into the lower beaker in a series of steps, between which the stockcock was closed. The upper flask was pear-shaped, and the lower vessel was a tall narrow beaker. ‘Therefore, the water level appeared to fall more slowly than it rose. The subject was provided several identical drawings of the apparatus and asked to indicate the two water levels on a drawing each time the water flow was interrupted. The set of figures that the child had marked was used in subszquent reconstructions of the temporal sequence. In a study of children’s ability to reconstruct the series, the figures were shuffled and the subject was asked to recreate the actual order. First, the complete drawings were seriated. Then the drawings were cut between vessels, reshuffled, and seriated pairwise. Piaget found that the youngest subjects, about age 6, were unable to order either the cut or uncut drawings. The 7- and 8-yearo!ds succeeded in seriating the uncut drawings and the upper or lower sets of cut drawings, but either failed or had difficulty seriating the cut drawings pairwise. By about 9 years subjects succeeded on all the tasks. In a replication study Love11 and Slater (1960) found similar age trends for the cut drawings, but their 5- and 6-year-olds were somewhat better at seriating the uncut drawings. Piaget’s (197 1) results indicate that temporal reconstructions using figural

276

William J . Friedman

representations are not mastered until about the time children can seriate sticks according to their length (Inhelder & Piaget, 1964). Using the communicating vessels, the experimenter asked the subjects several questions that required them to compare durations of the periods during which the water flowed (Piaget, 1971). Until about the age of 9, subjects had considerable difficulty comparing two intervals for the same beaker and an interval for the top beaker with an interval for the bottom beaker. Often, younger subjects claimed it was impossible to make the comparison because the water could only be at one level at a time. In other cases, younger children thought that synchronous intervals were unequal because either dropping or rising took longer than the other. Love11 and Slater’s (1960) younger subjects were more successful than Piaget’s at comparing synchronous intervals. However, since the question was of the form, “Does it take the same time.. .,” it is possible that many young children simply answered “yes” without understanding the equality of the intervals. When the same children were asked to equate the total intervals of emptying and filling using the diagrams, only one child in each age group below 8 years equated durations. Most of the rest thought that the tall jar took longer to fill. The latter answers probably more accurately indicated the competence of the younger subjects. In both of these studies, as in the experiments of Longobardi and Wolff (1973) and Wed 1(1975),young children’s duration judgments were influenced by spatial features of the task.

3 . Tasks Requiring the Coordination of Succession and Duration The second set of tasks (Piaget, 1971) included variants of two basic demonstrations. In one, two race cars were moved along parallel paths at different speeds. The cars started together and stopped either successively or simultaneously. Sometimes the race was run in a series of synchronous intervals. In the second demonstration a beaker emptied through an inverted Y-shaped tube. A stopcock above the branching regulated the duration of flow so that it was equal in each branch of the tube. In three of these studies Piaget investigated the ability of children to deduce the relative duration of motion of the objects from the succession or simultaneity of their starting and stopping times. Children were classified into three stages on the basis of their responses to questions. During stage I, successions and durations remain undifferentiated from distances: ‘‘longer’’ is equivalent to further: “first” may mean “before” or “after,” and differences in speed are thought to preclude synchronous processes and lead to confused estimates of duration. During the second stage the initial intuitions slowly become differentiated or articulated either because “before” and “after” in time and space become differentiated from each other, or else because simultaneity becomes recognized independently of positions or velocities, or finally because duration is understood to be inkersely proportional to velocity.. . . Finally, at stage DI the subjects become capable of applying the techniques of operational grouping to all the relations

Time Concepts in Children

271

involved, and go on to construct a coherent system involving both durations and successions. (Piaget, 1971, p. 92)

There was a general decrease in Stage I responses and increase in Stage 111 responses between 5 and 8 years. Fraisse and Vautrey (1952, cited in Fraisse, 1963) replicated part of Piaget’s findings using a similar race car task. They found that 88% of their 5-year-olds equated the durations when the start and stop were simultaneous and motions were at the same speed in opposite directions, but only 17% equated the durations when motions were in the same direction at unequal speeds. Love11 and Slater (1960) found Stage I responses predominant in all age groups, though the frequency decreased between ages 5 and 9. Three additional studies dealt with children’s ability to make logical inferences about equality, colligation, transitivity, and iteration of durations (Piaget, 1971). In each case the youngest children (about ages 5-6) were unable to make the comparision or were misled by irrelevant spatial cues. Slightly older children (about ages 7-8) were limited to partial comparisons and seldom reached the correct conclusion. By 8 or 9 years most subjects were able to reasons systematically about component durations and arrive at a reconstruction that preserved the temporal relationships. These older subjects were also able to use an independent motion to time an event, thus marking the understanding of metric time. Piaget believed that this achievement rests on the ability to equate synchronous intervals and to colligate standard intervals into a longer duration. In the final set of tasks the investigators showed that the limitations and capabilities demonstrated with perceptible events also apply to children’s judgments of relative age and short periods of time. For example, young children incorrectly attribute greater age to a taller tree that was planted later. Similarly, young subjects are swayed in their judgments of short durations by such factors as the amount of effort expended during an interval (Piaget, 1971; see also Section IV). 4 . Related Studies Other investigators have attempted to extend Piaget’s findings or to deal with theoretical issues raised by the studies. A first study dealt with temporal cognition in 4- to 7-year-olds while varying the difficulty of verb tense forms used in the instructions. Cromer (1971) attempted to construct a temporal task analogous to Piaget’s Three Mountain Test of spatial perspective. Subjects were pretrained to learn the conventions of the comic strip format and then were asked to point to the frame indicated by the verb tense of a statement made by one of the characters. If the subject ignored the tense and “centered” on the content he would fail to choose the correct frame. Results, listed by mental age on the Peabody Picture Vocabulary Scale, revealed that only subjects with a mental age of 6 or more

278

William J . Friedman

were able to decenter consistently. Linguistic form appeared to be a minor influence on performance, and most subjects at each mental age group were able to imitate all the tense forms. Cromer’s study provides independent evidence that the early understanding of a temporal sequence is limited by an apparent inability to recompose the relations mentally. In a previously mentioned study, Longobardi and Wolff (1973) contrasted children’s verbal responses with nonverbal responses to equal time and equal rate dsmonstrations. The authors found that verbal assessments alone masked cognitive progress that could be demonstrated in the imitation of actions. Second graders (mean age 7 years 10 months) were able to imitate equal time or equal rate responses in spite of dissonant spatial cues, while nursery-school subjects were bound to the cues of the track alignment. However, when asked whether a demonstration was of equal time or equal rate, the second graders and nurseryschool groups both answered at below chance levels. Logical justifications of the sort reported by Piaget (197 1 ) were uncommon in either group. In a related study, also previously mentioned, Weil (1975) examined children’s ability to learn to discriminate between the duration of motion of two race cars. Weil found that even subjects as old as 11 years responded at little over chance levels in the course of 40 trials. Following the first phase, subjects were given 19 training trials during which the experimenter called attention to relevant cues and explicitly stated relevant rules. Posttest discrimination performance improved in 80% of the oldest group, and 20% of the first and second graders, while the kindergarteners failed to benefit from the training. The results of the two studies indicate that below about 8 to 9 years subjects d o not give logical explanations for equal or unequal duration in term of starting and stopping times, and their performance does not improve much when such explanations are provided for them. Furthermore, unless equal duration is an easier concept than greater or shorter duration, it appears that imitation responses precede discrimination responses, which may, in turn, precede logical justifications. Longobardi and Wolff’s second graders were better at imitating equal durations than were Weil’s second graders at discriminating on the basis of unequal duration. Weil’s second graders, however, seemed to have some success at learning the discrimination, while the second graders in both studies were unable to give logical justifications that related duration to starting and stopping times. In a fourth study, Berndt and Wood (1974) attempted to induce operational responses to Piagetian paired velocity problems in 5- to 7-year-olds by pitting visual-spatial cues against auditory cues. The authors reasoned that if a child’s perception of the relative length of two durations could be brought into conflict with his judgments in the usual spatial tasks, the child might overcome his dependence on distance cues. A pretest showed that all of the children performed like Piaget’s preoperational subjects; their judgments of the time inverval during which a pair of toy trains were moving was influenced by the distance covered. During a second phase the trains were hidden by a tunnel, and the subjects

Time Concepts in Children

279

compared successive and simultaneous durations solely on the basis of time intervals during which the train whistles sounded. A large proportion of the children in each age group met a strict criterion of success. Next, the subjects were given a series of trials in which they first judged by sound alone, and then, after the tunnel was lifted, by spatial position as well. In cases where the spatial cues conflicted with their previous judgments, most children in each age group were misled by the spatial cues and reversed their judgments. Similarly, a posttest failed to show a significant overall improvement on judgments without the tunnel. However, some individual children did improve, presumably as a consequence of the conflict trials, and such improvements was more common among 7-year-olds than 5-year-olds. As i n the previous two studies, logical justifications for the judgments were nearly absent. 5 . Summary The studies reviewed provide considerable support for the general features of Piaget’s description of the development of time concepts. In particular, several studies have shown that before about 8 or 9 years children have considerable difficulty deducing relative duration from starting and stopping times and vice versa. Also well supported is the finding that young children are likely to confuse temporal succession and duration with spatial order and distance or size. Other studies lend credence to Piaget’s conclusion that the ability to reconstruct a temporal sequence goes beyond simple intutions-that young children are limited in their ability to consider multiple points in time. D . LOGICAL TIME CONCEPTS OF OLDER CHILDREN

A so far neglected area of research is the development of logical time concepts in late childhood and adolescence. Piaget’s end-state description of logical time seems to be approximated by the competence of the 9-year-old. At least i n the realm of observable events the concrete operational child can integrate discrete temporal series into a unitary sequence from which he can mutually reckon relations of succession and duration. However, the application of these capabilities to more remote content areas may take place in the succeeding years. K . C. Friedman (1944a, 1944b) found improvement from Grades 4-7 in children’s ability to order a set of important events in the life of an imaginary child. Other progress was seen i n the ordering of events of the calendar and events in history. The latter clearly require not only knowledge of particular facts but also knowledge of conventional time schemes. Inhelder and Piaget’s (1958) work on adolescent thought suggests several ways in which logical time concepts might be expected to change during that age period. First, since formal operations can be “content free” (Flavell, 1963) we would expect the adolescent to perform operations upon hypothetical temporal series. He may become interested in successions of infinitesimal magnitude or

280

Willirirn J . Friedrnun

infinitely long durations. Second, as the adolescent comes to differentiate the real from the possible he may begin to consider potential aberrations of logical time, for example, reversing or speeding the flow of time. The achievement of formal thought may also be a prerelquisite for understanding temporal systems that violate logical time, for example, Einstein’s relativity theory (see also Riegel, 1977).

111. Conventional Time Conventional time refers to a cultural product, the symbols and symbol systems used by a group of people to designate temporal cycles and sequences. Particular time systems vary across cultures and clearly influence the nature of the time concepts which members of a society can attain. The developmental literature to be considered is concerned exclusive!y with children’s understanding of the conventional time systems used in Western society. For historical reasons these systems have varying and somewhat conflicting featbres. Some systems (e.g., time of day, seasons, months of the year) are linked to natural cycles; others (e.g., days of the week, year numbers) are entirely arbitrary; and still others (yesterday, today, tomorrow) require subjective definition. Some have recurrent elements (times of day, days of the week) and others are strictly sequential (dates). This complex of features would be expected to present the child formidable problems and, indeed, it is not until adolescence that the cultural end state is attained. The acquisition process, however, appears less a random series of pitfalls than an orderly progression, with considerable consistency of age changes from study to study. Lt is, therefore, Liseful to relate the progression to concomm itant cognitive development. The studies to be reviewed vary in methodology and in particular in the rigor with which results were analyzed. It may be helpful to point to a few general problems for the interpretation of the data. First, the reporting of ages of acquisition is erratic, with the percentage of subjects meeting criteria varying and occasionally unspecified. Second, subjects differ from study to study in IQ and sometimes in generation. Third, the standards for correct answers vary in stringency, while incorrect answers are often not reported. The failure to report errors is unfortunate since errors are often more sensitive indicators of the child’s true understanding (W. Friedrnan & Seely, 1976). A.

CONVENTIONAL TIME CONCEPTS IN INFANCY AND EARLY CHILDHOOD

I . Early Tense Reference Perhaps the first conventional time system the child encounters is that of tense contrasts in adult speech. Foir this reason researchers have looked for similar

Time Concepts in Children

28 1

distinctions in early spontaneous utterances. In terms of actual tense usage, utterances before 29‘2 years are tense-free. At this time the present progressive form becomes inflected (and probably used to indicate immediate intention), and at about 3 4 years, the past irregular and past regular appear (Brown, 1973). However, other lexical and contextual time cues in early speech have led several authors to postulate an order of emergence of past, present, and future temporal reference. Stern (1930) noticed that his daughter used “at once,” “now,” “tomorrow,’’ and “soon’‘ half a year before using “just then,” and “yesterday,” and concluded that the present and future are indicated before the past. Relying on contextual cues, Lewis (1937) observed that reference to both past and future events emerged during the latter half of the second year. However, Lewis found no evidence that the infant could distinguish past and future, and suggested that they were differentiated only later under adult influence. Ames (1946) collected spontaneous utterances and answers to questions from 1%- to 4-year-old children and coded the use of time words, tense, and the apparent referent time as determined by the context. She concluded, along with Stern (1930), that words indicating the present (“now,” “today”) precede future words (“in a minute,” “gonna”), which are prior to past words (“yesterday,” “last week”). A similar developmental pattern was found in the percentage of statements referring to each time. Before about 2 years all statements dealt with present events; during the third year future reference became more frequent; and past reference was uncommon before the middle of the fourth year. These data rely heavily 011 the experimenter’s interpretation of the child’s intent and raise the possibility that the distinctions made by the author are not real for the child. Two points support this criticism. First, many of the words referring to future and past are inserted in present or present progressive sentences, so that there is no nonlexical confirmation i n tesne contrast. Second, the children often use time words that are incompatible with the tense: “I had a bath tomorrow,” “We will do it yesterday” (Ames, 1946, p. 116). In general, the claims for an orderly emergence of tense reference are dubious. It is more likely, as Lewis (1937) has proposed, that this time system is not embedded in the intuitions of early language learners but is acquired later as an aspect of conventional time.

2 . Other Conventional Concepts During the preschool period the child comes into increasing contact with words and objects which represent conventional time systems. At the same time his speech comes to include a variety of words and phrases concerning days, minutes, seasons, and clocks (Ames, 1946). Much of children’s thinking about conventional time during the 3- to 5-year age period is similar in form to other world concepts studied by Piaget (1929). For example, young children often attribute physical reality to abstract entities. A

282

William J . Friedman

5-year-old claimed, “Time is the minutes that go: in the morning they come, in the night they go. . . . The minutes come from the clock.” A subject of 4 years stated, “Time. . . it means the clock strikes” (Bromberg, 1938, p. 144). When asked to identify the season on a warm March day young children are likely to assert that it is summer because it is so warm (Oakden & Sturt, 1922). Springer (1952) asked children why a clock has two hands. Four-year-olds were apt to answer transductively (“Because it is like a girl”) or circularly (“Has to because it’s a clock”). The 3- and 4-year-olds, who were unable to tell time, still knew it is appropriate to look at the clock when wondering if it is time for some activity (Schechter, Symonds, & Bernstein, 1955). Young preschool children begin to recognize certain recurrent times, apparently by anchoring them to personal action. “When” or “what time” questions are answered by 3- and 4-year.olds by stating a concurrent activity or a sequence of activities leading up to a time (Ames, 1946; Oakden & Sturt, 1922; Schecter et al., 1955; Springer, 1952). Children use terms which designate times (“this noon,” “tonight,” “Sunday morning”) before they can adequately describe the referent (Ames, 1946). Questions about duration are more difficult for children of this age (Ames, 1946; Oakden & Sturt, 1922). Occasionally, duration questions are answered by listing intervening activities--“We take a rest and then.. .” (Ames, 1946, p. 113) or “Two more ‘sleeps’ until I go to school”-but more often answers give some unreasonable estimate of minutes or hours. in a study in which children were conditioned to respond to duration, 4- to 5-year-olds frequently described a 4-second waiting period as “a1 couple of minutes” or “10 minutes” (W. Friedman, 1973). These children appear to use a verbal formula which links conventional units with a “how long,” question. The studies reviewed thus far depended upon children’s answers to questions posed by the experimenter. Nonverbal methods can provide additional information about the strengths and limits of the conventional time concepts of preschool children. W. Friedman (1977) employed a set of card-arrangement tasks to test 4to 10-year old’s understanding of conventional and nonconventional cycles and series. Children were asked to order, judge, or coordinate sets of cards that named or depicted events i n a temporal cycle. Some of the card sets represented times of day, others annual events, and still others days of the week or months of the year. Below about 5 years of age subjects generally failed to show an understanding of the temporal order of the sets. However, by 5 years most children succeeded on a task requiring them to order pictures of familiar events that normally occur at different times of day. It is not clear to what extent this performance depends upon conventional representations, but it does show an early ability to organize the events of an extended period of time. One of the major limitations of this ordering ability, and one which applies as well to the succeeding age period to be considered, is the inflexibility in accepting multiple

Time Concepts in Children

283

starting points of cycle sets such as the times of day set. The author assumed that this inflexibility reflects an inability to conceptualize the recurrent features of temporal cycles. B . CONVENTIONAL TIME CONCEPTS OF 5- TO 7-YEAR-OLDS

1. Identtfving Points in Time and Learning about the Clock

Much of the progress of 5- to 7-year-olds in understanding conventional time can be viewed as learning to identify and order hours, days of the week, months, and seasons. By about 5 years most children can name days of the week and know that there is no school on Saturday (Ames, 1946; Bradley, 1947; Oakden & Sturt, 1922). While some 4- and most 5-year-olds use “yesterday,” “today,” and “ tomorrow,” 6-year-olds are the first to specify them using days of the week (Schechter et a l . , 1955). The 6 to 7-year-olds can generally list the months and seasons of the year aloud and identify the current month (Ames, 1946; Oakden & Sturt, 1922; Schechter et al., 1955). The hours of a number of important daily activities (getting up, going to school, eating lunch, etc.) are also learned during this period, but most children have difficulty telling time and setting the clock until about 7 or 8 years (Ames, 1946; Oakden & Sturt, 1922; Springer, 1952). Six-year-olds are generally successful at identifying whole hours, while parts of hours and questions about the composition of an hour present difficulties (Ames, 1946; Springer, 1952). These difficulties suggest that the latter depend on more complex abilities than the former. Children may be able to learn to tell whole hours by practice in recognizing static states and by relating whole hours to counting numbers. They may then expand their competence to several other times by applying generative rules, such as “5 minutes before/after -,” or “30.” A full understanding of the clock, however, may presuppose a logical temporal achievement, the ability to coordinate unequal velocities. Since the distance of the clock’s circumerence is fixed, the faster moving hand must complete the cycle i n a shorter time. Furthermore, there is a fixed relationship between advancement of the minute hand and advancement of the hour hand, one which is not perceptible but must nonetheless be conceptualized. Piaget’s (197 I ) work on the coordination of velocities suggests that this level of understanding of the clock may be an achievement of the concrete operational stage.

2 . Construction of Temporal Order and Duration A number of the studies using verbal methods showed progress during the 5- to 7-year age period in children’s ability to order days of the week, months, and seasons. W. Friedman’s results (1977) showed similar age trends. In addition, a composite index of success in ordering several of the temporal sets was significantly related to performance on a Piagetian spatial seriation task, even when the variance attributable to general age improvement was partialled out. This correla-

284

Willicim J . Friedman

tion raises the possibility thal common cognitive structures serve the construction of spatial and temporal order. One of the limitations apparent in verbal recitations of these temporal cycles is that they are usually treated as fixed lists and not cycles. Children begin with the conventional first element and end with the conventional last element, and when asked to start near the end of the list (e.g., November, December, January.. .) may state that such a procedure would be incorrect. Similar limitations were apparent i n W. Friedman’s nonverbal tasks. In the case of times of day. seasons, holidays, days of the week, and months certain responses thought to indicate an understanding of cyclic recurrence became predominant only at about 8-9 years. In addition to learning to identify and order conventional time markers, 5- and 7-year-olds begin to relate units of duration to counting numbers. While 3-year-olds are able to cite their age, 5-year-olds correctly tell the age at their next birthday, and 6-year-olds know what their sge will be in 4 years (Ames, 1946; Bradley, 1947). The 6- and 7-year-olds frequently and persistently ask how many minutes there are until an activity of interest begins. However, their estimations of durations in conventional units are s h l l markedly inaccurate (Bradley, 1947; Qakden & Sturt, 1922). C. CHANGES IN MIDDLE CHILDHOOD AND ADOLESCENCE

In several studies using question-and-answer procedures the authors describe improvements during the 7- to 9-year age period in children’s understanding of the calendar. For example, 8-year-olds are able to identify the day of the month and the year and know the number of days in a week (Ames, 1946; Bradley, 1947; Oakden & Sturt, 1922). It seems likely that a full understanding of the calendar depends on the ability to construct recurrent cycles and to embed one cycle within another. Children show progress on both of these abilities during the same age period (W. Friedman, 1976, 1977). By 8 4 % years most children distinguish correct from incorrect orders of times of day, days of the week, months, seasons, and holidays, but accept several starting points as correct. This index of the ability to conceptualize cyclic recurrence shows significant correlations between performance on several different cycles. Eight-year-olds are also able to make accurate predictions about the operation of a physical device which demonstrates the relationship of one temporal cycle, the times of day, embedded within another cycle, the days of the week (W. Friedman, 1976). Also during this age period children begin to understand and use the principles of measuring time. For example, Goldstone and Goldfarb (1966) found that 8-year-olds were the youngest subjects to measure an interval by counting, and in a conditioning to duration task, 7- and 8-year-olds, i n contrast to younger subjects, counted in order to remember the interval and explained how they could use a watch to perform better (W. Friedman, 1973). Piaget (197 1) found that 7-

and 8-year-olds are the first to assert the uniformity of motion of sand through an hourglass regardless of their own rate of activity. Eight-year-olds are also able to predict that the same amount of sand will run through during a given interval of clock time. Piaget believes that the cognitive operations underlying logical time also allow the construction of iterable temporal units and that this quantification is instrumental in solving these uniform velocity problems (Piaget, 1971). The ability to iterate a standard duration can, in principle, be generalized to include any conventional unit of duration. The assimilation of conventional units to the mental operations of time measurement is an integration of a culturally provided system and an individual’s cognitive competence; as it occurs, the child’s concepts of minutes, days, and years become enriched. For the younger child the terms could be inserted only in a verbal formula to express duration. Now the terms connote socially shared units of uniform duration. While temporal measurement allows a new level of understanding of conventional time, the child’s ability to use w i t s of duration is still very much dependent upon concrete, observable change. First, his accuracy in making duration estimates without a clock is poor until after age 10 (Bradley, 1947; Oakden & Sturt, 1922; see also Section IV). Second, until adolescence children are likely to believe that their age is altered when clocks are set ahead or back for daylight saving time (Michaud, 1949, cited in Fraisse, 1963). There is relatively little information on the development of conventional time concepts between about 8 years and adolescence. Apparently during this period the scale of historical time becomes increasingly meaningful (Jahoda, 1963). For example, by about 9 years the child comes to realize that Robin Hood lived before his grandmother’s time (Bradley, 1947). Preadolescents may also be the first to coordinate certain ordinal and cyclical features of time systems. Bradley (1947) found that not until age 12 could 75% of the subjects name the seventh month, and K. C. Friedman (1944a) observed that fourth through sixth graders often gave confused orderings of annual events because they thought the beginning of the school year should come before the end. Similar problems were shown in W. Friedman’s (1976) study but only with somewhat younger subjects. By 10 years most children could coordinate a set of cards representing events i n the school year and a set naming the months. The W. Friedman study indicated other abilities that were predominant i n the 10-year-old but not i n the 8-year-old group. The older subjects were the first to consistently order the seasons and holiday sets using an implicit January to December order. This achievement suggests that diverse annual events are integrated into a unitary, conventionally determined representation. Ten-year-olds are the youngest group to order the set of relational days (day before yesterday, yesterday, today, etc.) alone; 8-year-olds could arrange them correctly only by coordinating the set with another set naming the days of the week. The 10-year-

286

Willinm J . Friedman

old subjects also showed considerable progress in making quantitative predictions using the device that demonstrated the relation between daily and weekly cycles.

IV , Experiential Time Experiential time refers to the perception and memory of succession and duration i n the absence of logical or conventional cues. The negative aspect of this definition corresponds to the usual operational distinction between perception and cognition. Perceptual phenomena are distinguished from more purely cognitive processes by their persistence in the absence of logical cues. By analogy, experiential time is assumed to allow consistent temporal judgments when measurement or other logical solutions are precluded. In contrast to most perception paradigms, however, temporal judgments of more than very brief stimuli lack the immediacy of the usual visual or auditory judgments because, obviously, the stimulus is extended i n time. Experiential time may therefore by assumed to depend on perceptionlike and memorylike processes. In the present chapter I shall deal with only those aspects of experiential time that have been treated in the developmental research literature-judgments of about 1 second or longer. While logical and conventional time have cultural forms that serve as end-state descriptions for a developmental analysis, experiential time has no such obvious referent. Two approaches are suggested by the adult literature on time judgments (Fraisse, 1963; Ornstein, 1969). Judgments may be assessed as more or less veridical by clock time or, alternatively, susceptibility to certain illusions may be considered characteristic of mature functioning. Both of these end states are compatible with one another and with several interesting developmental patterns. Increasing veridicality is certainly the rule in perceptual development; however, there are some perceptual tasks in which older subjects are more likely to make errors, presumably of overcompensation, than younger subjects (Piaget, 1969; Wohlwill, 1970). In either case it is important to determine the age period during which accuracy or susceptibility to effects changes most if inferences are to be drawn about the contributing factors. Because the definition of experiential time is partly methodological, it is important to be explicit about the methodology of the research on time judgments. We will restrict the discussion to experiments in which children make temporal judgments of stimuli presented in a controlled setting (for studies of children’s judgments of real-life durations, see Cohen, 1964; K . C. Friedman, 1944a). Such judgments are usually in the form of one of the temporal psychophysical methods: comparison, estimation, production, or reproduction (Fraisse, 1963). Some of the methods are easier to use than others with young children, but only a single sludy has been conducted with children of less than 4

Time Concepts in Children

287

years. Thus, we are presently limited in our ability to detect developmental changes in experiential time during infancy and very early childhood. A second methodological contrast is that between simple judgment tasks, in which relatively meaningless and constant stimuli are presented during the intervals to be judged, and complex judgment tasks, in which either the stimuli vary in rate of change, velocity, or content, or the subject’s activity is varied. These two methods will be considered separately because they are used to study different illusions. A thud methodological issue, and one which unfortunately has been largely ignored in the research to be reviewed, is the possible differences in judged duration between tasks that focus the subject’s attention on his experience during the interval and tasks that focus attention on the memory of the interval. Piaget (197 1) argued convincingly that some experiences that seem interminable at the time are remembered as being brief, and Ornstein (1969) has demonstrated the malleability of remembered duration. In the present studies the issue appears to be confounded with task features, such as the length of the interval and immediacy of the judgment, and subject attributes, such as the age-linked ability to distinguish between the immediate and retrospective perspectives. Because of such confounds the issue cannot be systematically discussed here, but nevertheless remains in the background. A.

SIMPLE JUDGMENTS

The methods of comparison, estimation, production, and reproduction have all been used in simple judgment tasks to determine children’s accuracy and susceptibility to certain adult illusions. Comparison and reproduction procedures have generally been intended to tap an early and fundamental competence, while tasks that require production or estimation in minutes and seconds necessarily assess both experiential and conventional aspects of the child’s competence. Three temporal illusions found in adults (Fraisse, 1963) have been sought in children: (1) the tendency to overestimate short durations while underestimating long ones; ( 2 ) the tendency to overestimate durations filled by a stimulus (filled) and to underestimate intervals that have only the beginning and end indicated (empty); and (3) the tendency to overestimate the durations of auditory stimuli and underestimate those of visual stimuli. 1 . Judgments of Young Children Even the youngest subjects tested make generally accurate judgments but succumb to certain illusions suffered by adults. E. Friedman (1976) trained a group of 2%- to 5Yi-year-olds (mean age 4 years) to reproduce 15-second intervals by providing two demonstration trials and four feedback trials using a stopwatch. On six subsequent test trials without feedback the group produced a mean esti-

288

Willic~rnJ . Frieclrncin

mate of 15.5 seconds with a standard deviation of 4.5. A control group that had received only a single demonstration and no feedback trials produced significantly less accurate judgments. Crowder and Hohle (1970) taught 5-, 7-, and 9-year-olds to reproduce intervals of 2.7 and 5.4 seconds, which corresponded to the time it took for a concealed toy lion to reach his den. In a feedback condition a screen was removed to show the subjects where the lion was when they stopped his motion, and they were told to stop sooner or later the next time. Control subjects did not receive this information. Between reproduction trials the screen was removed and children i n both groups were able to watch the lion move for the correct amount of time. Ely the end of 15 trials, experimental subjects at each age level were responding accurately, as were the 7- and 9-year-old control subjects. However, the 5-year-old children without feedback continued to perform erratically. Gardner (1966) found the 5- to 12-year-olds did not improve significantly in the accuracy of reproducing brief auditory- and visual-filled intevals. There was no significant interaction of age with the modality effect (overestimating auditory stimuli, underestimating visual stimuli). Within the limits of negative inference, these findings tend to support the accuracy of 5-year-olds’ judgments and indicate their susceptibility to the auditory-visual illusion. Gardner’s graphs also indicate that 5-year-olds overestimated short intervals (1.5 seconds) and underestimated long intervals (4.5 seconds). Hermelin and O’Connor ( 197 I ) demonstrated the ability of 5-year-olds to compare the durations of successively presented visual stimuli, but only under very limited conditions. The subjects were able to correctly respond “same” or “different” significantly ablove chance level only if (1) they were cued to the relevant dimension-length of time, (2) the duration differed, and (3) the short stimulus ( 2 seconds) preceded the long stimulus (6 seconds). Berndt and Wood’s (1974; see also Section 11) 5-year-olds were successful in comparing both successive and simultaneous filled auditory intervals ranging from 2 to 9 seconds. AS a selection procedure for another task, Piaget (1971) asked 4- and 5-year-olds to compare empty intervals of 20 and 25 seconds. Though the results were not tabulated, it appears from the cases reported that this was not an insurmountable task for these young subjects. Six-year-olds have been the youngest subjects tested with several more difficult procedures, using estimation, production, or reproduction of longer durations, and their performance is typically given a negative characterization. Smythe and Goldstone (1957) used a I-second estimation task; Goldstone, Boardman, and Lhamon (1958) used a 30-second production task; and Fraisse (1948, cited in Fraisse, 1963) and Fraisse and Orsini (1958, cited in Fraisse, 1963) used 20- and 30-second reproduction tasks. All described 6-year-old’s performance as extremely variable. Fraisse (1948, cited in Fraisse, 1963) found that reproductions of a 20-second interval ranged from 1 to 60 seconds. How-

Time Coriceprs

iti

Children

289

ever, these 6-year-olds did show evidence of susceptibility to the third illusion: They tended to overestimate intervals filled by sounds and underestimate empty intervals. By the age of 7 , subjects learned to reproduce short intervals without feedback (Crowder & Hohle, 1970) and could be taught to reproduce a 30-second interval with greater accuracy than untrained adults when a training procedure using feedback was provided over a 3-week period (Orsini, cited in Fraisse, 1963). Performance of subjects of this age is still quite variable i n estimation and production tasks (Goldstone et a l . , 1958; Goldstone & Goldfarb, 1966; Smythe & Goldstone, 1957); however, 7- to i2-year-olds d o show a consistent overestimation of auditory stimuli and underestimation of visual stimuli (Goldstone & Goldfarb, 1966).

2. Judgments of Children of 8 Years and Older Smythe and Goldstone (1957) and Goldstone et al. ( 1958) found 8 years of age to be a transition point in several respects. First, there was a considerable decrease in the variability of the estimates from 6 to 7 or 8 years. In addition, when permitted to count in seconds, 8-year-olds produced intervals that were more accurate than those of younger subjects. These subjects were also the youngest to be able to count seconds silently, perhaps indicating practice in measuring durations in this way. Finally, 8-year-olds, but not younger subjects, gave more accurate estimates of a second after several standard I-second intervals were demonstrated. Both studies showed continued increases in accuracy and decreases in variability through the age of 14 years, when adult levels of performance were reached. Elkine (1928, cited in Fraisse, 1963) using longer intervals showed that accuracy of estimation in minutes and seconds improved at least through the age of 15. Gilliland and Humphreys (1943) compared performance of fifth graders with college subjects while varying several task dimensions. The authors found that the younger subjects were less accurate but that both groups benefited from counting. For both groups reproduction was easier than estimation or production. The latter finding parallels the finding that reproduction methods are easiest to use with younger subjects. Finally, the fifth graders as well as the adults showed a tendency to overestimate the shorter intervals and underestimate the longer intervals. The 9-second intervals were overestimated and 180-second intervals underestimated by both groups, but the magnitude of the error was about double for the younger group. 3 . Developmental Trends Except when assessed by estimation and production tasks that require knowledge of conventional units of time, there is evidence for an early ability to judge intervals accurately and for the early presence of bias i n the direction of adult illusions. The 4- through 6-year-olds, given sufficient practice and feedback, are

290

W i h m J . Friedman

able to accurately judge brief intervals (Crowder & Hohle, 1970), and 7year-olds can be trained to reproduce a 30-second interval with only a 5-second error. While there is considerable improvement even in reproductions beyond the age of I 1 (Gilliland & Humphrey, 1943), it appears that duration can first be judged as early as subjects have been tested. In addition, 5- and 6-year-olds’ judgments are affected by some of the same variables as those of adults and in a similar direction: overestimation of auditory relative to visual stimuli (Gardner, 1966), overestimation of filled and underestimation of empty intervals (Fraisse, 1948, cited in Fraisse, 196311, and overestimation of short and underestimation of long durations (Gardner, 1966). The last finding depends upon an adequate distinction between short and long, one which is difficult to arrive at empirically in either the developmental or adult studies (Doob, 1971; Gardner, 1966; Gilliland & Humphreys, 1943; see also Ornstein, 1969). It is not possible to discern developmental trends in the magnitudes of the length of duration illusion or the filled versus empty interval illusion, but the modality illusion appears fairly constant in magnitude at least from 8 through 13 years (Goldstone & Goldfarb, 1966). E;. COMPLEX JUDGMENTS

Several investigators have studied the temporal judgments of children by controlling complex stimulus features or the subject’s activity. These experiments deal less with veridicality than the influences of four illusions present in adults (Fraisse, 1963; Ornstein, 1969; Piaget, 1969): (1) interesting tasks seem to go more quickly than dull task.s, (2) difficult tasks seem to go more quickly than easy tasks, (3) vigorous activity seems to take longer than calm activity, and (4) stimuli moving rapidly seem to persist longer than those moving slowly. The illusions are difficult to dissociate empirically; indeed each of the authors cited above interprets all four illuisions as different aspects of a single illusion. Related to this problem is the imprecise formulation of each illusion, especially when viewed in a developmental context. The dimensions of how interesting or difficult a task is or how many changes are perceived are clearly intertwined with cognitive developmental variables, as is the concept, though apparently not the intuition, of velocity. Given these limitations, the developmental course of the illusions will be examined.

I . Children’s Judgments Piaget (197 1) tested children of 4 and 5 years and older on four complex judgments tasks. In each task the subject was asked to compare the length of two durations that were objectively equal. In the first procedure children were instructed to draw strokes rapidly during one interval and slowly during another.

Time Concepts in Children

29 I

Children below the age of 7 apparently focused on the tangible results of their actions, that is, the numerosity of the strokes, to the exclusion of introspective assessments. These subjects consistently claimed that the rapid drawing interval took longer and they sometimes used the greater number of strokes as a justification. About half of the 7- and 8-year-olds were influenced solely by the results of their actions. The remaining subjects in the older groups were able to equate the durations, often adding that the slower work seemed to take longer. With both the successful and unsuccessful subjects it seems that the experience of duration is not being tapped directly. The unsuccessful subjects appear unable to differentiate between subjective impressions gained during the task and concrete aspects of the results, while subjects who equate the intervals are probably compensating for their impressions with inferences about the relative velocities of work. In this case it is difficult to distinguish perceptionlike abilities from logical or sublogical functioning. Using a variant procedure Piaget found that when asked to draw at each speed for an equal time many subjects judged the rapid drawing interval as longer. Though this procedure was a more direct test of the subject’s experience, it was difficult to use with younger children. A second task required the comparison of intervals of fast and slow metronome motion. Two-thirds of the 4- to 8-year-olds but only one-sixth of the 10- to 12-year-olds judged the rapid interval as longer. The older subjects tended to equate the durations or to claim that the slow condition was longer. Using an analogous procedure with beads moving at different speeds for 5 seconds, Piaget (1969) found that adults erred i n the same direction as the young children in the above study: The fast presentation was overestimated. This is, perhaps, another instance of an illusion being compensated by older children but demonstrable even in adulthood under appropriate circumstances. Subjects in the third task transferred small pieces of lead or wood from one container to another with a pair of tongs. The lead pieces were not only heavier but more difficult to grasp because of their triangular shape. Subjects predictably transferred more wood than lead pieces during equal durations. In a final task children sat either doing nothing or looking at an amusing picture. In the transferral task 4- to 8-year-old subjects seemed to focus on their impressions during the action, since they judged the lead task as longer. Some attributed the longer time to the fact that the lead was heavier, indicating the influence of a concrete attribute. Other subjects incorrectly claimed that the lead interval was longer because more pieces were transferred. The 9- to 12-year-olds compensated for similar impressions, judging the intervals equal and arguing that although the lead transfer seemed longer, fewer pieces were moved. In this case difficulty seemed a more potent factor than the vigor of the subject’s activity. Finally, i n the waiting task, where no concrete features were available to either compete with the younger subject’s impressions or allow logical compensation by the older subjects, the illusions persisted from 4 years through 12 years of age.

Wi//iiiin J

292

. Friedinon

Younger and older children alike judged the boring condition to be longer than the interesting condition. Axel (1924, cited i n Fraisse, 1963) investigated the effect of 9- to 14-yearolds’ activity on their estimations of a 20-second interval. The tasks varied in both difficulty and interest, but the two dimensions were not entirely separable. For example, counting by increments of 7 was probably more difficult and more interesting than crossing out signs or writing letters “1’s.” However, waiting may have been both more difficult and less interesting than the other tasks. The results showed that the 9-year-olds drastically overestimated all the intervals with respect to clock time, but the order of the magnitude of the means of the different tasks was the same as at t h e other ages: Mental arithmetic received the shortest estimate and the empty duration the longest. This pattern is consistent with both illusions if the mental arithmetic task is interesting but not difficult. Clearly, this kind of hairsplitting will not be fruitful until the principles are better formulated. But the study does make plausible the assumption that the phenomena Piaget ( 197 I ) observed in younger children also characterizes pre- to midadolescents.

2. Drveloimentul Trends The complex judgment studies display, in general, a trend similar to that of the simple judgment studies. Evidence for each of the illusions is present for the youngest children tested and the effects appear to persist from childhood through adulthood. The 4- and 5-year-olds are affected by difficulty, velocity of changes, and interest value i n much the same manner as older subjects (Piaget, 1971). In some cases veridicality increases with age at the expense of the illusion, but this probably depends on the presence of concrete cues, such as the number of pencil marks o r the number of objects moved, to allow the subject to deduce a compensatory relationship. I n Piaget’s (1971) metronome and waiting tasks, where visible products of the activity were not available, the illusions persisted in older subjects. According to Piaget compensation also presupposes two cognitive achievements: ( 1) the ability to differentiate between personal impressions of duration gained during the interval and the tangible results, and ( 2 ) the ability to convert the rate of change to a velocity and recognize the inverse relation between velocity and duration. In the latter instance, concrete operational thought, Piaget argues, allows the conclusion that a greater number of products can be produced at a faster rate and yet the durations be equal. Thus, Piaget believes that the experience of time, like other perceptual judgments, has a constructed, cognitive element. C

DEVE.LOPMENT OF EXPERIENTIAL TIME

In the present chapter we have defined experiential time, by analogy to perception, as independent of logical cues. In suitably impoverished tasks where

cognitive solutions are precluded the illusions are found at an early age and persist into adulthood. Taking a broader perspective, however, it is obvious that progress in logical and conventional time concepts have major consequences for the experience of duration. In the work described above, Piaget has indicated logical influences i n older children. Craik and Sarbin (1963) have shown how conventional cues, in the form of clock rate, affect judgments and personal tempo responses of adults. Older children and adults are neither slavish to their impressions of duration nor to logical or conventional time. Rather, they are able to distinguish between subjective and objective aspects of duration and use whichever framework is appropriate. Adults discuss their feelings of duration, particularly when their impressions are inconsistent with objective reckoning, but they are also able to make a variety of accurate inferences when necessary. In summary, certain aspects of experiential time are well developed by early childhood, but performance in a variety of judgment tasks continues to improve through midadolescence. Each of the seven duration illusions considered was demonstrated in young subjects, as was the ability to accurately reproduce short intervals. There is also some evidence that 5-year-olds can compare the length of pairs of 20- to 30-second empty intervals. Other studies tended to show increases in children’s ability to relate their judgments to clock time, mainly between 8 and 15 years. However, with extensive practice even 7-year-olds could accurately reproduce a standard 30-second interval. Probably practice with the clock is the most important factor in the normal improvement in making judgments in conventional units, although the units must surely be more meaningful to the concrete operational children who understand the principles of measuring time. By contrast, compensation for illusions may depend mainly on the attainment of cognitive structures that allow the regulation of judgments. Finally, there is a hypothetical case in which an illusion would be expected to increase with cognitive development. Ornstein ( 1969) attempted a theoretical synthesis of several of the illusions discussed above within an informationprocessing framework. By influencing dift‘erent adults’ understanding of identical stimuli he demonstrated that subjective duration of an interval depends on the amount of “stored information” about that interval. If seemingly random stimuli could be conceptually integrated by a simple rule, the subjective length of the duration should be shortened. The model is, obviously, relevant to retrospective, not concurrent judgments. The theory is of developmental interest because it implies that many events that are meaningful for adults but seem random for children may be experienced as endless for the latter and brief for the former. The phenomenon could be experimentally demonstrated by comparing preoperational and concrete operational subjects’ judgments of the relative length of two intervals when the stimuli of one interval could be organized according to principles of classification or seriation. Ornstei n’s theory would predict that the concrete operational subjects, but not the preoperational subjects, would underestimate the organizable interval.

William J . Friedman

294

V.

Summary and Conclusion

Research on the development of time concepts has revealed somewhat different patterns in the cases of logical, conventional, and experiential time. The experiential mode, as studied by the perception of duration, appears well developed in young children. Evidence suggested that each of seven illusions demonstrated in adults also iinfluences the duration judgments of preschool children. The 5- and 6-year-oldr; successfully learned to reproduce and distinguish certain intervals, though the ability to estimate in conventional units usually required an additional 10 years to reach adult levels. Conventional time probably has little influence on the thought of children younger than 5 years. Younger children begin to use tense contrasts and to represent the order of familiar activities but their understanding of a variety of temporal terms is characteristically action-bound. Over the next several years children learn several conventional series and associate time units with counting numbers; however, it is not until about 8 years that time measurement and some of the cyclic aspects of the calendar are mastered. During middle childhood, children become able to coordinate multiple cycles and series and to conceptualize cyclic recurrence, but they are generally unable to distinguish arbitrary from natural features of conventional time before midadolescence. Research i n the development of logical time shows that before about 8 years succession and duration are unstable concepts, and that children are misled by perceptual aspects of many tasks. From infancy onward there is a purposeful ordering of action, and by 5 years children can imitate and verbalize relations of temporal succession. Comparison of duration of motions and verbalization of relative duration, however, are still quite poor at ages 5 and 6. These patterns indicate that experiential time is a primitive mode, while conventional and logical time are poorly developed before the age of about 8 years. The 8-year-old transition in the latter two may depend on concurrent cognitive developmental changes, or alternatively on cultural experiences first provided at this age. The cultural explanation seems less likely for several reasons. First, where specific verbal and nonverbal training is given to younger children, as in Weil’s ( 1975) study of duration discrimination, there is negligible improvement. First and second graders are more likely to benefit from the training and older children improved dramatically. Second, children beginning school early, such as the subjects of the Ames (11946)study, showed about the same rate of progress as children tested upon entering school at the usual time. Third, conventional time words and temporal relational terms are used spontaneously but inaccurately by 5-year-olds, indicating that young children have considerable exposure to conventional and logical time before they enter school but that exposure alone is not enough. Finally, it is obvious that many of the concepts on which 8-year-olds improve, especially those studied by Piaget (1971), are simply not taught sys-

Time Concepts in Children

295

tematically in school. First and second graders rarely receive instruction on inferences from paired velocities or measuring time. Further indirect support for the influence of general cognitive changes on temporal abilities comes from W. Friedman’s (1977) finding that ordering ability in temporal tasks is correlated with seriation performance even when the variance attributable to overall age improvement is partialled out. Piaget’s (1971) research on the understanding of succession, duration, and measurement and W. Friedman’s (1976, 1977) work on temporal cycles both suggest ways in which abilities essential to the construction of logical time may contribute to the understanding of conventional time as well. Furthermore, Piaget’s work on the regulation of duration illusions indicates an important contribution of concrete operations to experiential time. Older children in these studies could differentiate between their own impressions and objective measures of duration and thus compensate for the former where possible. All this seems to suggest an integration of logical, conventional, and experiential time during middle childhood but an integration in which impressions are subordinated by the child to logical inference and measurement, and particular aspects of conventional time are subordinated to the general concepts of duration and measurement. (The latter is shown by the finding that the 8-year-old realizes that he cannot time events only by the clock but by the hourglass or sun as well.) This developmental pattern is well-described by Werner’s orthogenetic principle (Werner & Kaplan, 1967) that development proceeds by differentiation and hierarchic integration. The concrete operational transition includes the differentiation of the three aspects and their integration into a new level of awareness of time. There is considerable progress in the understanding of conventional time systems well into adolescence. Children only gradually come to appreciate the magnitude of intervals of historical time, and certain cycles and series of events are difficult to integrate. Furthermore, the distinction between arbitrary measurement systems and natural cycles may come late. However, these delays may be more the result of limited exposure than of inherent cognitive complexity. Estimates in conventional units also improve through midadolescence probably because of practice in timing events. While the systematic aspects of logical time are mastered in middle childhood, their application to remote content areas may also depend on learning specific facts. We would expect to see an expansion in the range of application of logical temporal inferences through later school years. Hopefully, future researchers will add to our understanding of changes in conventional time concepts during this age period and begin to explore the relative contributions of cognitive development and cultural learning to particular achievements. The clarification of the development of time concepts during late middle childhood and adolescence is but one of a number of problems deserving further

296

William J . Friedman

study. We have noted previously that the limited amount of research with infants and young children precludes an adequate description of experiential time during this age period. We have also observed the need for nonlinguistic studies of temporal ordering ability to bridge the gab between Piaget’s (1954) descriptions of sensorimotor time and his work on temporal seriation in middle childhood (Piaget, 1971). A fourth suggested direction for further work was the application of a cognitive developmental perspective to problems in experiential time. This approach could be used to clarify the relationship that Ornstein (1969) proposed between the cognitive complexity of experiences and their subjective duration. Finally, very little is known about the relationship between spatial imagery and the development of time concepts. In the 1920s Guilford (1926) found that a large proportion of his adult subjects were able to draw spatial figures relevant to various temporal frameowrks and concluded that, “Spatial images, which are probably often of the habitual type, carry the meaning of time and furnish a frame of reference for personal and historical time” (p. 423). W. Friedman (1976) found evidence that by late middle childhood subjects frequently constructed and understood circular spatial representations of temporal cycles. In spite of these suggestive findings, it remains to be shown how important such spatialization is in the mastery of time and hiow the nature of temporal imagery changes with development. REFERENCES Ames, L. B. The development of the sense of time in the young child. Journal of Genetic Psychology. 1946, 68, 97-126. Amidon, A , , & Carey, P. Why five-year-olds cannot understand before and after. Journal of Verbal Learning and Verbal Behavior, 1972, 11, 417423. Berndt, T., & Wood, D. The development of time concepts through conflict based on a primitive duration capacity. Child Development, 1974, 45, 825-828. Brackbill, Y., Fitzgerabld, H. E., & Lintz, L. M. A developmental study of classical conditioning. Monographs of the Society for Research in Child Development, 1967, 32(8, Serial No. 116). Bradley, N. C. The growth of the knowledge of time in children of school age. British Journal of Psychology, 1947, 38, 67-68. Brombeg. W. The meaning of time for children. American Journal of Orthopsychiatry, 1938, 8, 142-147. Brown, R. Afirst languange, the early stages. Cambridge, Mass. Harvard University Press, 1973. Cavanagh, P., & Davidson, M . L . Operant conditioning and secondary circular reactions in human infants. Unpublished manuscript, University of Rochester, 1974. Clark, E. V. On the acquisition of the meaning of before and after. Journal of Verbal Learning and Verbal Behavior, 197 I , 10(3), 266-275. Cohen, J. Psychological time. Scientijic American, 1964, 211(5), 116-124. Coker, P. L. On the acquisition of temporal terms: Before and afrer. Paper presented at the Stanford Child Language Research Forum, April 1975. Craik, K. H ., & Sarbin, T. R . The effect of covert alterations of clock rate upon time estimates and personal tempo. Perceptual and Motor Skills, 1963, 1, 597-610.

Time Concepts in Children

297

Cromer, R. .F. The growth of temporal reference during the acquisition of language. Unpublished doctoral dissertation, Harvard University, 1968. Cromer, R. F. The development of the ability to decenter in time. British Journal of Psychology, 1971, 62(3), 353-365. Crowder, A. M . , & Hohle, R. H. Time estimation by young children with and without informational feedback. Journal of Experimental Child Psychology. 1970, 10(3), 295-307. Doob, L. W. Patterning of time. New Haven, Conn.: Yale University Press, 1971. Ferreiro, E., & Sinclair, H. Temporal relationships in language. International Journal of Psychology, 1971, 6(1), 39-47. Flavell, J. H. The developmental psychology of Jean Piaget. New York Van Nostrand-Reinhold, 1963. Fraisse, P. The psychology of time. New York: Harper, 1963. Friedman, E. The interval judgments of young children. Paper presented at the meeting of the Eastern Psychological Association, New York, April 1976. Friedman, K. C. Time concepts of elementary school children. Elementary School Journal, 1944, 44, 337-352. (a) Friedman, K. C. Time concepts of junior and senior high school pupils and of adults. School Review, 1944, 52, 233-238. (b) Friedman, W. The development of time judgments. Unpublished research, University of Rochester, 1973. Friedman, W. The development of children’s understanding of temporal cycles. Unpublished doctoral dissertation, University of Rochester, 1976. Friedman, W. The development of children’s understanding of cyclic aspects of time. Child Development, 1977, in press. Friedman, W., & Seely, P. The child’s acquisition of spatial and temporal word meanings. Child Development, 1976, 47, 1103-1 108. Gardner, D. B. Intersensory aspects of children’s judgments of short time intervals. Paper presented at the meeting of the American Psychological Association, New York, September 1966. Gilliland, A. R., & Humphreys, D. W. Age, sex, method and interval as variables in time estimation. Journal of General Psychology. 1943, 63, 123-130. Goldstone, S., Boardman, W. K., & Lhamon, W. T. Kinesthetic cues in the development of time concepts. Journal of Genetic Psychology, 1958, 93, 185-190. Goldstone, S., & Goldfarb, J . L. The perception of time by children. In A . H. Kidd & J. L. Rivoire (Eds.), Perceptual development in children. New York: International Universities Press, 1966. Guilford, J. P. Spatial symbols in the apprehension of time. American Journal of Psychology. 1926, 31, 420-423. Hermelin, B. M., & O’Connor, N . Children’sjudgments of duration. British Journal of Psychology, 1971, 62(1), 13-20. Inhelder, B., & Piaget, J. The growth of logical thinkingfrom childhood to adolescence. New York: Basic Books, 1958. Inhelder, B., & Piaget, J. The early growth of logic in the child. London: Routledge & Kegan Paul, 1964. Jahoda, G. Children’s concepts of time and history. Educational Review, 1963, 15, 87-104. Keller-Cohen, D. Children’s verbal imitation, comprehension and production of temporal structures. Paper presented at the meeting of the Society for Research in Child Development, Denver, April 1975. Lewis, M. M. The beginning of reference to past and future in a child’s speech. British Journal of Educational Psychology, 1937, I, 39-56. Longobardi, E. T., & Wolff, P. A. A comparison of motoric and verbal responses on a Piagetian rate-time task. Child Development, 1973, 44, 4 3 3 4 3 7 .

298

William J . Friedman

Lovell, K., & Slater, A. The growth of the concept of time: A comparative study. Journal of Child Psychology and Psychiatry, 1960, 1(3), 179-190. Oakden, E. C., & Sturt, M. The development of the knowledge of time in children. British Journa! of Psychology, 1922, 12(4), 309-336. Ornstein, R. On the experience of time. London: Penguin Books, 1969. Piaget, J. The child's conception of the world. London: Routledge & Kegan Paul, 1929. Piaget, J. The origins of intelligence in children. New York: International Universities Press, 1952. Piaget, J. The construction of reality in the child. New York: Basic Books, 1954. Piaget, J. The development of time concepts in the child. In P. Hoch & J. Zubin (Eds.), Psychopathology of childhood. New York: Grune & Stratton, 1955. Piaget, J. Play, dreams and imitation in childhood. New York: Norton, 1962. Piaget, J. The mechanisms of perception. London: Routledge & Kegan Paul, 1969. Piaget, J. The child's conception of time. New York: Ballantine Books, 1971. Riegel, K. F. The dialectics of time. In N. Datan & H. W. Reese (Eds.), Life-span developmental psychology: Dialectical perspectives on experimental research. New York: Academic Press, 1977, in press. Schechter, D. E., Symonds, M., & Bernstein. I. Development of the concept of time in children. Journal of Nervous and Mentai' Disease, 1955, 121, 301-310. Smythe, E. J . , & Goldstone, S. The time sense: A normative genetic study of the development of time perception. Perceptual and Motor Skills, 1957, 7, 49-59. Springer, D. Development in young children of an understanding of time and the clock. Journal of Genetic Psychology, 1952, 80, 83-96. Stern, W. Psychology of early child,hood up to the sixth year of age. New York: Holt, 1930. Weil, J . Effectiveness of training on the acquisition of an abstract concept. Paper presented at the meeting of the Society for Research in Child Development, Denver, April 1975. Werner, H., & Kaplan, B. Symbolformation. New York: Wiley, 1967. Wohlwill, J . Perceptual development. In H. W. Reese & L. P. Lipsitt (Eds.), Experimental child psychology. New York: Academic Press, 1970. Zelkind, I . , & Sprug, J . Time research: 1172 studies. Metuchen, N.J.: Scarecrow Press, 1974.

AUTHOR INDEX Numbers

in

italics reter to the page\ on which the complete reterences are listed

A Abraham, S . , 9, 13, 32, 50 Abramov, I., 140, 159, 160. 173 Abravanel, E., 118, 169 Acheson, R. M . , 6 , 5 / Adkinson, D., 134, 169 Adrian, E. D., 136, 169 Agarwal, K. N . , 42.54 Alfred, B. M . , 8, 51 Allen, I . , 8, 42, 56 Ames, E. W . , 152, 164, 179 Ames, L. B., 281, 282, 283, 284, 294,296 Amidon, A , , 272, 273,296 Anderson, J. E., 150, 174 Anderson, J. R . , 203, 221,226 Antrobus, A. C. K . , 6, 50 Arabie, P., 222, 226 Aristotle, 144, 169 Ariza, J., 27, 55, 57 Arkhipoya, G. P., 6, 19, 50 Arlitt, A. H . , 122, 169 Armington, J. C . , 131, 136, 167. 170, 176 Arnheim, R., 165, 170 Ascoli, W . , 3, 58, Asenjo, C. F., 32, 5 2 Ashcroft, M. T., 6, 23, 31.50 Aslin, R. N . , 162, 170 Asnes, D., 10, 54 Azrin, N . , 252, 261

B Backstrom, L., 14, 50 Baertl, J . M . , 3 0 , 5 1 Bailyn, B., 235, 261 Baker, P. T., 6, 53

Bakitova, 2 . . 19. 50 Balazs. E. A , , 130, 170 Baldini. G . , 15, 59 Baldwin, J . M . , 119, 120, 121, 131. 155, 157, 163, 170, 189, 191, 192, 194, 195, 212,226 Ballester, D., 27, 51 Baltia, A , , 15, 53 Bandura. A., 216, 217,226 Bane, M. J . , 232, 261 Banbegyi, M . , 34, 52 Banik, N . D. D., 18, 21, 23, 27, 34, 38, 39, 42, 50, 51 Banks, M. S . 128, 131, 170, 178 B a j a , I.. 27,51 Barnet, A. B., 131, 136, 167, 170, 176 Bartelson, C. J . , 160, 170 Bartlett, F. C . , 193, 198, 202, 203, 208, 218,226 Baumel, H. A , . 119, 176 Baurnrind, D., 254, 256, 257, 258,261 Bayer, R. H . , 221, 222,227 Baylor, G. W.,64, I13 Beales, R. W . , 238,261 Beare, A. C . , 143, 144, 145, 149, 170 Bearzotti, L., 15, 52 Beck, G. J . . 25. 51 Beck. R., 232, 261 Becker, J . D., 65, 113 Becker, W . C., 254. 259. 261 Beebe-Center, J . G . . 159, 170 Belniont, J . M . . 155, 170 Belogorskii. V . Y . , 34, 55 Belousov, A. 2.. 19, 51 Benson, K . , 154. 165, 172 Berg, W . K., 134, 169 Berger, P., 241. 242, 261 Beri, S . . 40. 53 299

300

Author

Berkeley, G . , 118, 163. 170 Berlin, B., 145, 152, 159, 170 Bemdt, T., 278, 288, 296 Bernstein, I..282, 283. 2YX Beny,F. B.,5,6, 7 . 8 , II. 13.21. 30.51 Billewicr. W. Z., 6. 56. 58 Binet, A., 122, 170 Birkbeck, J . A , . 8, 51 Blanca-Adrianzen, T., 30, 51 Blanco, R. A,, 6. 51 Blatt, M . , 74, I14 Boardman, W. K . , 28X. 289, 297 Bobrow, D. G . , 65, 1 1 1 Boettner, E. A , . 130, 170 Bogachanov, N . D., 35,54 Bond, E. K . , 126, 170 Boothe, R., 137, 139, 170 Bornstein, M . H.,119, 128. 129, 130. 131. 137, 138. 139, 140. 142, 143, 144. 145, 146, 148, 149. 151. 152, 153, 154, 155, 156. 157, 1.58. 159. 160. 161, 164, 165, 166. 167. 170, 171 Boring, E. G . , 185.226 Borstelmann, L. J . , 234, 245, 261 Bosse, K. K . , 120, 158. 175 Bouchalova. M . . 18. 53 Bourne, L. E . , 165, 171 Boutourline, E . , 14, 51 Bower, G. H.,78. 114, 221,226 Bower, J. H., 203, 226 Bower, T. G . R.. 118. 123. 171 Boycs-Braem. P., 219, 22X Boynton, R. M . , 124, 127. 143, 144. 145. 147, 148, 150. 151. 171, 1x0 Brackbill, Y . , 135, 1 7 / . 268. 296 Bradbury, H., 158, 1 7 / Bradley, N . C . , 283, 284, 285. 296 Braine, M . D.S., 104, 114 Bransford, J . D.. 219, 227 Braverman, H.,242, 261 Bremner, R. H., 234, 23.5, 262 Brian, C. R., 154, 164, 165, 171 Bridges, A . , 235, 265 Bridges, W . E.. 243, 262 Broadbent, D. E . , 9 5 , 114 Bromberg, W . , 282, 296 Bronfenbrenner, U., 260, 262 Bronson, G., 136, 171 Broughton, J . M . , 118, 171. 184, 185. 226 Brown, A. L . , 212. 214, 226 Brown, C. J.. 157, 171

Indu

Brown, J . L., 127, 171 Brown, M . L . , 3, X, 25, 5X Brown, R., 28 I, 296 Brown, R. D.. 239. 262 Brown, R. W., 152, 171 Bruch, H. A , . 3, 42, 53, 58 Bruner, J . S . , 105, 114, 118, 171 Bryant, P. E , 118. 154, 163. 164. 171 Bubis. E. A., 221. 227 Bukatko. D., 154, 165, 172 Bullis. G. E . , 155, 171 Burgess, A., 12, 5 / Burgess, H. J . L . . 12. 51 Burgos, J . C., 32, 52 Burkc. D.. 105, I14 Burton, R. S.. 32. 52 Bussadori, G . , I S . 5 1

C Caltlwell, B.. 232, 262 Calhoun, A . W., 240262 Calhoun, D., 234,262 Calvin. A . D.. 127, 162. 172 Campione, J . C., 214, 226 Canosa. C . . 6, 51 Cant;ilini. C., 15. 59 Cardu, B., 131. 176 Carey, P.. 272. 273. 296 Carroll, J . B.. 150, 172 Casper, M . . 126. 173 Cavanagh. P., 27 I, 296 Centuribn, C . , 13, 52 Chapinmi, R. H . . 105, 111 Champncys, F. H..125, 172 Chapanis. A,. 139, 167, 172 Chase, H. P., 32. 52 Chase, W. G., 96, 114 Chase, W. P.. 133, 141. 162, 172 Chen. J . S.. 13.52 Chen, M . L . . 13. 52 Cherry, F. F., 30. 52 Chevallerau. A,, 130. 172 Chi. M . T. C.. 105. I14 Chisng, C. H.. 13, 52 Child, 1. L.. 159. 160, 172, 245, 265 Chipnian, S . F., 221, 222, 226 Chiriac. I.. 13, 58 Chiurazzi, L . , 34. 53 Choopanya. K . , 2 I. 54

Author 1nde.r

Chouhan, C. N . , 21, 55 Christ, R. E., 168, 171 Chrzastek-Spruch. H., 18, 19,52 Clanton, C . , 118, 161, 181 Clark, C. E., Jr., 239, 262 Clark, E. V . , 165, 172, 272, 273. 2Y6 Clark, W. H.. 152, 174 Clarke-Stewart, K . A . , 254,262 Clauson, J. A., 255, 262 Clifford, L. T., 127, 162, 172 Coates, B.,216. 217, 226 Cohen, J . . 286, 296 Cohen, L. B.. 118, 123, 152, 153, 154. 164. 167. 172, 178 Cohen. S . , 262 Coker, P. I . , 272, 273, 296 Colavita. F. B.. 117. 172 Colby. M . C . , 165. 172 Collins. A , , 65, 114 Collis, W. R. F . , 6 , 52 Colombini, M . , 15, 55 Conger, J . J . . 123, I 7 8 Conklin. H . C . , 151. 172 Cook. C . , 130. 139, 173 Coonibs. C., 125, 172 Cooper. L., 126. 173 Corah. N . L., 163. I72 Coren, S . . 130, 172 Cornell, E. H . . 148, 172 Coulson, A . H., 137, 174 Court, S. D. M . , 31.56 Courtney, G. R . , 155, 172 C o u i , M . , 15, 52 Coy. J. F . . 16.52 Craik, K. H.. 293, 296 Cranier, P.,22 I. 224, 226 Croiner, R. F., 273, 277, 297 Crowder, A. M . , 288, 289, 290, 297 Cruise, M . 0.. 40, 52 Cumming, W. W., 150, I82 Curri. P., 34.53 Cunis, D. W . , 223, 226

30 I

Ihshiell. J . F., 155. 156, 158. 172 Dattoli. A . , 36. 52 David, J. H. S . , 5. 52 Davidoff. J . . 124, 163, 164, 172 Davidson, M. L . . 27 I , 296 Davila, G. H., 3.57 Day, J., 215, 227 Day, R. H . , 166. 177 Dayal, R. S . , 15,57 Dearborn, G. V . N., 121, 172 DeGala, A . B., 14, SS Degler, C . , 249, 262 DeLoache, J . S . . 152. 173 DeLuca, G . , 15, 52 Dema, I., 6 , 52 DeMause, L., 234,262 Demos, J., 238. 245, 262 Demos, V., 238, 262 Deodhar, N. S . . 18, 20. 38. 57 Desai, P., 23, 52. 58 DeValois, K. K., 139. 140. 150, 151. 160, I73 DeValois, R. L . . 137, 139. 140, 150, 151. 159, 160, 166, 173 Dick, S.. 215. 216. 227 DiGiorgi. R.. 34. 53 Dtnimick. F. L., 130. 144, 173, I80 Dobson, V . . 128, 129. 130, 142, 147, 173 Dodd, C., 119, 173 Ilodds. J. M . , 32, 52 Domilescu, M . . 13. 58 Donoso, G . , 27. 51 Doob, L. W . , 290, 297 Doornbos, L.. 7 , 52 Doris, J., 126, 173 Dorozhnova, K. P.. 36,52 Dreitzel, H . P., 258. 262 Drenhaus. U . , 8. 52 Droz, R . , 2 I I , 226 Dugdale, A . E.. 22. 52 Duke-Elder. S . , 130. 139, 173 Duvall. R., 234, 262

D Daehler. M . . 154, 165, 172 Dale, P. S . . 151, 172 Danzinger, K . , 245, 262 Darby. W. J . , I I, 52 Darwin, C. H . , 119, 125, 172

E Earle, A. M . , 238, 262 Ebhinghaus, H . , 151. 173 EdoLien, J . , I I , 52

Author Index

302

Edwards, A. S., 166, 173 Eiben, 0.. 34.52 Eichberg, R., 165, 181 Ekman, G., 144, 145, 173 Eksmyr, R . , 23, 52 Ekstrand, B., 165, 181 Eliot, W. G . , 241, 262 Entwisle, D. R., 215, 226 Eppright, E. S., 16, 53 Eska, S. J . , 18, 55 Essock, S . M . , 150, 173 Eysenck, H. J . , 157, 158, 160, 17-3

F

Friedman, S., 148, 174 Friedman, W , 272, 273, 280, 282, 284, 285. 295, 296, 2Y7

Frisancho, A. R . , 6, 53 Frost, N . , 215, 227 Fryer, B. A , . 16. 53 Furstenberg, F., 235, 262 Furth, H. G., 196,228

G

Gabrielli, G., 34. 53 Gadir, A . M. A , , 5, 58 Gadlin. H.. 234, 236, 248, 255, 256,262, 264

Fagan, J . F., 133, 134, 145, 173 Fantz, R. L., 148, 153, 155, 161, 163, 173, I77

Farnham-Diggory, S., 163, 164, 165, 173 Featherstone, J., 232, 260, 262 Fegy, M. J., 159, 176 Feinman, S., 215,226 Ferguson, A. D., 30.59 Fernandez, M. D., 22, 57 Fernbdez-Fernandez, M. D., 22, 53 Fernandez, N . A,, 32, 52 Ferreiro, E., 272, 273, 297 Fichsel, H., 136, 174 Fieve, R. R., 166, 177 Fisk, S. C., 9, I I , 13, 14, 52 Fitzgerald, H. E., 268, 2Y6 Fitzhardinge, P. M., 40, 53 Flavell, J. H. 105,114, 225,227, 270, 279, 297

Fleiss, J. L., 166, 177 Flory, C. D., 126, 179 Forbes, A. P., 33, 53 Forgy, C., 79, I15 Forster, E. S., 144, 169 Fox, H. M., 16.53 Foxx, M. M., 252,261 Fraisse, P., 268, 277, 285, 286, 287, 288, 289, 290, 292, 297

Franckx, H., 34,53 Franks, J. J . , 219, 227 Frankel, D. G.,95, 115 Freeman, D. N . , 130, 161, 177 Freidman, E., 287, 297 Friedman, K . C., 279, 285, 286,297

Gaines, R., 141, 163, 174 Garbini, A,, 120, 122, 174 Garcia-Almansa, A,, 22, 53, 5 7 Garcia-Robles, R., 22, 57 Gardner, D. B., 288, 290, 297 Gardiner, P. A,, 166, 174 Garner, W. R . , 163, 174 Garry, P. J . , 9, 13, 5 7 Garth, T. R., 158, 174 Gascon, J., 64, 113 Gastol, B., 1953 Gaylord, H. A., 144, 175 Gelber, E. R., 152, 153, 164, 172 Geldard, F. A., 117, 123, 137, 174 Gelman, R., 84, 114 George, M., 6. 50 Gervasio, C. C., 14, 55 Gerylovov6, A , , 18, 53 Gesell, A., 118, 155, 174 Geubelle, F., 15, 53 Ghai, 0. P., 27, 53 Ghosh, S., 40. 53 Gibson, E. J . , 118, 174 Gibson, J . J . , 166, 174 Gilbert, J . A., 142, 160, 174 Girgus, J. S., 130. 172 Gilliland, A . R., 289, 290, 297 Girshick, M. A , . 47, 56 Glanzer, M . , 152, 174 Glaser, R., 64, 114 Glegg, R. A,, 20.59 Goesling, W . J . , 164, 174 Gokulanathan, K. S., 18, 31,53 Goldtarb, J. L., 268, 284, 259, 290, 297

Aurhor Index

Goldfeld, A. Y., 7, 8.9, 19, 25, 26. 34. 35, 3 , 3n,53 Goldstein, H., 41, 53 Goldstein, P. J . , 130, 161, 177 Goldstone. S . , 268, 284. 288, 289. 290, 297, 298

Goodenough, F. L., 150, 154, 164, 165, 167, 171, 174

Goodnow, J . J . , 118, 174 Goodson, F. E., 166, 177 Gopalan, C., 18,58 Gordon, J . , 124, 143, 144, 145, 147, 148, I 5 I , 171

Gordon, J. E., 3, 42, 53, 58 Gothlin, G. F . , 145, 174 Gottesman, M . , 118. 174 Graham, B. V . , 140, 144, 174 Graham, C. H., I 18, 123, 124, 137. 140, 143, 144, 149, 150, 159, 161. 166, 174, 175 Graham, G. G., 30,51 Granger, G. W . , 160, 174 Gray, W. D., 219, 228 Green, R. Z., 165, 175 Greeno, J . G., 64, 114 Greenspon, T. S . , 151, 171 Gregg, L. W., 163, 164, 165. 173 Grether, W. F., 137, 150, 166, 174 Greven, P., 236, 231, 238, 262, 263 Gross, J . B., 163, 172 Grusec, J . E., 216, 226 Griitzner, P.. 139, 175 Guilford, J . P., 156, 157, 158, 159, 160. 175, 181, 296. 297 Guilford, R. B . , 158, 181 Guillaumin, J., 209, 227 Gurson, C. T . , 22, 56 Gutman, H., 233, 240,263 Guy, K. C., 221,228 Guzman, M. A , , 3, 42,53, 58 Guzman, R. D., 155. 178

H Habicht, J . P . , 5, 59 Hagen, J . W . , 212, 213, 214, 216. 227 Haith, M. M., 123, 176 Hale, G . A , , 165, 175 Hall, G. S., 126, 175

303

Ham. M., 165, 181 Hammes, L. M . , 8, 54 Hansen. J . A , , 159, 160, 172 Hareven, T. K . . 242. 263 Harinasuta, C., 2 I, 22, 56 Harris, P., 162, 175 Harrison, J. F. C., 240, 263 Hartup. W. H., 216, 217, 226 Harvey, R. G., 6, 54 Hawley, T. G., 8, 54 Hay, J . C., 117, 178 Heath, G. G., 155, 164, 166, 172. 175, 181 Hegedus, G., 34, 52 H e l m h o k . H., 162, 175 Heller, C. A,, 8, 54 Helson, H., 160, 175 Hemaidan, N., 14, 51 Hemphill, W., 35,54 Henke, J., 35, 58 Henry, J . , 257, 263 Hermelin, B. M . , 288, 297 Hernandez, L. E . , 30, 54 Hering, E., 151, 162, 175 Hershenson, M . , 125, 156, 175 Hess, V . L., 95, 115 Hiernaux, J . , 10, 54 Hochberg, J . , 154, 175 Hoffman, C. D., 215, 216, 227 Hoffman. M . L., 253, 254, 258. 263 Hoffman, R. F . , 159, 176 Hohle, R. H.. 288, 289. 290.297 Holden, W . A ., 120, 158. 175 Holibka, V . , 18, 54 Holibkova, A.. 18, 54 Holmes, S. K.. 161. 175 Hongthong, K . , 21, 22,56 Hopkins, J . W . , 31, 54 Hornbeck, F. W . , 159, 160, 172 House, B. J . , 95, 116, 163, 182 Howe, L. K., 255, 258, 260, 263 Hsia, Y . , 137, 140. 143. 150, 166,174, 175 Huang, C . S . . 13, 52 Hubbard, M. R . , 145, 173 Hull, E. M . . 137, 173 Humphreys, D. W . . 289. 290,2Y7 Hunt, D.. 235. 263 Hunt, E., 78, 114 Hunt, E. P . , 47, 56 Hunter, R. M., 32, 52 Hurlock, E. B., 123, 175

Author Index

304

Hurst, D. C., 140, 144, 174 Hurvich, L. M., 143, 167, 175 Huttenlocher, J., 105, /f4

I Ilg, F. L., 155, 174 lllsley, R., 6, 58 Indian Council of Medical Research, 35, 46, 54 Inhelder, B., 67, 105, 114, 196, 201, 203, 206,227, 228, 276, 279.2Y7 Inkeles, A , , 239, 240, 263 Ipinza, M., 31, 56 Irwin. O., 126, 175

J Jackson, S . , 66, 114 Jacobs, G . H., 123. 144, 159, 160, 173, 175 Jahoda, G.,268, 285.2Y7 Janes, M. D., 8.54 Jansen, A. A. J., 8, 54 Jastrow, J., 157, 158, 175 Jeffrey, K., 241, 242, 263 Jeffrey, R. W . , 217, 226 Jelezneac, I., 13, 5X Jersild, A. T., 123, 176 Johnson, D., 5.58 Johnson, D. M., 219,228 Jones, A. D., 252, 263 Jones, D. L., 35, 54 Jones, T., 166, 177 Jones-Molfese, V., 156, 176 Jongeward, R. H., Jr., 212, 213, 214, 216, 227 Jonxis, J. H. P., 7, 52 Jordan, T. E., 34, 54 Jorgensen, J. B., 8, 52 Judd, C. H.,248,263 Judd, D. B., 124, 141, 176 Jusczyk, P. W., 221, 227 Jyothi, K. K., 18, 58

K Kagan, J., 123, 148, 177, 178

Kaganovich, D. I.,35, 54 Kail. R. V.. 212, 213, 214, 216,227 Kallal, Z..14, 51 Kalmus. J., 137, 176 Kambara, T.. 15, 41, 54 Kantero, R. L . , 14, 50 Kaplan, B., 295, 2% Kardashenko, V. N., 19, 5 / Karniel. B. Z., 159, 160, 176 Karr, A . C., 166, 177 Kasatkin, N. I.,135, 176 Katz, D., 164, 176 Kaufman, L.. I 18, 123. 176 Kay. P.. 145, 152, 159, /70 Kee. I>. W . , 22 I, 22X Kecle, S . W . . 219. 22X Keller-Cohen, D., 272, 273, 297 Kelton, J . J . , 161, 17.5, 176 Kemler, D. G . , 22 I, 227 Kenney. H., 105, 114 Kerr. G . R., 14, 51 Kessen. W . . 119. 123, 128. 171, 176, 254, 260, 263 Kett. J. F., 238, 263 Kettle, E. S.. 8, 54 Kevany, J. J., 3, 58 Khanjana5thiti. P., 21, 54 Khurana, V . . 42, 54 Kiefer. M.. 250, 263 Kiniura, K.. 30, 46, 54 Kiss, K.. 34, 52 Klahr, D., 64. 77, 84, 104, 114 Klapper. Z. S., 254, 263 Klein, R. E., 5, 59 Klein, R. M . , 117, /7Y Kling, J. W . , 118, 123, 176 Kluckhohn. F. R., 240, 263 Knox, E. G . , 3 I,56 Kock, H. P., 35, 58 Koftka, K., 122, 176 Kogan, R. B., 34, 54 Kondakova-Varlamova, L. P., 19, 51 Kopp. J.. 143. 150, /76 Koriat, A , , 164, 177 Koslowski, B . , 118, 17/ Kosslyn, S. M . , 216, 222, 224.226, 227 Kourim, J., 18. 55 Korhonazarov, K. K.. 9, 54 Kraeba, N . A , . 35. 54 Kram. K. M . , 9, 13, 57

A rithor Index

Kravchenko, A . G . , 19.54 Kreutrer, M . , 225, 227 Kreysler, J . . 14. 55 Krishna, R., 18, 21, 23. 27. 34, 3X. 39. 42, 50. 51

Kublt, K., 18.55 Kudaiarov, D. K.. 9, 55 Kuhn. D.. 74, 115 Kulkarni, H. D., 18, 20. 22. 38. 57 Kumar, R . , 15, 57 Kumar, V.. 32, 52 Kurniewicz-Witczakowa, R.. 18, 19, 55 Kutsenko. V . V . , 19, 35. 36. 55 Kuz’niin. E. M.. 35. 55

L Ladd-Franklin, C., 122, 131, 147. 176 LaFuente, M. E., 27, 51 Lakhani. S. M.. 21. 55 Lakowski, R . , 142, 176 Lambrechts, L . , I S , 53 Lamkin, G. H . , 16, 53 Lane, H.. 143, 150, 176 L.dnatord, : R . , 160. 175 Lapitakii, F. G . , 34. 55 Lashley. K. S., 144, 176 Lasch. C., 232. 241, 244. 263 Laslett , B., 236. 263 Laslett. P.. 233. -763 Lawler. C. O . , 166. 176 Lawler. E. E., 166, 176 Lazar. M.. 153. 164, I72 Lechtig, A., 5 , 5 Y Lee, L. C.. 66, 67, 115, 16.5. 176 Lee, M . , 8 , 50 Leesuwan, V . . 21, 54 Lemond, L. C., 118, 17H Lenneberg, E. H., 149. 152, 171. 176 Leonard, C . , 225, 227 Lepork, F . . 131, 176 Lesi, F.. E. A,. 6, 52 Levikova, A . M . , 135, 176 Levin. H . , 257, 264 Lewis, I . C . , 16, 52 Lewis, M., 119, 173, 176 Lewis, M. M., 271. 281.297 Lhanion, W. T.. 288. 289. 297 Lian, 0. K.. 6 . 5 8

305

Liben, L., 196, 227 Liebert. R. M., 74, 115, 123, 176 Lim. R. K. H.. 22. 52 Limaye, C.. 2 1 , 5 5 Lintz, L. M., 268, 296 Liong-Ong. T. W., 6. 5H Lipsitt. L. P . . 123, 17Y Lipset, S . M.. 249, 263 Little. A . C., 141, 174 Lodge, A , , 131, 136. 167. 170, 176 Loftus, E. F., 220, 227 Loniax, E.. 247. 263 Longmore. E. A , . 16, 52 Longobardi. E. T . . 274, 276. 278, 2Y7 Loosley, E. W.. 240. 264 Lorenz. A . B.. 167, I76 Loveday, T., 144. 16Y Lovell, H.G . , 6, 23, 31, 5 0 Lobell. K., 275. 276. 277. 2YH Lovell, K. A,, 66, 115 Low. W . D.. 30. 55 Lowe, J. E.. 9, 13. 57 Lowenstein. F. W . . 9, 13. 3 2 . 50 Lowenthal, L.. 249. 263 Lubin, A . H . , 9 , 10, 13.57 Lucchetta. G . . 34, 53 Luckey, G . W . A , . 162. 177 Ludbigh, E.. 129, 177 Luna-Jape, H.. 27. 55. 57 Luyken, R.. 7 . 5 5 Luyken-Koning, F. W. M., 7. 55

Mc MacKay. D. A , . 22.52 MacFarlane. A . , 151, 162, 175 McBride, A , , 257. 263 McCain, C . N . , 166, 177 McCall, R. B., 138, 177 M c C m h y , E. F., 129, 177 McClure, W. E.. 167, I76 McDermott, J . , 79. 89. 115 MeDougall. W.. 120, 177 McEwen, W. J . , 14.57 McFarland, W.. 151. 177 McCanity, W. J . . 9, 55 McGregor, I.A., 6. 56, 58 McKenzie. B . E.. 166. 177 McKusick, V . A , , 166, 177

306

Author Index

M Macoby, E. E . , 257. 264 Madhavan, S., 18, 55, 58 Mager, R. F . , 64, 115 Mair. C. H., 16, 52 Maisel, E. B., 159, 160, 176 Majd, M., 33, 53 Major, D. R., 120, 156, 177 Malcolm, L. A , , 3, 5 . 55. 59 Malina, R . , 5, 59 Malnio, R. 9..131, 176 Mandler, G.. 213. 227 Mandler, J . M . , 215, 227 Mandola, J., 166, 167, 177 Mane, S. I . S., 18, 23. 27, 34, 38, 39. 42, 50, 51

Manwani, A . H.. 42, 54 Marg, E . , 130, 161, 177 Marks, L. E., 117, 131, 171, 177 Marmor, G . S., 221, 222, 227 Marrocco, R. T., 160, 173 Marsden, R. E., 120, 134, 177 Martin, E., 152, 177 Martin, R. M . , 152, 177 Martin, W . J . , 8, 42, 56 Martirosian, R. B., 22, 55 Marzot, G . , 15, 55 Matthews. R. W . . 137, 181 Matawaran, A . J . , 14, 55 Mazurczak, T., 18, 19, 55 Mead, W . R . , 137, 140. 173 Mechling, 232 263 Mekanandha, P . , 21.54 Meli, G., 36, 5 7 Melkman, R., 164, 177 Melton, A . W . , 152, 177 Mendelson, M. J . , 222, 226 Mendlewicz, T., 166, 177 Menlove, F. L., 216, 226 Menzies, F.. 31, 56 Meredith, H. V . , 3, 31, 40, 41, 56 Mergen, B., 238, 263 Merkova, A . M . , 7, 8, 9, 19, 25, 26, 34, 35, 36, 37.53 Mervis. C. B.. 219. 228 Meyer, H., 137, 177 Meyers, E. S.. 35, 54 Miall, W . E., 23, 52, 58 Miesowicz, I., 18. 19, 45.55, 56

Migasena. P . , 2 I , 56 Milewski, A , , 135, 153. 164, 177 Millar, S.. 215, 227 Miller, D. J . , 163, 177 Miller, D. R . , 234, 245, 247, 249, 250. 25 I, 254, 255, 263

Miller, F. J . W . , 31. 56 Miller, G. A , , 95. 115, 151, 177 Miranda, H . , 13, 52 Miranda, S . B., 153, 155. 177 Modell. J . , 242. 263 Mot'fett, A . , 152, 164, 179 Mohanta, K. D.. 18. 30. 56 Monckeberg, F . , 27. 51 Monda, M . , 34.52 Montoya, C., 3 I, 56 Moore, J . , 65, 115 Moore, M. K.. 118, I 7 1 Mom. J . O., 27, 55, 5 7 Moreno-Estaban, 9..22.57 Morgan, G. A , , 166, 177 Morgan, H. C., 137, 173 Morley, D. C., 8, 42, 56 Mortison, J . , 39, 56 Moyer, R. S., 221, 222,227 Mueller, C. G . , 127, 171 Munn, N. L., 128. 129, 177 Munsinger, H., 131, 170. 178 Munz, F. W . , 151, 177 Mussen, P. H., 123, 178 Myers, C . S., 120, 122, 155, 178 Myers, G. S., 8 , 5 1 Myers, N . , 154, 165. 172 Myers, N. A , , 152, 178, 215, 228

N Nagel. V . A . , 120, 121, 147. 178 Nayar. S.. 21, 27, 51 Nekisheva, Z. I . , 22, 36, 43,56 Nelson. A. K . , 128, 17Y Nelson, K. E . , 216, 222, 226, 227 Nelson, T. M . , 158, 171 Nemzer. M. P . , 34, 55 Neuman, R. P . , 245, 264 Neumann. C. G., 18,56 Neumann. P. G., 219, 228 Newcomb, C. N.. 136, 167, 176 Newell, A , , 65, 77, 79. 81, 89, 104, 115

Author

Newhall, S. M . , 124, 145, 178 Newton, I . , 144, 178 Neyzi, 0.. 22. 24, 56 Nickerson, D., 124, 145, 178 Nissen. M . J . , 117, 179 Norman, D. A , , 78, I 1 6 Norren, D. V . , 129, 178 Norsworthy. N., 150, 178 Notaney, K . H . , 22,52 Nouth-Savoeun, 15, 56 Novakova, M., 18,55 Nunnally, J . C . , 118. 178

0

1nrle.r

301

Payne. M . C.. 165, 178 Pedron, G . . 30, 58 Peeke, S . C., 164. 178 Peeples. D. R . , 126, 128, 129, 134, 135, 136, 137, 138. 140, 141. 132, 147, 167. 178, 181 Peiper, A . . 122. 126, 128. 129, 131, 135. I78 Peltzman, P., 130, 161. 177 Periera, S . M . , 23, 29, 57 Perlmutter, M., 152, 178, 215, 228 F‘hadke, M., 18. 20, 21, 22, 38,55, 57 Piaget, J . . 67, 74, 105. 114. 115, 185, 189, 191. 195. 196. 198, 201. 203, 204, 206, 207. 2 I I . 228, 270, 21 I, 212, 214. 275, 216, 271. 278. 279. 28 I 283, 284. 285, 286. 2x7, 288. 290, 291, 292, 294. 295. 296. 297. 2Y8 Pick, A. D . . 95, 115 Pick, H. L . , 117, I 7 8 Pire, E . , 36, 57 Platt, E. M.. 255, 265 Podgorny. P.. 22 I, 228 Pogorely. I . A , . 34. 55 Pokatilo, V . M . , 19, 38. 57 Polack. 4.. 130. 172 Polikanina, R. I., 136. 17Y Pollack. R. H . , 161, 175, 176, 179, 180 Polson. M. C.. 137. 173 Polyak, S., 118, 179 Pongpaew. P., 21, 22. 56 Poresky. R.. 126, 173 Porter, E . , 158. 174 Posnansky. C. J . , 219, 228 Posner. M. I . . 117, 152. 17Y. 219. 228 Potter, D.. 233. 264 Poulos, R. W . , 123. 176 Praliaraj, K . C.. 18, 30. 56 Prasad. R . , 15, 5 7 Pratt, K . C . , 128, 17Y Preyer. W . . 119, 126. 163. 17Y Prokhorova. M . V . . 19. 5 / Pylyshyn, Z. W . . 223. 228 PyE~ik,M . , 18, 56. 57, 5Y ~

Oakden, E. C., 282, 283, 284, 285, 298 O’Brien, R., 47, 56 Ochmisch, W . , 14. 5 7 Ochsenfarth, A , , 35,58 O’Connell, D. E . , 9, 13. 32, 50 O’Connor, N . , 288,297 Odom, R. D., 155, 178 Olson, G . M.. 214. 216, 227 Omran, A. R., 14, 5 7 O’Neill. N., 255, 264 O’Neill. 0.. 255, 264 Oppers, V . M.. 35, 57 Ordy, J . M . , 161. 173 Ornstein. R . , 286, 287, 290, 293, 296, 2YH Osgood, C. E.. 118. 178 Oster, H.. 135, I 7 8 Owen, G. M.. 9, 13.57 Oyarna, T., 166, 178

P Paibio, A , , 221, 222. 227. 22H Palacios, J . M., 22, 5 7 Palacios-Mateos, J. M., 22, 53, 57 Pancratz. C. N.. 154, 178 Pande, T. H., 18, 30, 56 Pardo. F.. 27, 55, 5 7 Pardu, K.. 164. 177 Parke, R . D., 118. 181, 254, 259. 264 Parry, M . H., 153, 158, 17Y Pascual-Leone. J . A., 78. 95. 115 Paul, I . H., 218, 228 Paulos, M. A , , 223, 226

Quek. K. M . , 15, 5Y

308

Author Index

R Rahman, A. K., 6, 56, 58 Rahmy, M., 21 I,226 Raj, L., 18, 21, 23, 27, 34, 38, 39. 42,50, 51

Rao. N. P., 18, 20.57 Rapson. R. L., 239,264 Rasmussen, T., 131, 176 Ratkowsky, D. A,, 16, 52 Ray. V . F.. 150, 179 Rea. J. N., 18.57 Reddy, R., 65, 115 Redfield, J., 126, 179 Reese. H. W.. 123, 179 Regan, D., 167, 179 Reiner, M. L., I I.58 Resnick, L. B., 64, 65. 115 Richardson. B. D., 18, 25, 57 Riegel, K. F.. 184, 226. 280,2Y8 Riesnian, D., 248, 249, 264 Riggs, L. A., 118, 123, 176 Roback, A . A , , 185.228 Robertson. J. B . , 165, 172 Robinson, J. P., 256, 264 Roche, A. J., 3, 5 7 Rogers, C., 255. 264 Rohwer, W. D.. 22 I , 228 Ronaghy, H. A , . 3 3 , 5 3 Rosa, I. R.. 32, 52 Rosch, E., 219. 228 Rosch, E. H., 152, 179 Rose, C. S., 6, 58 Rosenherg, J. H.,I I, 5 7 Rosinski, R. R., 123, 179 Ross, B. M.. 196,228 Rothman, D. J . , 234, 235, 264 Rubin, S.. 248, 255, 256, 262 Ruddock, K . H.. 130, 17Y Rueda-Williamson, R.. 27, 55. 5 7 Rule, S. J., 223, 226 Rutishauser, I . H. E., 30, 57 Ruzskaya, A. G., 151, 182 Rychener, M. D., 79, 115

S Saayman, G . , 152, 164, 179 Sabatier, J . , 15, 53

Sackett, G. P.. 137, 139, 170 Salapatek, P . , 118, 123, 128, 161. 162, 167, 170, 172, 176. 179

Salomon, J . B., 6, 51 Saltzstein, H . D., 254. 258. 263 Sandhu, R. K., 27, 53 Saner, G.. 24, 56 Sarhin. T. R., 293, 2Y6 Sauberlich, H. E., 32, 52 Sawrey, J. M., 123, 176 Schaffer, H. R., 153, 158, / 7 9 Schallenberger, M . , 120, / 7 9 Schaller, M. J . , 135, 136, /7Y Schechter, D. E.. 282, 283, 298 Schlage, C., 14, 55 Schneibner, H. M. 0.. 150, 179 Schoen, E. J.. 25, 41,59 Scott, E. M., 8, 54 Scott. J . A , , 34, 58 Scott, R. B., 30, 59 Scrimshaw, N . S.. 3. 42, 53, 58 Sears, R. R.. 247, 257. 260.264 Seeley, J . , 240, 264 Seely, P.. 272, 273. 278,297 Sekel, M., 137, I 8 1 Senipe, M., 30, 58 Sempe, P., 30.58 Senn, M . J . E.. 234, 260, 264 Sevcikovi. A , , 19.50 Sewny, V . D., 188. 228 Sgromskaia, E. P.. 19, 5 1 Shakir, A , , 29, 58 Shah. P. M., 18, 27, 58 Shanker, H., 18.56 Shanks, B. L.. 136. 167, 176 Sharma, J . C., 18. 30. 58 Shattock. F. M . , 7, 58 Shearron, G. F.. 166, 167, 179 Sherman, I . C., 126, 179 Sherman, M., 126, 17Y Shepard. R. N., 221, 222. 224, 228 Shiddhaye, S., 18, 27.58 Shinn, M. W., 120, 151, 163. 179 Shirley. M. M . . 122, 179 Shorter. E., 233, 236, 254, 255, 258, 264 Siegel, D. E., 140, 152, 179 Siegel. M. H., 152, 179 Siegler, R. S., 64, 66, 67, 69, 74, 75, 95, 96, 97, 105. 115

Sim, R. O., 240. 264

309 Simon, H . A , . 81, 96, 114. 115 Simon. T., 122, 170 Sinclair. H . , 212. 273, 2Y7 Singh, D., 18. 20. 57 Singh. R . . 18. 5X Sinsheimer, R . L . . 117, /KO Sinopalnikov, 0. Y . , 34. 55 Sinson. J . , 155. / X U Siqueland, E. R . , 135. 153. 164, 177 Skoff, R . , I6 I, I80 Skrobak-Kacznski, J . , 8. 52 Slater, A . , 215. 216, 211. 29X Sloan. L. L . , 166. I80 Sniedslund, J . , 104. 116 Smith. D. P.. 140, 144, 180 Smith. D. S . , 3, 8, 2 5 , H Smith, H. C.. 122, 163, 1x0 Smith, J . M.. 126, 128. 129. / X U Smith, M., 234, 264 Smith, P. C. A , . 160, 175 Sniythe. E. J . . 288. 289. 2Y8 Snyder. Y. N., 166, 1x0 Sokolow, Y . N., 153, 180 Spalding, V . , 32, 5 2 Spaner, S. D.. 34.54 Spears, W . C., 133, 134. 155. 156. 1%. 163, 180 Spekreijse. H., 167. 174, Sperling, H. G . , 137, 174 Spranger. J . , 35. 5X Springer. D., 282. 283. 298 Sprug. J.. 261, 2YX Sriniusikapodh. V . . 21, 54 Srivastava, G .. 42, 54 Stacey, J. T . , 196. 228 Stinciulescu, E., 13, 58 Staples, R . . 123. 155, 1.56. 1%. I80 Stare, F. J . . 14, 51 Staub, E . , 254, 251, 259, 264 Stave, M . , 164, 165, 1x1 Steinhardt, J., 127, 1x0 Stendler, C.. 245, 246. 247.264 Stem, W.. 121, 151, 163, 180, 281. 298 Sternheini. C. S . . 145. IN0 Steven, E. M., 40. 53 Stewart. A , , 252, 263 Stiles, W . S., 118. 123, I82 Stiinpll, J . , 120, 1 x 1 Stimimann, F., 158, 159, I N 0 Stone, G . C., 164, 17X

Stracker, 0. A , . 36. SX Strauss, G . D., 123, 176 Strickland, C., 2 3 8 , 264 Strong, B.. 23Y. 264 Stubbs. D. F.. 160. I80 Suchman, R. G.. 154, 164, 165. 1x0 Sttirt, M . , 282, 283. 284, 285, 2YX Sukkar, M. Y.. 5. 5X Sun, K . H . , 128, 174, Supachaturas, P., 2 I , 54 Susheela, T. P.. I X . 55 Swaak. A . J . . 41. 5X Swaminathan, M. C . , 18. 20, 55. 57, 58 Swanson. G. E . . 234, 245. 241. 249. 250, 25 1 , 254. 255. 263 Symonds, M . , 282, 283, 298 Standard. K. L.. 23. 52, 5K Stannard. D.. 233. 238, 264 Syrovitka. A , , I X . 55 Sysfeins Development Project Stall, 32, 36, 5x Szajner-Milait. I . . I X . 19, 52

T Taback, M.. 40. 58 Takaishi, M . . 34. 58 Tinisecu. C.. 13, 58 Tanman, F., 23, 56 Tanner, J. M . , 34, 5X Tasker, A . D.. I X . 3 8 . 42. 51 Tasnidi. I . . 34. 52 Tataliore. E.. 34, SX Teill'er. E.. 215. 228 Teixeira, G., 30. 59 Telford. C. W . , 123. /76 Teller. D. Y . , 126. 128, 129, 133. 135. 136, 137, 138. 13Y. 141, 142. 147. 170, / 7 x , 181

Terrosi. F . . IS, 58 Tesi, G . , 14. 5 / Thomas. H.. 12.5, 156, lXl Thomson. A . M . . 6. 56, 58 Thompson, B . . 6. 56, 5X Thompson. E. P., 234. 264 Thuline, H. C.. 166. 181 Tliurnhani. D. I.. 21, 22, 56 Tie. L. T . . 6, 5X Tier, H., 130, / H I

310

Author Index

Tonelli, E., 15, 59 Trabasso, T., 154, 164, 165, 180, 181 Tracy, F., 120, 181 Trincker, D., 128, 129. 133, I81 Trincker, I., 128, 129, 133, 181 Tronick, E., 118, 161, 162, 181 Tsai, C. M., 30,54 Tseimlina, A. G . , 7, 8, 9, 19, 25, 26, 34, 35, 36, 38,53 Turiel, E., 74, I16 Turki, M . , 14, 51 Turner, M. E., 140, 144, 174 Tuxford, A . W . , 20.59 Tversky, B . , 215, 228 Twiesselmann, F., 34, 59 Twombly, R. L., 244, 264 Tye, C. Y . , 15,59

U Uberoi, I . S . , 18, 56 Udani, P. M . , 18, 21, 23, 27, 58, 59 Udelf. M. S . , 161, 173 Underwood. B. J . , 165, 181 United States Center for Disease Control. 32,59

v Vago. S . , 105, 115 Valentine, C. W., 120, 122, 163, I81 van der Berg, B. J . . 25. 51 Van der Kuyp, E., 7, 8, 14.59 van Hekken, S. M. J . , 216,229 Vannas, S . , 130, 181 van Wieringen, J . C., 14, 59 Vedovello, R., 15. 55 Verghese, K. P . , 18, 30, 31, 5 3 , 59 Vernon, M . D., 121, 159, 181 Verriest, G . , 142, 181 Vesi, G . , 15, 59 Vincent, C., 245, 260, 264 Vis, H. L., 9, 59 Visser. H . K. A , , 7. 52 Vivanco, F., 22. 57 Vivian, V . M . , 16, 53 Vizzoni, L., IS, 59 von Frisch, K., 135, 150, 181

von Helverson, 0.. 140, 181 Vos, J . J . , 129, 178

W Wade, M., 144, 176 Wald, G . , 127, 181 Walkowitz, D. J . , 244, 264 Wall, R., 233, 263 Wallace, J . G . , 64, 84, / I 4 Walls, G . 1.. 118, 130, 137, 151, 166, 168, 181

Walters, R. H.. 118, I8f Walton, W. E., 158, 181 Walton, W. S., 31,56 Wark, L., 5, 59 Warren, D. H., 117, 178 Warren, J . M . . 164. 181 Waterman, D. A., 104, 116 Watkin, D., 13, 52 Watson, J . B., 126, 181, 243, 264 Watts, J. E., 254. 265 Waugh, N. C . , 78, 116 Weale, R. A . , 130, 181 Weil, J . . 274, 276, 278. 294, 298 Weinbaum, B . , 235, 265 Weinstein. F., 255, 265 Weisman, S. A , , 31. 59 Weiskopf, S., I 19, 128. 171 Weiss, L., 126, 175 Welch, M. J . , 152, 182 Wellman, H. M . , 225,227 Welter, 0 . . 243, 265 Werner, H . , 164, 182, 295, 298 Wetherick, N. E., 155, 180 Wheeler, E. F., 12. 51 White, B. L., 254, 265 Whitehouse, R. H., 34, 58 Whiting, J . , 245, 265 Whitley, M. T . , 150, 178 Whorf. B. L., 150, 182 Wickelgren, L. W., 134. 182 Wickelgren, W . A . . 214,229 Wiebe. R . , 240, 265 Wiggin, K., 248. 251, 265 Williams, A , , 6 , 50 Williams, H . . 166. 182 Williams, L. G . , 161, 182 Wilson, R. S., 40, 59

31 I

Aurhor Index

Wingerd, J . , 25, 40, 41. 59 Winter, D. G., 252, 263 Wishy, B., 238, 265 Wittgenstein, L., 224, 225, 229 Wohlwill, J . , 286, 298 Wolanski, N., 18, 57, 59 Wolfenstein, M., 245, 246. 251, 252. 265 W O I ~ TP. , A , , 274, 276. 278,297 Wolter, J . R . , 130, 170 Wong, H . B., 15, 59 Wood, -D., 278, 288,296 Woodbury, R. M . , 5, 20.59 Woodland, M . , 8, 42,56 Wooley, M . T., 120, 182 Wooten, B. R . , 128, 129, 134, 182 Wright, A . A , , 130, 140, 144, 150, I82 Wright, W. D., 129, 140, 144, 167, 182

Wyszecki, G.. 118, 123, 182

XYZ Yamamura, T., 166, 178 Yarbrough, C., 5 . 59 Yendovitskaya, T. V.. 151. 182 Young, R . M . , 64, 116 Yousif, M . K . , 5, 58 Zaini, S . , 29, 58 Zaki, M . H., 14,57 Zauli-Naldi, G., 15, 59 Zeaman, D., 95, 116, 163, 182 Zelkind, I . , 267, 298 Zhaglina, A. K . , 35, 59 Zinchenko, V . P., 151, 182 Zolotareva, M. P., 34, 55

This Page Intentionally Left Blank

SUBJECT INDEX A Accommodation, memory and, 190-193 Assimilation, memory and, 190-193 Awareness, memory and, 205-208 topical research in, 224-226

B Birth(s), single and twin, standing height and, 4 0 41 Birth order, standing height and, 40 Birth weight, standing height and, 3 9 4 0 Brightness, perception of, 125-127

Ethnic groups, standing height compared among, 3-1 1

F Family, ideology of, discipline and, 240-244

H Health, standing height and, 4 1 4 2 Health care, standing height and, 4 2 4 3 Hue, organization of, 144-151

I C Child rearing, 244-248 discipline and, 248-259 Color preference, 155-161 Color vision, 117-1 19 constancy, 162 defects in, 166-167 ontogeny of, 119-125 brightness and spectral sensitivity, 125-132 chromatic vision, 132-167 perceptual cues in, 165-166 perceptual salience and, 162-165 peripheral, 161-162

D Decision tree, 73-74 Discipline colonial, 236-238 ideology of family and, 240-244 in Jacksonian era, 238-240 modern, 248-259 in nineteenth century, 235-236

Imagery, memory and, topical research in, 220-224 Imitation, memory and, 188-190 topical research in, 214-217

K Knowledge behavioral and cognitive objectives and, 63 -64 representation of, 66-69 assessing initial knowledge and, 69-74 choosing, 64-65 decision tree, 73-74 encoding and, 95-102 evaluation of, 94-95 production system, 77-82, 85-94 protocol analysis, 82-85 training and, 74-77 L

Language, memory and, 204-205

M

E Elevation, standing height and, 41 Encoding, knowledge representations and, 95-102

Memory affective, 193-195 assimilation and accommodation and, 190193 313

3 14

Subjecr /ndex

Memory (continued) assumptions and logical distinctions, 186-188 awareness and, 205-208 color vision and, 151-155 comparison of theories of, 210-212 imitation and, 188-190 language and, 204-205 recognition, 193-1 95, 199-20 1 reconstructive, 201-204 schemes, 198-199 topical research in awareness and, 224-226 imagery and, 220-224 possible theoretical contacts and, 2 1 2-214 recognition and imitation and, 214.-217 schema theory, 217-220

N Nationality, standing height and, 1 1 - 1 6 Neutral zone, 136-140

P Pattern vision, 117-1 19 Physical abnormality, standing height and, 41

S Sequence, concept in infancy, 270-272 Socioeconomic status, standing height and, 20-25 Smoking. maternal, standing height and, 4 0 4 1 Spectral sensitivity, 127-132 Standing height birth weight and birth order and, 39-40 ethnic group comparisons of, 3-1 I female and male groups compared on, 36-39 intercity comparisons of, 25-36

intracity comparisons of, 33-36 ethnic group and, 25-27 socioeconomic status and, 27-33 maternal smoking and, 4041 national and intranational comparisons of, 11-16 physical abnormality, elevation, health status, and health care and, 4 1 4 3 population variability estimates of methods, 4 3 4 5 standard deviation in, 4 5 4 8 rural and urban comparisons of, 17-20 sscioeconomic group and, 20-25 single and twin births and, 4 0 4 1

T Time concepts conventional of 5- to 7-year-olds, 283-284 in infancy and early childhood, 280-283 in middle childhood and adolescence, 284-286 experiential, 286-287 complex judgments, 290-292 development of, 292-293 simple judgments, 287-290 logical, 269-270 in early childhood, 272-274 in infancy, 270-272 in older children, 279-280 in school-age children, 274-279

V Visual acuity, chromatic, 161 W Wavelength discrimination, 140-144

Contents of Previous Volumes Volume I Responses of Infants and Children to Complex and Novel Stimulation Gordon N . Cantor Word Associations and Children’s Verbal Behavior David S . Palermo Change in the Stature and Body Weight of North American Boys during the Last 80 Years Howard V . Meredith Discrimination Learning Set in Children Hayne W . Reese Learning in the First Year of Life Lewis P . Lipsitr Some Methodological Contributions from a Functional Analysis of Child Development Sidney W . Bijou and Donald M. Baer The Hypothesis of Stimulus Interaction and an Explanation of Stimulus Compounding Charles C . Spiker The Development of “Overconstancy” in Space Perception Joachim F . Wohlwill Miniature Experiments in the Discrimination Learning of Retardates Betry J . House and David Zeuman AUTHOR INDEX-SUBJECT

INDEX

Volume 2 The Paired-Associates Method in the Study of Conflict Alfred Castaneda Transfer of Stimulus Pretraining in Motor Paired-Associate and Discrimination Learning Tasks Joan H. Cantor The Role of the Distance Receptors in the Development of Social Responsiveness Richard H . Walters and Ross D. Parke Social Reinforcement of Children’s Behavior Harold W . Stevenson

Delayed Reinforcement Effects Glenn Terrell A Developmental Approach to Learning and Cognition Eugene S. Collin Evidence for a Hierarchial Arrangement of Learning Processes Sheldon H . White Selected Anatomic Variables Analyzed for Interage Relationships nf the Size-Size, SizeGain, and Gain-Gain Varieties Howard V . Meredith AUTHOR INDEX-SUBJECT

INDEX

Volume 3 Infant Sucking Behavior and Its Modification Herbert Kaye The Study of Brain Electrical Activity in Infants Robert I . Ellingson Selective Auditory Attention in Children Eleanor E . Maccoby Stimulus Definition and Choice Michael D. Zeiler Experimental Analysis of Inferential Behavior in Children Tracy S . Kendler and Howard H . Kendler Perceptual Integration in Children Herbert L. Pick, Jr., Anne D . Pick, and Robert E. Klein Component Process Latencies in Reaction Times of Children and Adults Raymond H . Hohle

Volume 4 Developmental Studies of Figurative Perception David Elkind The Relations of Short-Term Memory to Development and Intelligence John M . Belmont and Earl C . Butterfield 315

316

Contents of Previous Volumes

Learning, Developmental Research, and Individual Differences Frances Degen Horowitz Psychophysiological Studies in Newborn Infants S.J. Hutt, H.G. Lenard, and H.F.R. Prechtl Development of the Sensory Analyzers during Infancy Yvonne Brackbill and Hiram E . Fitqerald The Problem O f Imitation Justin Aronfreed

Volume 5 The Development of Human Fetal Activity and Its Relation to Postnatal Behavior Tryphena Humphrey Arousal Systems and Infant Heart Rate Responses Frances K. Graham and Jan C. Jackson Specific and Diversive Exploration Corinne Hutt Developmental Studies of Mediated Memory John H. Flavell Development and Choice Behavior in Probabilistic and Problem-Solving Tasks L. R. Goulet and Kathryn S . Goodwin AUTHOR INDEX-SUBJECT

INDEX

Volume 6 Incentives and Learning in Children Sam L. Witryol Habituation in the Human Infant Wendell E. Jeffrey and Leslie B . Cohen Application of Hull-Spence Theory to the Discrimination Learning of Children Charles C. Spiker Growth in Body Size: A Compendium of Findings on Contemporary Children Living in Different Parts of the World Howard V . Meredith Imitation and Language Development James A. Sherman Conditional Responding as a Paradigm for Observational, Imitative Learning and VicariousReinforcement Jacob L. Gewirtz AUTHOR INDEX-SUBJECT

INDEX

Volume 7 Superstitious Behavior in Children: An Experimental Analysis Michael D. Zeiler Learning Strategies in Children from Different Socioeconomic Levels Jean L. Bresnuhan and Martin M . Shapiro Time and Change in the Development of the Individual and Society Klaus F. Riegel The Nature and Development of Early Number Concepts Rochel Gelman Learning and Adaptation in Infancy: A Comparison of Models Arnold J . Sameroff AUTHOR INDEX-SUBJECT

INDEX

Volume 8 Elaboration and Learning in Childhood and Adolescence William D. Rohwer, Jr. Exploratory Behavior and Human Development Jum C . Nunnally and L. Charles Lemond Operant Conditioning of Infant Behavior: A Review Robert C. Hulsebus Birth Order and Parental Experience in Monkeys and Man G.Mitchell and L. Schroers Fear of the Stranger: A Critical Examination Harriet L. Rheingold and Carol 0.Eckerman Applications of Hull-Spence Theory to the Transfer of Discrimination Learning in Children Charles C . Spiker and Joan H . Cantor AUTHOR INDEX-SUBJECT

INDEX

Volume 9 Children's Discrimination Learning Based on Identity or Difference Betty J . House, Ann L . Brown, andMarcia S . Scott Two Aspects of Experience in Ontogeny: Development and Learning Hans G.Furth

Contents of Previous Volumes

The Effects of Contextual Changes and Degree of Component Mastery on Transfer of Training Joseph C. Campione and Ann L. Brown Psychophysiological Functioning, Arousal, Attention, and Learning During the First Year of Life Richard Hirschman and Edward S . Katkin Self-Reinforcement Processes in Children John C. Masters and Janice R. Mokros AUTHOR INDEX-SUBJECT

INDEX

Volume 10 Current Trends in Developmental Psychology Boyd R. McCandless and Mary Fulcher Geis The Development of Spatial Representations of Large-Scale Environments Alexander W . Siege1 and Sheldon H. White Cognitive Perspectives on the Development of Memory John W . Hagen, Robert H . Jongeward. Jr., and Robert V . Kail, Jr. The Development of Memory: Knowing, Knowing About Knowing, and Knowing How to Know Ann L. Brown Developmental Trends in Visual Scanning Mary Carol Day The Development of Selective Attention: From Perceptual Exploration to Logical Search John C . Wright and Alice G . Vliefstru AUTHOR INDEX-SUBJECT

INDEX

Volume 11 The Hyperactive Child: Characteristics, Treatment, and Evaluation of Research Design Gladys B. Barley and Judith M. Leblunc

A E

8 9

D € F G H I 1

1 2 3 4 5 6 7

c o

317

Peripheral and Neurochemical Parallels of Psychopathology: A Psychophysiological Model Relating Autonomic Imbalance toHyperactivity , Psychopathy, and Autism Stephen W . Porges Constructing Cognitive Operations Linguistically Harry Beilin Operant Acquisition of Social Behaviors in Inlancy: Basic Problems and Constraints W . Stuart Millar Mother-Infant Interaction and Its Study Jacob L. Gewirtz and Elizabeth F. Boyd Symposium on Implications of Life-Span Developmental Psychology for Child Development Introductory Remarks Paul B. Baltes 'Theory and Method in Life-Span Developmental Psychology: Implications for Child Development Aletha Huston-Stein and Paul B. Baltes The Development of Memory: Life-Span Perspectives Hayne W . Reese Cognitive Changes during the Adult Years: Implications for Developmental Theory and Research Nancy W . Denney and John C. Wright Social Cognition and Life-Span Approaches to the Study of Child Development Michael J . Chandler Life-Span Development of the Theory of Oneself Implications for Child Development Orville G . Brim, Jr. Implications of Life-Span Developmental Psychology for Childhood Education Leo Montada and Sigrun-Heide Filipp AUTHOR INDEX-SUBJECT

INDEX

This Page Intentionally Left Blank